Bioresource Technology 172 (2014) 150–155
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Sequential operation of a hybrid anaerobic reactor using a lignocellulosic biomass as biofilm support Mohamed Ali Wahab a,b, Frédéric Habouzit b, Nicolas Bernet b, Jean-Philippe Steyer b, Naceur Jedidi a, Renaud Escudié b,⇑ a b
University of Carthage, Water Research and Technologies Centre (CERTE), Wastewater Treatment Laboratory, Tunisia INRA, UR50, Laboratoire de Biotechnologie de l’Environnement, Avenue des Etangs, Narbonne F-11100, France
h i g h l i g h t s Hybrid anaerobic reactor operated sequentially to treat liquid and solid waste. Lignocellulosic biomass used as both carrier support and substrate. High treatment efficiency during fast start-up and restart-up phases. Acclimation phase depends of the soluble fraction of lignocellulosic support. During non-feeding period, biofilm conserved its biological activity.
a r t i c l e
i n f o
Article history: Received 17 June 2014 Received in revised form 28 August 2014 Accepted 31 August 2014 Available online 8 September 2014 Keywords: Anaerobic digestion Lignocellulosic biomass Biofilm Start-up Hybrid reactor
a b s t r a c t Agro-industries are facing many economic and environmental problems associated with seasonal generation of liquid and solid waste. In order to reduce treatment costs and to cope with seasonal variation, we have developed a hybrid anaerobic reactor operated sequentially by using lignocellulosic biomass (LB) as biofilm carrier support. Six LBs were tested to evaluate the treatment performance during a succession of two start-up periods, separated by a non-feeding period. After a short acclimation phase of several days, all the reactors succeeded in starting-up in less than 1 month to reach an organic loading rate of 25 gCOD L 1 d 1. In addition, they restarted-up successfully in only 15 days after a 3 month non-feeding period, indicating that biofilms conserved their biological activities during this last phase. As a consequence, the use of LB as a biofilm support gives the potential to sustain seasonal variations of wastewater loads for industrial application. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The agro-industries are considered as water and energy intensive consumers and their activities have been always associated with economic and environmental problems due to the generation of large amounts of wastewater and solid waste, in addition to the increase of their carbon footprint. Thus, agro-industries should focus on improving their energy efficiency, water use and waste management. Hence, in a world where water scarcity and climate change are a reality, actions to protect the environment and enhance renewable energy are mandatory for any type and size of industries.
⇑ Corresponding author. Tel.: +33 468425173. E-mail address: [email protected] (R. Escudié). http://dx.doi.org/10.1016/j.biortech.2014.08.127 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
The high biodegradability of agro-industries waste often justifies the use of conventional biological treatments. Among them, anaerobic digestion (AD) is the most adequate technology for the treatment of wastewaters with high organic carbon concentration. AD ensures the degradation of organic matter simultaneously into a valuable biogas composed of methane (CH4) and carbon dioxide (CO2) and into a nutrient-rich digestate with agronomic qualities (Karthikeyan and Visvanathan, 2012). A major limitation of AD process is the long duration and instability of their transition phases (starting, re-starting, organic load increase) because of the low growth rates of anaerobic microorganisms and their sensitivity to perturbations such as organic overloads (Arnaiz et al., 2003; Cresson et al., 2009; Escudié et al., 2011). Moreover, application of AD for the treatment of agro-industrial waste and wastewater is hindered by several limitations, such as the heterogeneity and the seasonal supply of waste, low biogas production rate and high
151
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
investment cost (Ren and Wang, 2004; Sasaki et al., 2009). Thus, it is of high interest to develop an AD reactor with low investment and operational costs that could ensure the treatment of both liquid and solid waste in one single reactor with high treatment efficiency and high methane yield. The operation of this reactor has to tolerate seasonal fluctuations and OLR perturbations that can happen in agro-industries over the year. Anaerobic fixed bed reactors (AFBR) present a good potential for the treatment of wastewaters. In fact, they are known to obtain high solid retention times for high system efficiency and stability, with low hydraulic retention times for system economy. Furthermore, these processes are inherently stable and resistant to organic and hydraulic shock loadings (Zhao et al., 2013). AFBR performance is strongly depending on the nature of the support material (Habouzit et al., 2014). Different natural and artificial solid materials can be used as biofilm carriers. Commercial carrier elements which are usually made of polyethylene, polypropylene or polyurethane, are expensive, while natural carriers are essentially inorganic such as sand, gravel, pumice stone, porous glass beads and zeolite (Tarjányi-Szikora et al., 2013). Natural lignocellulosic biomasses (LBs) can be an alternative to these inorganic or plastic support materials. LBs are abundant in agro-industries and represent the main part of the solid waste produced. They have low costs, while having high porosity, low specific gravity and higher bacterial adsorption/adhesion (Nabizadeh et al., 2008). The LBs are essentially composed of cellulose, hemicelluloses and lignin in different proportions. The variation of their chemical composition can influence the treatment performance and methane production of the reactor, especially during the start-up period which constitutes an important step in the anaerobic biological treatment process (Cresson et al., 2009; Habouzit et al., 2014). However, the efficiency of LB to be used as biofilm carrier has not been well studied. Andersson and Björnsson (2002) and Mshandete et al. (2008) have investigated the use of wheat straw and sisal fiber waste, respectively, as biofilm supports in methanogenic bioreactors digesting crop waste leachate, in a two-stage anaerobic digestion. To our knowledge, there is no study which has investigated the efficiency of LB for the treatment of high strength wastewater in one-stage reactor. Moreover, because of the seasonal production of effluents in the agro-industries (Zhao et al., 2008), the reactor operation is seasonally unstable. In fact, when the process is not fed (i.e., ‘‘inactivity’’ period), the biological activity of the biofilm is affected and the reactor becomes difficult to be restarted again: this sequential operation can lead to the low removal of organic matters or to a prolonged period of acclimation (Sowmeyan and Swaminathan, 2008). The use of LB as a biofilm support could be a very interesting solution to keep the biofilm biologically active during the inactivity period. The main objective of this study was to explore the potential and performance of using LB as biofilm support and for the seasonal treatment of high strength wastewater. Since the start-up phase is a critical step in the process operation, we have first investigated the efficiency of using six LBs (having different chemical compositions and biodegradabilities) as biofilm support during this decisive period. These reactors were restarted after an inactivity period of 3 month in order to simulate a sequential operation of the process.
2. Methods 2.1. Characterization of lignocellulosic biomasses Six LBs were used to investigate the efficiency of using natural materials as biofilm carrier supports. LBs used were: wheat straw
(WS), sunflower stalk (SS), grape stalk (GS), cactus fiber (CF), luffa fiber (LF) and cypress cones (CC). These LBs were used without grinding or any other pretreatment. The following methods were applied in order to identify their biochemical characteristics and their biodegradability. The Total Solid (TS) and Volatile Solid (VS) contents were measured as described previously by Motte et al. (2013). The distribution of lignocellulosic compounds (soluble compounds, cellulose, hemicelluloses and lignin) was estimated according to a Van Soest fractionation adapted by Motte et al. (2014). The LBs were firstly grinded and sieved to achieve a particle size of 1 mm. Biomethane potential (BMP) of each LB was evaluated by following the methane produced under mesophilic conditions (35 °C). The inoculum was an industrial sludge sampled from a UASB (Upflow Anaerobic Sludge Blanket) process treating sugar factory effluent. The assays were conducted in triplicate in 600 mL sealed flask (working volume of 400 mL) according to the procedure described by Motte et al. (2013). The methane production is expressed under standard condition and accounts for the variation of the gas content in the headspace of the reactors. The biochemical composition and the biomethane potentials of the six LBs are reported in Table 1.
