Perturbation-independent community ... - Wiley Online Library

ARTICLE Perturbation-Independent Community Development in Low-Temperature Anaerobic Biological Wastewater Treatment Bioreactors Pa´dhraig Madden,1 Fabio A. Chinalia,1 Anne-Marie Enright,2 Gavin Collins,2 Vincent O’Flaherty1 1

Microbial Ecology Laboratory, Department of Microbiology, National University of Ireland, Galway, Ireland; telephone: +353 (0) 91 493734; fax: +353 (0) 91 494598; e-mail: [email protected] 2 Microbial Ecophysiology Research Group, Department of Microbiology and Environmental Change Institute, National University of Ireland, Galway, Ireland Received 13 April 2009; revision received 5 August 2009; accepted 7 August 2009 Published online 17 August 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bit.22507

Introduction ABSTRACT: The reproducibility and stability of lowtemperature anaerobic wastewater treatment systems undergoing transient perturbations was investigated. Three identical anaerobic expanded granular sludge bed-based bioreactors were used to degrade a volatile fatty acid and glucose-based wastewater under sub-ambient (158C) conditions. The effect of a variety of environmental perturbations on bioreactor performance was assessed by chemical oxygen demand removal. Temporal microbial community development was monitored by denaturation gradient gel electrophoresis (DGGE) of 16S rRNA genes extracted from sludge granules. Methanogenic activity was monitored using specific methanogenic activity assays. Bioreactor performance and microbial population dynamics were each well replicated between both experimental bioreactors and the control bioreactor prior to, and after the implementation of most of the applied perturbations. Gene fingerprinting data indicated that Methanosaeta sp. were the persistent, keystone members of the archaeal community, and likely were pivotal for the physical stability and maintenance of the granular biofilms. Cluster analyses of DGGE data suggested that temporal shifts in microbial community structure were predominantly independent of the applied perturbations. Biotechnol. Bioeng. 2010;105: 79–87. ß 2009 Wiley Periodicals, Inc. KEYWORDS: low-temperature; anaerobic digestion; methanogenic consortia; wastewater

Fabio A. Chinalia’s present address is Centre for Resource Management and Efficiency, School of Applied Science, Cranfield University, College Road, Cranfield, Bedfordshire MK43 0AL, UK. Correspondence to: V. O’Flaherty Contract grant sponsor: Irish Research Council for Science and Engineering Technology (IRCSET)

ß 2009 Wiley Periodicals, Inc.

Anaerobic digestion is now an established and proven technology for the treatment of a wide variety of industrial wastewaters (Bouallagui et al., 2005; Elmitwalli et al., 2001; Macarie, 2000; Rincon et al., 2006). However, the majority of full-scale applications and research effort has been concentrated on anaerobic digestion within the mesophilic (25–458C) or thermophilic (45–658C) temperature ranges. This was mostly due to the belief that low-temperature (90% was achieved within 5 days (Fig. 2).

Origin and closest relatives of excised DGGE bands.

DGGE Band 1 2 3 4 5

arithmetic averages (UPGMA) similarity dendrograms with Jaccard’s co-efficient as the distance co-efficient using the MVSP 3.1 statistical package (Kovach, 1999). Non-metric multi-dimensional scaling (NMDS) scatter plots were constructed using the palaeontological statistics software package (PAST Ver 1.82; Ryan et al., 1995). Both analytical tools were used to infer temporal archaeal and bacterial community structure succession in biomass from the sludge beds.

Genbank accession DQ867002 DQ903695 FJ639251 FJ639252 FJ639253

Biomass R1 day Day 0 R3 day R1 day R3 day

253 217 217 217

Biotechnology and Bioengineering, Vol. 105, No. 1, January 1, 2010

Closest relative from blastn (accession number)

Similarity (%)

Uncultured Methanosaeta sp. Methanospirillum hungatei Uncultured Methanomethylovorans sp. Uncultured Methanomicrobiales archaeon Methanocorpusculum bavaricum

100 100 99 97 100

Figure 2. COD removal efficiency of R1 (O), R2 (&), and R3 (~) following each of the five perturbations applied. A: Three-times and six-times OLR increases (P3, P4); (B) HRT reduction (P5); (C) aeration perturbation (P6); (D) temperature reduction (P7).

Specific Methanogenic Activity The SMA of the seed sludge against each of the substrates was higher when tested at 378C than at 158C (Table II). Highest activity under both temperature conditions was recorded against H2/CO2, with the exception of the 378C assay with ethanol as the sole substrate. Assays performed using biomass sampled during P2 (before the first perturbation) also indicated higher activity at 378C than at 158C against each of the substrates tested. Nevertheless, increased methanogenic activity was observed at both assay temperatures compared to the seed sludge; the only exception to this was in the case of propionate-fed assays, which indicated a replicated reduction in the potential activity of propionate-oxidizers (Table II). In R1 biomass, the highest activity at both temperatures was recorded against acetate, whereas higher activity against H2/CO2 was observed in R2 and R3 (Table II). No development of the SMA of R1 biomass against acetate or H2/CO2 was apparent by P6 at 158C, compared to tests conducted on day 202 at the same temperature. (Table II). On the other hand, however, enhanced

propionate-degrading activity was measured at both assay temperatures (Table II). R1 methanogenesis was stillpredominantly acetoclastic. However, no trend was apparent from the R2 or R3 data produced from acetoclastic or hydrogenotrophic assays. SMA assays with biomass from the conclusion of the trial (P7, day 361) again indicated higher activity at 378C than at 158C against all substrates tested (Table II). The only exception to this was in the case of H2/CO2-fed assays with R3 biomass (Table II), which indicated higher potential hydrogenotrophic methanogenic at 158C than at 378C. An obvious deviation between acetoclastic and hydrogenotrophic methanogenic potential was not apparent in R2 biomass at 158C, but H2/CO2-mediated activity appeared dominant at 158C in R3 biomass (Table II).

Microbial Community Development First, for the Archaea, a replicated shift in the structure of the communities of all three bioreactors was apparent after P1 (Figs. Fig. 3A and Fig. 4A). A further shift was observed after

Madden et al.: Reproducibility of Low-Temperature Anaerobic Digestion Biotechnology and Bioengineering

83

Figure 3.

Non-metric multi-dimensional scaling (NMDS) illustrating (A) archaeal and (B) bacterial DGGE for R1 (O), R2 (&) and R3 (~).

P2, coinciding with the conclusion of the first 200 days of operation, but this was not closely replicated across R1–R3 (Figs. 3A and 4A). Community succession during P3 and P4 did not appear linked to the applied perturbations. A replicated shift in the R2–R3 community was observed after P5, indicating a response to the HRT perturbation (Figs. 3A and 4A). However, the archaeal communities of R1–R3 appeared similar after P6, thus suggesting no specific impact on the two experimental bioreactors. Nevertheless, the R1 and the well replicated, R2–R3 communities appeared clearly different after P7 (temperature shock; Figs. 3A and 4A). Relatively well-replicated bacterial communities, relative to the seed sludge, were observed at P1, and the community structure at P2 was similar across R1–R3 (Figs. 3B and 4B). No relationship was apparent between the perturbations and the R2–R3 bacterial communities at P3, P4, or P6, as all bioreactors are grouped independent of the other

84


Figure 4. Unweighted pair-group methods using arithmetic averages (UPGMA) dendrograms, with associated banding patterns, illustrating (A) archaeal and (B) bacterial DGGE samples. Excised bands, 1–5, are indicated.

(Figs. 3B and 4B). However, a replicated response to the HRT reduction was apparent by the end of P5, due to the fact that both experimental bioreactors were grouped close to each other, away