2.2. Laboratory-scale reactors 2.2.1. Experimental set-up The anaerobic fixed bed reactor employed in this study is a PVC column with an internal diameter of 0.20 m and a height of 0.56 m. The total and working volume of the reactor were 16.7 L and 14.8 L, respectively. The reactor was equipped with hot water jackets to maintain a mesophilic temperature of 35 °C. Six reactors were packed with the different LBs used as carrier supports for biomass immobilization and retention. The influent was pumped to the top of the reactor by means of a peristaltic pump. The reactor liquid was recirculated within the reactor by means of a recirculation pump fixed at the bottom inside the reactor. The effluent was discharged at the top through a U-tube for separation of gas. pH was measured in the liquid outlet with a Mettler Toledo 1100 Calimatic pH meter, regulated at 6.8 by automatic NaOH addition. Biogas produced passed through a moisture trap and then to a milligas counter fitted with a 4–20 mA output (Ritter, Bochum, Germany).
2.2.2. Influent preparation and reactor inoculation The influent was a wine-based reconstituted wastewater, complemented with nitrogen as NH4Cl and phosphorus as NaH2PO4, in a ratio of COD/N/P equal to 400/7/1 (Cresson et al., 2009). The influent was daily prepared by diluting wine to different final COD concentrations ranging from 0.5 to 24 g L 1. The reactor was inoculated with 1 g VS L 1 of sludge originating from a large-scale anaerobic reactor treating distillery vinasse.
Table 1 Van soest fractionation, TS, VS content and biochemical methane potential of the six LBs. Supports
LF
CC
GS
SS
WS
CF
TS (%) VS (%TS) Soluble fraction (%TS) Hemicellulose (%TS) Cellulose (%TS) Lignin (%TS) BMP (mL CH4/g VS)
90 89 19.2 24 46 10.8 434.11
82 77 17.7 16.3 37 29 68.18
79 74 32.3 15 27.5 25.2 220.93
92 83 26.8 14.2 31.3 26.7 325.22
94 87 23 23.8 31.7 20.5 347.95
89 84 16.6 23 45 15.4 281.71
152
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
2.2.3. Analytical methods During the reactor operation, COD, VFA and biogas composition were daily analyzed. Soluble COD was measured by colorimetric method using Hach 0–1500 mg L 1 vials. VFA concentrations were measured with a gas chromatograph Perkin Clarus580 with an Elite-FFAP crossbond carbowax 15 m column connected to an FID detector at 280 °C and nitrogen at 6 mL min 1 as carrier gas. The biogas production was analyzed by biogas counter and then normalized according to the ambient temperature. The Biogas composition was determined using a gas chromatograph (Varian GC-CP4900) equipped with two columns: the first (Molsieve 5A PLOT) was used at 110 °C to separate O2, N2, CH4, the second (HayeSep A) was used at 70 °C to separate CO2 from other gases. The injector temperature was 110 °C and the detector 55 °C. The detection of gaseous compounds was done using a thermal conductivity detector.
2.3. Experimental strategy The reactors were operated sequentially during 146 days in three consecutive stages in order to investigate (1) the efficiency of LBs, with different chemical compositions and biodegradabilities, to be used as carrier support during a short start-up period (30–40 days), (2) the biodegradation of LB supports during the inactivity period (3 months), (3) and finally the capacity of these reactors to be restarted again in very short period (15 days).
2.3.1. Phase 1: start-up phase During the first 40 days, the reactors were operated as continuous fixed bed reactors for the treatment of winery wastewater. Reactors were fed continuously at a low initial OLR of 0.65 gCOD L 1 d 1 by applying a short hydraulic retention time (HRT) of 18 h with diluted wastewater to wash out the planktonic biomass coming from the inoculum. The start-up strategy proposed by Cresson et al. (2008) was adopted in order to achieve a short start-up period of about 1 month. When a COD removal of 80% was achieved at the initial OLR of 0.65 gCOD L 1 d 1, the OLR was then daily increased by 15%, by increasing COD of the wastewater to maintain the same HRT, till reaching 25 gCOD L 1 d 1 in 23 days. The final OLR was then maintained for 5 days to showcase the stability of the process.
2.3.2. Phase 2: non-feeding phase The reactor feeding with influent was stopped and three reactors packed with sunflower stalk (SS), grape stalk (GS) and luffa fiber (LF) were followed to investigate the treatment performance and the biogas production during the second and the third stages. During this second stage, the reactors were operated as solid anaerobic digesters in batch mode for 3 months during which the lignocellulosic supports remained as the only substrate.
2.3.3. Phase 3: restart-up phase The three reactors were fed with wastewater and operated again as continuous fixed bed reactors (HRT of 18 h) by applying a high OLR in a short period (15 days) after a long period of inactivity. The reactors were initially fed with a higher OLR than during the first stage (2 gCOD L 1 d 1) and OLR was then increased daily by 30% till reaching 27 gCOD L 1 d 1. This OLR was maintained for 4 days to showcase the stability of the process.
3. Results and discussion Continuous assays using an anaerobic fixed bed reactor were performed to investigate the efficiency of LB to be used as carrier support through a short start-up period and the effect of the successions of start-up, including non-feeding periods on the treatment performance and biogas production of the reactor. LBs selected for the study cover a wide range of biochemical compositions and, as a consequence, biodegradabilities (Table 1 and Supplementary text). 3.1. First phase: start-up period of anaerobic fixed bed reactors Reactors loaded with the six different LBs were started up following the same strategy. Table 2 presents the main parameters followed during this period. First, it is important to notice that the targeted OLR of 25 gCOD L 1d 1 at the end of the start-up period was reached for all the LBs tested, except for CF (maximal OLR of 19 gCOD L 1 d 1). According to the start-up strategy, an OLR increase requires a COD removal efficiency higher than 80%. For the reactor filled with CF, the COD removal efficiency decreased rapidly from 96% on day 27 to values lower than 80% on day 34 (Fig. 1). The OLR increase was thus stopped in order to prevent an overloading shock and a perturbation of the biofilm activity. In fact, compared to the other LBs, CF is characterized by a lower specific surface available for biofilm growth which can minimize the quantity of microorganisms entrapped within the support matrix. This physical structure, inducing lower microorganisms abundance, could explain the lower treatment capacity of this specific LBs support. Depending on the LB, the start-up period varied from 32 days for LF to 44 days for WS. This difference is mostly attributed to the ‘‘acclimation’’ phase, corresponding to the time required to reach a COD removal efficiency > 80% when an initial OLR of 0.65 gCOD L 1 d 1 was applied (Fig. 1). The durations of acclimation phase were about 2 weeks for GS, WS and SS, whereas they were lower than one week for the three other LBs (Table 2). Variations in the duration of the acclimation phase can be related to biochemical characteristics of LBs, mainly to the content in soluble fraction (Table 1). During the first days of contact with a liquid fraction, LBs having the highest content in soluble fraction can emit additional soluble COD. For instance, the COD of the effluent was higher than the one of the influent (0.5 gCOD L 1) during the first 6 days for GS and WS and first 8 days for SS, creating an organic overload which may increase the time required to reach 80% of COD removal (calculated on the basis of the COD of the effluent inlet). As a consequence, VFA concentration (Fig. 2) was higher during the first days for reactors loaded with LBs of high content of soluble fraction (e.g., 1.8 g L 1 for SS). For the 5 supports operated at a final OLR of 25 gCOD L 1 d 1, COD removal efficiencies ranged from 82% to 94% at the end of the first phase, as they were also related to the biochemical characteristics of the LBs. Higher final COD removal efficiencies were obtained for LBs characterized by high lignin and low hemicellulose content (i.e., SS, GS and CC). In fact, as it was mentioned in many previous studies, the presence of high content of lignin
Table 2 OLR and treatment performance at the end of the start-up period for the six reactors. Supports
LF
CC
GS
SS
SW
CF
Final OLR (gCOD L 1 d 1) Acclimation phase period (d) Start-up period (d) Final COD removal (%) Final VFA (g L 1) SW
25 2 32 87 0.74
25 5 34 90 1.08
25 13 40 92 0.45
25 14 41 94 0.4
25 11 44 86 0.38
19 7 40 82 1.72
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
low cost biofilm carrier supports, alternative to plastic carriers, in order to allow a fast start-up with a high COD removal efficiency. To sum up, the use of LB as carrier support allows a short startup phase to initiate the formation of an active biofilm capable of treating liquid influent under high performances. Biochemical composition and the resulting degradability potential of LBs are the main parameters influencing the start-up phase. They must be considered to reduce the start-up period and allow higher treatment performance without perturbing the biofilm activity.
100 90
COD removal (%)
80 70 60
LF
50
CC CF
40
153
WS 30 SS 20
3.2. Phase 2: biodegradation of biofilm carrier during the inactivity period
GS
10 0 0
3
7
10
13
16
19
22 25 Time (d)
28
31
34
37
40
Fig. 1. COD removal efficiency during the start-up period.
2
LF
VFA concentration ( g L-1)
1,8
CC
1,6
CF
1,4
WS
1,2
SS 1 GS 0,8 0,6 0,4
0,2 0 1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 Time (d)
Fig. 2. VFA accumulation during the start-up period.
represents a physical barrier for hydrolytic enzymes inducing slow biodegradation level and kinetics. On the opposite, for LBs characterized by low lignin and high hemicelluloses contents, the biofilm can also use the LB as a second substrate, which increases the global COD load. Furthermore, the residual COD in the effluent is higher in the reactors packed with LBs which have high amount of hemicelluloses than in the reactors packed with LBs which have high amount of lignin. As far as there is more lignin content, the LB would behave as an inert support like plastic or mineral support. Except with CF (final OLR of 19 gCOD L 1 d 1), VFA concentration was lower than 1 g L 1 at the end of the start-up phase, which confirms the high COD removal efficiencies. For all reactors, acetic acid remained the major VFA produced, in addition to a low concentration of propionic acid. Moreover, in the stable phase during which the OLR was maintained constant for 5 days after the increase of OLR, the VFA concentration decreased while the COD removal efficiency increased. This may explain the stability of the reactors which can be attributed to adequate methanogenic activity of the biofilm. Table 3 summarizes the start-up parameters of anaerobic fixed bed reactors packed with inert supports for the treatment of wastewater. The global efficiencies (final OLR, COD removal and duration of the start-up period) of the start-up period for the reactor loaded with LBs are in adequacy with the performances observed with efficient inert supports (Habouzit et al., 2014). The present study thus demonstrates that LBs, which can be considered as solid wastes in the agro-industries, can be used as reliable and
In order to simulate extreme seasonal variations and to investigate the biodegradation of the colonized biofilm carrier support in absence of the liquid influent, the feeding of the reactors packed with SS, GS and LF has been stopped at the end of the start-up period for 3 months. During this phase, these three reactors were operated as solid anaerobic digester in batch mode. The residual soluble COD and VFA were consumed during the first week after the shutdown of the influent feeding and no further accumulation of VFA and COD was then observed. Meanwhile, the three reactors continued to produce biogas with different rates according to the biodegradability of the LBs. LF had the highest methane production with 103 mL g VS1 followed by SS with 35 mL g VS1 and GS with only 12 mL g VS1. These results are in agreement with the BMP results which indicated the same order of biodegradability (Table 1). This indicates that the most biodegradable fraction of the LB was probably converted into biogas during the start-up period of 45 days. Andersson and Björnsson (2002) and Mshandete et al. (2008) indicated that about 60% of the lignocellulosic support was degraded during a start-up period of about 80 days. As a consequence, during the batch period, the biofilm may conserve its biological activity because of the biodegradation of LB (i.e., a second substrate). Therefore, the objective of the next step was to investigate whether the biofilm was active enough to cope with the application of wastewater at high OLR in short period. 3.3. Phase 3: reactor performance during the restart-up period The reactors packed with SS, GS and LF were fed at an initial OLR of 2 gCOD L 1 d 1 without any acclimation phase. Fig. 3 presents the OLR and the COD removal rate throughout the restart-up period. According to this figure, the reactor performances were not affected by a drastic increase of the OLR. In the first 8 days (OLR up to 12.6 gCOD L 1 d 1), the COD removal efficiencies were almost constant (93%, 88% and 80% for SS, GS and LF, respectively) and the VFA concentrations were lower than 1 g L 1 in the three reactors (Fig. 4). These high performances demonstrate that the biofilm conserved its activity even after a non-feeding period of 3 months. Nevertheless, when increasing the OLR from 12.6 to 27 gCOD L 1 d 1, the COD removal efficiency decreased to 71% and 77% for LF and GS, respectively, while it remained high for SS with 87%. The sharp decrease of COD removal efficiency in the reactors packed with GS and LF was associated with the increase of VFA concentration which reached 1.4 g L 1 and 2.5 g L 1, respectively, at the end of the restart-up period. However, methane production was not affected by the increase of VFA concentration. Acetic acid remained the major VFA produced in addition to a low concentration of propionic acid. At the end of the restart-up period, OLR was kept constant at 27 gCOD L 1 d 1 for 4 days and the COD removal efficiency increased to reach 90%, 82% and 75% for SS, GS and LF, respectively. Such COD removal efficiencies were lower than the ones observed at the end of the first start-up period. In addition, the VFA concentration was
154
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
Table 3 Start-up performance of biofilm reactors packed inert supports. Supports
OLR (gCOD L
Polyethylene Polyvinyl Chloride Polypropylene Clay ceramic particles Active carbon fiber Polyethylene String-shaped plastic
2.6 21.3 30 8 6.7 42 37.7
1
d
1
)
Start-up period (d)
Final COD removal (%)
References
35 34 31 45 40 140 70
83 96 97 76 80 80 82
Habouzit et al. (2014) Habouzit et al. (2014) Habouzit et al. (2014) Han et al. (2013) Zhao et al. (2013) Ganesh et al. (2010) Yu et al. (2006)
100
50 45
90
35 OLR
70
30
LF 25
GS
60
20
SS
OLR (g L-1d-1)
COD removal (%)
40 80
15
50
10 40 5 30
0 1
2
3
4
5
6
7
8 9 Time (d)
10
11
12
13
14
15
Fig. 3. COD removal efficiency and OLR evolution during the restart-up period.
3 LF
VFA concentration ( g L-1)
2,5
GS 2
SS
1,5
1
0,5
0 1
2
3
4
5
6
7 8 Time (d)
9
10
11
12
13
14
Fig. 4. VFA accumulation during the restart-up period.
reduced (Fig. 4), indicating a robustness of the reactor which could be attributed to the stable methanogenic activity of the biofilm which resisted to the extreme OLR increase in a very short period, even after a long period without feeding the reactor with influent. 3.4. Prospect of using the hybrid anaerobic reactor as a sustainable treatment technology in agro-industries The proposed hybrid reactor can ensure the treatment of both liquid and solid waste in one reactor is capable of resisting seasonal variations. First, this study demonstrated that LBs can be used as efficient biofilm carrier supports for treating liquid influent under high performances. However, biochemical composition and resulting degradability potential of the LBs play a crucial role in the start-up phase of anaerobic fixed bed reactor. LBs characterized by high level of soluble easily degradable biodegradable fraction are associated to long acclimation phase of about 2 weeks, because
the soluble fraction released in the reactor represents a supplementary load which has to be degraded in addition to the influent COD. The reactor efficiency during the start-up period is also affected by the presence of hemicelluloses and cellulose. Actually, during this period of several weeks, hydrolysis of these complex carbohydrate compounds produces additional metabolites which have to be degraded by the methanogenic consortium as another supplementary load. As a consequence, LBs having higher content of hemicelluloses and cellulose (i.e., SS, GS and CC) can induce organic overloads which have to be considered to define the initial effluent organic loading rate. In addition, after a non-feeding period of 3 months, we have also demonstrated that the hybrid reactors were able to restartup successfully in only 15 days. In fact, biofilms conserve a good level of biological activity by using the lignocellulosic biomasses as a second substrate. As a consequence, the use of lignocellulosic waste (LB) as low cost support (alternative to plastic material) gives the potential to ensure three different actions in relation with the seasonal variations of the industry activity. Firstly, during the harvesting and the processing of the raw material, characterized by the generation of large quantity of wastewater with high flow rate and high OLR for relatively short period, such reactor can be started-up rapidly and ensure the continuous treatment of these effluents with high COD removal efficiency. Secondly, when the agro-industry activity slows down and the effluent flow rate is reduced, the biofilm can maintain a high biological activity by the co-digestion of both the solid and liquid wastes and, as a consequence, will be less affected by the reduction of substrate in the effluent due to the presence of solid waste as a second substrate. This flexible mode of operation can reduce the perturbation of biofilm activity in case of seasonal fluctuations and successions of feeding and shut-down operations. Finally, at the end of the production process, characterized by the high accumulation of the solid residue and no effluent generation, the hybrid reactor can be operated as a dry solid waste digester with highly active and acclimated biofilm able to use the solid waste as only substrate. Such a hybrid reactor could contribute to the sustainable development of agro-industries by reducing their environmental impact and their waste treatment costs while producing their own clean renewable free energy through the year.
4. Conclusion The use of lignocellulosic biomass as both carrier support and substrate allowed the treatment of both liquid and solid wastes in the same reactor with less operational cost. Such hybrid reactor ensured high treatment efficiency in addition to fast start-up phase. The efficiency of the reactors over the start-up period depends on the chemical composition of LBs. The reactor was successfully operated sequentially to response to seasonal variation by the fast restart-up of the reactor. Biofilm activity seemed to be more stable during start-up and restart up, including periods of inactivity, when using LB as carrier support.
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
Acknowledgements This work was carried out at the ‘‘Laboratory of Environmental Biotechnology - Narbonne, France’’ and supported by Grants from the ‘‘Ministry of Higher Education and Scientific Research of Tunisia’’ and from ‘‘the Agence Universitaire de la Francophonie (AUF)’’. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.08. 127. References Andersson, J., Björnsson, L., 2002. Evaluation of straw as a biofilm carrier in the methanogenic stage of two-stage anaerobic digestion of crop residues. Bioresour. Technol. 85, 51–56. Arnaiz, C., Buffiere, P., Elmaleh, S., Lebrato, J., Moletta, R., 2003. Anaerobic digestion of dairy wastewater by inverse fluidization: the inverse fluidized bed and the inverse turbulent bed reactors. Environ. Technol. 24, 1431–1443. Cresson, R., Dabert, P., Bernet, N., 2009. Microbiology and performance of a methanogenic biofilm reactor during the start-up period. J. Appl. Microbiol. 106, 863–876. Cresson, R., Escudié, R., Steyer, J.P., Delgenès, J.P., Bernet, N., 2008. Competition between planktonic and fixed microorganisms during the start-up of methanogenic biofilm reactors. Water Res. 42, 792–800. Escudié, R., Cresson, R., Delgenes, J.-P., Bernet, N., 2011. Control of start-up and operation of anaerobic biofilm reactors: an overview of 15 years of research. Water Res. 45 (1), 1–10. Ganesh, R., Rajagopal, R., Torrijos, M., Thanikal, J.V., Ramanujam, R., 2010. Anaerobic treatment of winery wastewater in fixed bed reactors. Bioprocess Biosyst. Eng. 33, 619–628. Habouzit, F., Gaëlle, S.-C., Jerôme, H., Steyer, J.P., Bernet, N., 2014. Biofilm development during the start-up period of anaerobic biofilm reactors: the biofilm Archaea community is highly dependent on the support material. Microb. Biotechnol. 7 (3), 257–264.
155
Han, W., Yue, Q., Wu, S., Zhao, Y., Gao, B., Li, Q., Wang, Y., 2013. Application and advantages of novel clay ceramic particles (CCPs) in an up-flow anaerobic biofilter (UAF) for wastewater treatment. Bioresour. Technol. 137, 171–178. Karthikeyan, O., Visvanathan, C., 2012. Bio-energy recovery from high-solid organic substrates by dry anaerobic bio-conversion processes: a review. Rev. Environ. Sci. Biotechnol., 1–28. Motte, J.-C., Escudié, R., Beaufils, N., Steyer, J.-P., Bernet, N., Delgenès, J.-P., Dumas, C., 2014. Morphological structures of wheat straw strongly impacts its anaerobic digestion. Ind. Crops Prod. 52, 695–701. Motte, J.-C., Escudié, R., Bernet, N., Delgenes, J.-P., Steyer, J.-P., Dumas, C., 2013. Dynamic effect of total solid content, low substrate/inoculum ratio and particle size on solid-state anaerobic digestion. Bioresour. Technol., 141–148. Mshandete, A.M., Björnsson, L., Kivaisi, A.K., Rubindamayugi, M.S.T., Mattiasson, B., 2008. Performance of biofilm carriers in anaerobic digestion of sisal leaf waste leachate. Electron. J. Biotechnol. 11 (1). Nabizadeh, R., Naddafi, K., Mesdaghinia, A., Nafez, H.A., 2008. Feasibility study of organic matter and ammonium removal using loofa sponge as a supporting medium in an aerated submerged fixed-film reactor (ASFFR). Electron. J. Biotechnol. 11 (4). Ren, N., Wang, A., 2004. Principle and Application of Anaerobic Biotechnology. Chemical Industry Press, 89–138. Sasaki, K., Morita, M., Hirano, S.-I., Ohmura, N., Igarashi, Y., 2009. Effect of adding carbon fiber textiles to methanogenic bioreactors used to treat an artificial garbage slurry. J. Biosci. Bioeng. 108 (2), 130–135. Sowmeyan, R., Swaminathan, G., 2008. Performance of inverse anaerobic fluidized bed reactor for treating high strength organic wastewater during start-up phase. Bioresour. Technol. 99 (14), 6280–6284. Tarjányi-Szikora, S., Oláh, J., Makó, M., Palkó, G., Barkács, K., Záray, G., 2013. Comparison of different granular solids as biofilm carriers. Microchem. J. 107, 101–107. Yu, H.Q., Zhao, Q.B., Tang, Y., 2006. Anaerobic treatment of winery wastewater using laboratory-scale multi- and single-fed filters at ambient temperatures. Process Biochem. 41, 2477–2481. Zhao, H., Li, J., Li, J., Yuan, X., Piao, R., Zhu, W., Li, H., Wang, X., Cui, Z., 2013. Organic loading rate shock impact on operation and microbial communities in different anaerobic fixed-bed reactors. Bioresour. Technol. 140, 211–219. Zhao, Q.B., Mu, Y., Wang, Y., Liu, X.W., Dong, F., Yu, H.Q., 2008. Response of a biohydrogen-producing reactor to the substrate shift from sucrose to lactose. Bioresour. Technol. 99, 8344–8347.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Sequential operation of a hybrid anaerobic reactor using a lignocellulosic biomass as biofilm support Mohamed Ali Wahab a,b, Frédéric Habouzit b, Nicolas Bernet b, Jean-Philippe Steyer b, Naceur Jedidi a, Renaud Escudié b,⇑ a b
University of Carthage, Water Research and Technologies Centre (CERTE), Wastewater Treatment Laboratory, Tunisia INRA, UR50, Laboratoire de Biotechnologie de l’Environnement, Avenue des Etangs, Narbonne F-11100, France
h i g h l i g h t s Hybrid anaerobic reactor operated sequentially to treat liquid and solid waste. Lignocellulosic biomass used as both carrier support and substrate. High treatment efficiency during fast start-up and restart-up phases. Acclimation phase depends of the soluble fraction of lignocellulosic support. During non-feeding period, biofilm conserved its biological activity.
a r t i c l e
i n f o
Article history: Received 17 June 2014 Received in revised form 28 August 2014 Accepted 31 August 2014 Available online 8 September 2014 Keywords: Anaerobic digestion Lignocellulosic biomass Biofilm Start-up Hybrid reactor
a b s t r a c t Agro-industries are facing many economic and environmental problems associated with seasonal generation of liquid and solid waste. In order to reduce treatment costs and to cope with seasonal variation, we have developed a hybrid anaerobic reactor operated sequentially by using lignocellulosic biomass (LB) as biofilm carrier support. Six LBs were tested to evaluate the treatment performance during a succession of two start-up periods, separated by a non-feeding period. After a short acclimation phase of several days, all the reactors succeeded in starting-up in less than 1 month to reach an organic loading rate of 25 gCOD L 1 d 1. In addition, they restarted-up successfully in only 15 days after a 3 month non-feeding period, indicating that biofilms conserved their biological activities during this last phase. As a consequence, the use of LB as a biofilm support gives the potential to sustain seasonal variations of wastewater loads for industrial application. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The agro-industries are considered as water and energy intensive consumers and their activities have been always associated with economic and environmental problems due to the generation of large amounts of wastewater and solid waste, in addition to the increase of their carbon footprint. Thus, agro-industries should focus on improving their energy efficiency, water use and waste management. Hence, in a world where water scarcity and climate change are a reality, actions to protect the environment and enhance renewable energy are mandatory for any type and size of industries.
⇑ Corresponding author. Tel.: +33 468425173. E-mail address: [email protected] (R. Escudié). http://dx.doi.org/10.1016/j.biortech.2014.08.127 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.
The high biodegradability of agro-industries waste often justifies the use of conventional biological treatments. Among them, anaerobic digestion (AD) is the most adequate technology for the treatment of wastewaters with high organic carbon concentration. AD ensures the degradation of organic matter simultaneously into a valuable biogas composed of methane (CH4) and carbon dioxide (CO2) and into a nutrient-rich digestate with agronomic qualities (Karthikeyan and Visvanathan, 2012). A major limitation of AD process is the long duration and instability of their transition phases (starting, re-starting, organic load increase) because of the low growth rates of anaerobic microorganisms and their sensitivity to perturbations such as organic overloads (Arnaiz et al., 2003; Cresson et al., 2009; Escudié et al., 2011). Moreover, application of AD for the treatment of agro-industrial waste and wastewater is hindered by several limitations, such as the heterogeneity and the seasonal supply of waste, low biogas production rate and high
151
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
investment cost (Ren and Wang, 2004; Sasaki et al., 2009). Thus, it is of high interest to develop an AD reactor with low investment and operational costs that could ensure the treatment of both liquid and solid waste in one single reactor with high treatment efficiency and high methane yield. The operation of this reactor has to tolerate seasonal fluctuations and OLR perturbations that can happen in agro-industries over the year. Anaerobic fixed bed reactors (AFBR) present a good potential for the treatment of wastewaters. In fact, they are known to obtain high solid retention times for high system efficiency and stability, with low hydraulic retention times for system economy. Furthermore, these processes are inherently stable and resistant to organic and hydraulic shock loadings (Zhao et al., 2013). AFBR performance is strongly depending on the nature of the support material (Habouzit et al., 2014). Different natural and artificial solid materials can be used as biofilm carriers. Commercial carrier elements which are usually made of polyethylene, polypropylene or polyurethane, are expensive, while natural carriers are essentially inorganic such as sand, gravel, pumice stone, porous glass beads and zeolite (Tarjányi-Szikora et al., 2013). Natural lignocellulosic biomasses (LBs) can be an alternative to these inorganic or plastic support materials. LBs are abundant in agro-industries and represent the main part of the solid waste produced. They have low costs, while having high porosity, low specific gravity and higher bacterial adsorption/adhesion (Nabizadeh et al., 2008). The LBs are essentially composed of cellulose, hemicelluloses and lignin in different proportions. The variation of their chemical composition can influence the treatment performance and methane production of the reactor, especially during the start-up period which constitutes an important step in the anaerobic biological treatment process (Cresson et al., 2009; Habouzit et al., 2014). However, the efficiency of LB to be used as biofilm carrier has not been well studied. Andersson and Björnsson (2002) and Mshandete et al. (2008) have investigated the use of wheat straw and sisal fiber waste, respectively, as biofilm supports in methanogenic bioreactors digesting crop waste leachate, in a two-stage anaerobic digestion. To our knowledge, there is no study which has investigated the efficiency of LB for the treatment of high strength wastewater in one-stage reactor. Moreover, because of the seasonal production of effluents in the agro-industries (Zhao et al., 2008), the reactor operation is seasonally unstable. In fact, when the process is not fed (i.e., ‘‘inactivity’’ period), the biological activity of the biofilm is affected and the reactor becomes difficult to be restarted again: this sequential operation can lead to the low removal of organic matters or to a prolonged period of acclimation (Sowmeyan and Swaminathan, 2008). The use of LB as a biofilm support could be a very interesting solution to keep the biofilm biologically active during the inactivity period. The main objective of this study was to explore the potential and performance of using LB as biofilm support and for the seasonal treatment of high strength wastewater. Since the start-up phase is a critical step in the process operation, we have first investigated the efficiency of using six LBs (having different chemical compositions and biodegradabilities) as biofilm support during this decisive period. These reactors were restarted after an inactivity period of 3 month in order to simulate a sequential operation of the process.
2. Methods 2.1. Characterization of lignocellulosic biomasses Six LBs were used to investigate the efficiency of using natural materials as biofilm carrier supports. LBs used were: wheat straw
(WS), sunflower stalk (SS), grape stalk (GS), cactus fiber (CF), luffa fiber (LF) and cypress cones (CC). These LBs were used without grinding or any other pretreatment. The following methods were applied in order to identify their biochemical characteristics and their biodegradability. The Total Solid (TS) and Volatile Solid (VS) contents were measured as described previously by Motte et al. (2013). The distribution of lignocellulosic compounds (soluble compounds, cellulose, hemicelluloses and lignin) was estimated according to a Van Soest fractionation adapted by Motte et al. (2014). The LBs were firstly grinded and sieved to achieve a particle size of 1 mm. Biomethane potential (BMP) of each LB was evaluated by following the methane produced under mesophilic conditions (35 °C). The inoculum was an industrial sludge sampled from a UASB (Upflow Anaerobic Sludge Blanket) process treating sugar factory effluent. The assays were conducted in triplicate in 600 mL sealed flask (working volume of 400 mL) according to the procedure described by Motte et al. (2013). The methane production is expressed under standard condition and accounts for the variation of the gas content in the headspace of the reactors. The biochemical composition and the biomethane potentials of the six LBs are reported in Table 1.
2.2. Laboratory-scale reactors 2.2.1. Experimental set-up The anaerobic fixed bed reactor employed in this study is a PVC column with an internal diameter of 0.20 m and a height of 0.56 m. The total and working volume of the reactor were 16.7 L and 14.8 L, respectively. The reactor was equipped with hot water jackets to maintain a mesophilic temperature of 35 °C. Six reactors were packed with the different LBs used as carrier supports for biomass immobilization and retention. The influent was pumped to the top of the reactor by means of a peristaltic pump. The reactor liquid was recirculated within the reactor by means of a recirculation pump fixed at the bottom inside the reactor. The effluent was discharged at the top through a U-tube for separation of gas. pH was measured in the liquid outlet with a Mettler Toledo 1100 Calimatic pH meter, regulated at 6.8 by automatic NaOH addition. Biogas produced passed through a moisture trap and then to a milligas counter fitted with a 4–20 mA output (Ritter, Bochum, Germany).
2.2.2. Influent preparation and reactor inoculation The influent was a wine-based reconstituted wastewater, complemented with nitrogen as NH4Cl and phosphorus as NaH2PO4, in a ratio of COD/N/P equal to 400/7/1 (Cresson et al., 2009). The influent was daily prepared by diluting wine to different final COD concentrations ranging from 0.5 to 24 g L 1. The reactor was inoculated with 1 g VS L 1 of sludge originating from a large-scale anaerobic reactor treating distillery vinasse.
Table 1 Van soest fractionation, TS, VS content and biochemical methane potential of the six LBs. Supports
LF
CC
GS
SS
WS
CF
TS (%) VS (%TS) Soluble fraction (%TS) Hemicellulose (%TS) Cellulose (%TS) Lignin (%TS) BMP (mL CH4/g VS)
90 89 19.2 24 46 10.8 434.11
82 77 17.7 16.3 37 29 68.18
79 74 32.3 15 27.5 25.2 220.93
92 83 26.8 14.2 31.3 26.7 325.22
94 87 23 23.8 31.7 20.5 347.95
89 84 16.6 23 45 15.4 281.71
152
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
2.2.3. Analytical methods During the reactor operation, COD, VFA and biogas composition were daily analyzed. Soluble COD was measured by colorimetric method using Hach 0–1500 mg L 1 vials. VFA concentrations were measured with a gas chromatograph Perkin Clarus580 with an Elite-FFAP crossbond carbowax 15 m column connected to an FID detector at 280 °C and nitrogen at 6 mL min 1 as carrier gas. The biogas production was analyzed by biogas counter and then normalized according to the ambient temperature. The Biogas composition was determined using a gas chromatograph (Varian GC-CP4900) equipped with two columns: the first (Molsieve 5A PLOT) was used at 110 °C to separate O2, N2, CH4, the second (HayeSep A) was used at 70 °C to separate CO2 from other gases. The injector temperature was 110 °C and the detector 55 °C. The detection of gaseous compounds was done using a thermal conductivity detector.
2.3. Experimental strategy The reactors were operated sequentially during 146 days in three consecutive stages in order to investigate (1) the efficiency of LBs, with different chemical compositions and biodegradabilities, to be used as carrier support during a short start-up period (30–40 days), (2) the biodegradation of LB supports during the inactivity period (3 months), (3) and finally the capacity of these reactors to be restarted again in very short period (15 days).
2.3.1. Phase 1: start-up phase During the first 40 days, the reactors were operated as continuous fixed bed reactors for the treatment of winery wastewater. Reactors were fed continuously at a low initial OLR of 0.65 gCOD L 1 d 1 by applying a short hydraulic retention time (HRT) of 18 h with diluted wastewater to wash out the planktonic biomass coming from the inoculum. The start-up strategy proposed by Cresson et al. (2008) was adopted in order to achieve a short start-up period of about 1 month. When a COD removal of 80% was achieved at the initial OLR of 0.65 gCOD L 1 d 1, the OLR was then daily increased by 15%, by increasing COD of the wastewater to maintain the same HRT, till reaching 25 gCOD L 1 d 1 in 23 days. The final OLR was then maintained for 5 days to showcase the stability of the process.
2.3.2. Phase 2: non-feeding phase The reactor feeding with influent was stopped and three reactors packed with sunflower stalk (SS), grape stalk (GS) and luffa fiber (LF) were followed to investigate the treatment performance and the biogas production during the second and the third stages. During this second stage, the reactors were operated as solid anaerobic digesters in batch mode for 3 months during which the lignocellulosic supports remained as the only substrate.
2.3.3. Phase 3: restart-up phase The three reactors were fed with wastewater and operated again as continuous fixed bed reactors (HRT of 18 h) by applying a high OLR in a short period (15 days) after a long period of inactivity. The reactors were initially fed with a higher OLR than during the first stage (2 gCOD L 1 d 1) and OLR was then increased daily by 30% till reaching 27 gCOD L 1 d 1. This OLR was maintained for 4 days to showcase the stability of the process.
3. Results and discussion Continuous assays using an anaerobic fixed bed reactor were performed to investigate the efficiency of LB to be used as carrier support through a short start-up period and the effect of the successions of start-up, including non-feeding periods on the treatment performance and biogas production of the reactor. LBs selected for the study cover a wide range of biochemical compositions and, as a consequence, biodegradabilities (Table 1 and Supplementary text). 3.1. First phase: start-up period of anaerobic fixed bed reactors Reactors loaded with the six different LBs were started up following the same strategy. Table 2 presents the main parameters followed during this period. First, it is important to notice that the targeted OLR of 25 gCOD L 1d 1 at the end of the start-up period was reached for all the LBs tested, except for CF (maximal OLR of 19 gCOD L 1 d 1). According to the start-up strategy, an OLR increase requires a COD removal efficiency higher than 80%. For the reactor filled with CF, the COD removal efficiency decreased rapidly from 96% on day 27 to values lower than 80% on day 34 (Fig. 1). The OLR increase was thus stopped in order to prevent an overloading shock and a perturbation of the biofilm activity. In fact, compared to the other LBs, CF is characterized by a lower specific surface available for biofilm growth which can minimize the quantity of microorganisms entrapped within the support matrix. This physical structure, inducing lower microorganisms abundance, could explain the lower treatment capacity of this specific LBs support. Depending on the LB, the start-up period varied from 32 days for LF to 44 days for WS. This difference is mostly attributed to the ‘‘acclimation’’ phase, corresponding to the time required to reach a COD removal efficiency > 80% when an initial OLR of 0.65 gCOD L 1 d 1 was applied (Fig. 1). The durations of acclimation phase were about 2 weeks for GS, WS and SS, whereas they were lower than one week for the three other LBs (Table 2). Variations in the duration of the acclimation phase can be related to biochemical characteristics of LBs, mainly to the content in soluble fraction (Table 1). During the first days of contact with a liquid fraction, LBs having the highest content in soluble fraction can emit additional soluble COD. For instance, the COD of the effluent was higher than the one of the influent (0.5 gCOD L 1) during the first 6 days for GS and WS and first 8 days for SS, creating an organic overload which may increase the time required to reach 80% of COD removal (calculated on the basis of the COD of the effluent inlet). As a consequence, VFA concentration (Fig. 2) was higher during the first days for reactors loaded with LBs of high content of soluble fraction (e.g., 1.8 g L 1 for SS). For the 5 supports operated at a final OLR of 25 gCOD L 1 d 1, COD removal efficiencies ranged from 82% to 94% at the end of the first phase, as they were also related to the biochemical characteristics of the LBs. Higher final COD removal efficiencies were obtained for LBs characterized by high lignin and low hemicellulose content (i.e., SS, GS and CC). In fact, as it was mentioned in many previous studies, the presence of high content of lignin
Table 2 OLR and treatment performance at the end of the start-up period for the six reactors. Supports
LF
CC
GS
SS
SW
CF
Final OLR (gCOD L 1 d 1) Acclimation phase period (d) Start-up period (d) Final COD removal (%) Final VFA (g L 1) SW
25 2 32 87 0.74
25 5 34 90 1.08
25 13 40 92 0.45
25 14 41 94 0.4
25 11 44 86 0.38
19 7 40 82 1.72
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
low cost biofilm carrier supports, alternative to plastic carriers, in order to allow a fast start-up with a high COD removal efficiency. To sum up, the use of LB as carrier support allows a short startup phase to initiate the formation of an active biofilm capable of treating liquid influent under high performances. Biochemical composition and the resulting degradability potential of LBs are the main parameters influencing the start-up phase. They must be considered to reduce the start-up period and allow higher treatment performance without perturbing the biofilm activity.
100 90
COD removal (%)
80 70 60
LF
50
CC CF
40
153
WS 30 SS 20
3.2. Phase 2: biodegradation of biofilm carrier during the inactivity period
GS
10 0 0
3
7
10
13
16
19
22 25 Time (d)
28
31
34
37
40
Fig. 1. COD removal efficiency during the start-up period.
2
LF
VFA concentration ( g L-1)
1,8
CC
1,6
CF
1,4
WS
1,2
SS 1 GS 0,8 0,6 0,4
0,2 0 1
3
5
7
9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 Time (d)
Fig. 2. VFA accumulation during the start-up period.
represents a physical barrier for hydrolytic enzymes inducing slow biodegradation level and kinetics. On the opposite, for LBs characterized by low lignin and high hemicelluloses contents, the biofilm can also use the LB as a second substrate, which increases the global COD load. Furthermore, the residual COD in the effluent is higher in the reactors packed with LBs which have high amount of hemicelluloses than in the reactors packed with LBs which have high amount of lignin. As far as there is more lignin content, the LB would behave as an inert support like plastic or mineral support. Except with CF (final OLR of 19 gCOD L 1 d 1), VFA concentration was lower than 1 g L 1 at the end of the start-up phase, which confirms the high COD removal efficiencies. For all reactors, acetic acid remained the major VFA produced, in addition to a low concentration of propionic acid. Moreover, in the stable phase during which the OLR was maintained constant for 5 days after the increase of OLR, the VFA concentration decreased while the COD removal efficiency increased. This may explain the stability of the reactors which can be attributed to adequate methanogenic activity of the biofilm. Table 3 summarizes the start-up parameters of anaerobic fixed bed reactors packed with inert supports for the treatment of wastewater. The global efficiencies (final OLR, COD removal and duration of the start-up period) of the start-up period for the reactor loaded with LBs are in adequacy with the performances observed with efficient inert supports (Habouzit et al., 2014). The present study thus demonstrates that LBs, which can be considered as solid wastes in the agro-industries, can be used as reliable and
In order to simulate extreme seasonal variations and to investigate the biodegradation of the colonized biofilm carrier support in absence of the liquid influent, the feeding of the reactors packed with SS, GS and LF has been stopped at the end of the start-up period for 3 months. During this phase, these three reactors were operated as solid anaerobic digester in batch mode. The residual soluble COD and VFA were consumed during the first week after the shutdown of the influent feeding and no further accumulation of VFA and COD was then observed. Meanwhile, the three reactors continued to produce biogas with different rates according to the biodegradability of the LBs. LF had the highest methane production with 103 mL g VS1 followed by SS with 35 mL g VS1 and GS with only 12 mL g VS1. These results are in agreement with the BMP results which indicated the same order of biodegradability (Table 1). This indicates that the most biodegradable fraction of the LB was probably converted into biogas during the start-up period of 45 days. Andersson and Björnsson (2002) and Mshandete et al. (2008) indicated that about 60% of the lignocellulosic support was degraded during a start-up period of about 80 days. As a consequence, during the batch period, the biofilm may conserve its biological activity because of the biodegradation of LB (i.e., a second substrate). Therefore, the objective of the next step was to investigate whether the biofilm was active enough to cope with the application of wastewater at high OLR in short period. 3.3. Phase 3: reactor performance during the restart-up period The reactors packed with SS, GS and LF were fed at an initial OLR of 2 gCOD L 1 d 1 without any acclimation phase. Fig. 3 presents the OLR and the COD removal rate throughout the restart-up period. According to this figure, the reactor performances were not affected by a drastic increase of the OLR. In the first 8 days (OLR up to 12.6 gCOD L 1 d 1), the COD removal efficiencies were almost constant (93%, 88% and 80% for SS, GS and LF, respectively) and the VFA concentrations were lower than 1 g L 1 in the three reactors (Fig. 4). These high performances demonstrate that the biofilm conserved its activity even after a non-feeding period of 3 months. Nevertheless, when increasing the OLR from 12.6 to 27 gCOD L 1 d 1, the COD removal efficiency decreased to 71% and 77% for LF and GS, respectively, while it remained high for SS with 87%. The sharp decrease of COD removal efficiency in the reactors packed with GS and LF was associated with the increase of VFA concentration which reached 1.4 g L 1 and 2.5 g L 1, respectively, at the end of the restart-up period. However, methane production was not affected by the increase of VFA concentration. Acetic acid remained the major VFA produced in addition to a low concentration of propionic acid. At the end of the restart-up period, OLR was kept constant at 27 gCOD L 1 d 1 for 4 days and the COD removal efficiency increased to reach 90%, 82% and 75% for SS, GS and LF, respectively. Such COD removal efficiencies were lower than the ones observed at the end of the first start-up period. In addition, the VFA concentration was
154
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
Table 3 Start-up performance of biofilm reactors packed inert supports. Supports
OLR (gCOD L
Polyethylene Polyvinyl Chloride Polypropylene Clay ceramic particles Active carbon fiber Polyethylene String-shaped plastic
2.6 21.3 30 8 6.7 42 37.7
1
d
1
)
Start-up period (d)
Final COD removal (%)
References
35 34 31 45 40 140 70
83 96 97 76 80 80 82
Habouzit et al. (2014) Habouzit et al. (2014) Habouzit et al. (2014) Han et al. (2013) Zhao et al. (2013) Ganesh et al. (2010) Yu et al. (2006)
100
50 45
90
35 OLR
70
30
LF 25
GS
60
20
SS
OLR (g L-1d-1)
COD removal (%)
40 80
15
50
10 40 5 30
0 1
2
3
4
5
6
7
8 9 Time (d)
10
11
12
13
14
15
Fig. 3. COD removal efficiency and OLR evolution during the restart-up period.
3 LF
VFA concentration ( g L-1)
2,5
GS 2
SS
1,5
1
0,5
0 1
2
3
4
5
6
7 8 Time (d)
9
10
11
12
13
14
Fig. 4. VFA accumulation during the restart-up period.
reduced (Fig. 4), indicating a robustness of the reactor which could be attributed to the stable methanogenic activity of the biofilm which resisted to the extreme OLR increase in a very short period, even after a long period without feeding the reactor with influent. 3.4. Prospect of using the hybrid anaerobic reactor as a sustainable treatment technology in agro-industries The proposed hybrid reactor can ensure the treatment of both liquid and solid waste in one reactor is capable of resisting seasonal variations. First, this study demonstrated that LBs can be used as efficient biofilm carrier supports for treating liquid influent under high performances. However, biochemical composition and resulting degradability potential of the LBs play a crucial role in the start-up phase of anaerobic fixed bed reactor. LBs characterized by high level of soluble easily degradable biodegradable fraction are associated to long acclimation phase of about 2 weeks, because
the soluble fraction released in the reactor represents a supplementary load which has to be degraded in addition to the influent COD. The reactor efficiency during the start-up period is also affected by the presence of hemicelluloses and cellulose. Actually, during this period of several weeks, hydrolysis of these complex carbohydrate compounds produces additional metabolites which have to be degraded by the methanogenic consortium as another supplementary load. As a consequence, LBs having higher content of hemicelluloses and cellulose (i.e., SS, GS and CC) can induce organic overloads which have to be considered to define the initial effluent organic loading rate. In addition, after a non-feeding period of 3 months, we have also demonstrated that the hybrid reactors were able to restartup successfully in only 15 days. In fact, biofilms conserve a good level of biological activity by using the lignocellulosic biomasses as a second substrate. As a consequence, the use of lignocellulosic waste (LB) as low cost support (alternative to plastic material) gives the potential to ensure three different actions in relation with the seasonal variations of the industry activity. Firstly, during the harvesting and the processing of the raw material, characterized by the generation of large quantity of wastewater with high flow rate and high OLR for relatively short period, such reactor can be started-up rapidly and ensure the continuous treatment of these effluents with high COD removal efficiency. Secondly, when the agro-industry activity slows down and the effluent flow rate is reduced, the biofilm can maintain a high biological activity by the co-digestion of both the solid and liquid wastes and, as a consequence, will be less affected by the reduction of substrate in the effluent due to the presence of solid waste as a second substrate. This flexible mode of operation can reduce the perturbation of biofilm activity in case of seasonal fluctuations and successions of feeding and shut-down operations. Finally, at the end of the production process, characterized by the high accumulation of the solid residue and no effluent generation, the hybrid reactor can be operated as a dry solid waste digester with highly active and acclimated biofilm able to use the solid waste as only substrate. Such a hybrid reactor could contribute to the sustainable development of agro-industries by reducing their environmental impact and their waste treatment costs while producing their own clean renewable free energy through the year.
4. Conclusion The use of lignocellulosic biomass as both carrier support and substrate allowed the treatment of both liquid and solid wastes in the same reactor with less operational cost. Such hybrid reactor ensured high treatment efficiency in addition to fast start-up phase. The efficiency of the reactors over the start-up period depends on the chemical composition of LBs. The reactor was successfully operated sequentially to response to seasonal variation by the fast restart-up of the reactor. Biofilm activity seemed to be more stable during start-up and restart up, including periods of inactivity, when using LB as carrier support.
M.A. Wahab et al. / Bioresource Technology 172 (2014) 150–155
Acknowledgements This work was carried out at the ‘‘Laboratory of Environmental Biotechnology - Narbonne, France’’ and supported by Grants from the ‘‘Ministry of Higher Education and Scientific Research of Tunisia’’ and from ‘‘the Agence Universitaire de la Francophonie (AUF)’’. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.biortech.2014.08. 127. References Andersson, J., Björnsson, L., 2002. Evaluation of straw as a biofilm carrier in the methanogenic stage of two-stage anaerobic digestion of crop residues. Bioresour. Technol. 85, 51–56. Arnaiz, C., Buffiere, P., Elmaleh, S., Lebrato, J., Moletta, R., 2003. Anaerobic digestion of dairy wastewater by inverse fluidization: the inverse fluidized bed and the inverse turbulent bed reactors. Environ. Technol. 24, 1431–1443. Cresson, R., Dabert, P., Bernet, N., 2009. Microbiology and performance of a methanogenic biofilm reactor during the start-up period. J. Appl. Microbiol. 106, 863–876. Cresson, R., Escudié, R., Steyer, J.P., Delgenès, J.P., Bernet, N., 2008. Competition between planktonic and fixed microorganisms during the start-up of methanogenic biofilm reactors. Water Res. 42, 792–800. Escudié, R., Cresson, R., Delgenes, J.-P., Bernet, N., 2011. Control of start-up and operation of anaerobic biofilm reactors: an overview of 15 years of research. Water Res. 45 (1), 1–10. Ganesh, R., Rajagopal, R., Torrijos, M., Thanikal, J.V., Ramanujam, R., 2010. Anaerobic treatment of winery wastewater in fixed bed reactors. Bioprocess Biosyst. Eng. 33, 619–628. Habouzit, F., Gaëlle, S.-C., Jerôme, H., Steyer, J.P., Bernet, N., 2014. Biofilm development during the start-up period of anaerobic biofilm reactors: the biofilm Archaea community is highly dependent on the support material. Microb. Biotechnol. 7 (3), 257–264.
155
Han, W., Yue, Q., Wu, S., Zhao, Y., Gao, B., Li, Q., Wang, Y., 2013. Application and advantages of novel clay ceramic particles (CCPs) in an up-flow anaerobic biofilter (UAF) for wastewater treatment. Bioresour. Technol. 137, 171–178. Karthikeyan, O., Visvanathan, C., 2012. Bio-energy recovery from high-solid organic substrates by dry anaerobic bio-conversion processes: a review. Rev. Environ. Sci. Biotechnol., 1–28. Motte, J.-C., Escudié, R., Beaufils, N., Steyer, J.-P., Bernet, N., Delgenès, J.-P., Dumas, C., 2014. Morphological structures of wheat straw strongly impacts its anaerobic digestion. Ind. Crops Prod. 52, 695–701. Motte, J.-C., Escudié, R., Bernet, N., Delgenes, J.-P., Steyer, J.-P., Dumas, C., 2013. Dynamic effect of total solid content, low substrate/inoculum ratio and particle size on solid-state anaerobic digestion. Bioresour. Technol., 141–148. Mshandete, A.M., Björnsson, L., Kivaisi, A.K., Rubindamayugi, M.S.T., Mattiasson, B., 2008. Performance of biofilm carriers in anaerobic digestion of sisal leaf waste leachate. Electron. J. Biotechnol. 11 (1). Nabizadeh, R., Naddafi, K., Mesdaghinia, A., Nafez, H.A., 2008. Feasibility study of organic matter and ammonium removal using loofa sponge as a supporting medium in an aerated submerged fixed-film reactor (ASFFR). Electron. J. Biotechnol. 11 (4). Ren, N., Wang, A., 2004. Principle and Application of Anaerobic Biotechnology. Chemical Industry Press, 89–138. Sasaki, K., Morita, M., Hirano, S.-I., Ohmura, N., Igarashi, Y., 2009. Effect of adding carbon fiber textiles to methanogenic bioreactors used to treat an artificial garbage slurry. J. Biosci. Bioeng. 108 (2), 130–135. Sowmeyan, R., Swaminathan, G., 2008. Performance of inverse anaerobic fluidized bed reactor for treating high strength organic wastewater during start-up phase. Bioresour. Technol. 99 (14), 6280–6284. Tarjányi-Szikora, S., Oláh, J., Makó, M., Palkó, G., Barkács, K., Záray, G., 2013. Comparison of different granular solids as biofilm carriers. Microchem. J. 107, 101–107. Yu, H.Q., Zhao, Q.B., Tang, Y., 2006. Anaerobic treatment of winery wastewater using laboratory-scale multi- and single-fed filters at ambient temperatures. Process Biochem. 41, 2477–2481. Zhao, H., Li, J., Li, J., Yuan, X., Piao, R., Zhu, W., Li, H., Wang, X., Cui, Z., 2013. Organic loading rate shock impact on operation and microbial communities in different anaerobic fixed-bed reactors. Bioresour. Technol. 140, 211–219. Zhao, Q.B., Mu, Y., Wang, Y., Liu, X.W., Dong, F., Yu, H.Q., 2008. Response of a biohydrogen-producing reactor to the substrate shift from sucrose to lactose. Bioresour. Technol. 99, 8344–8347.
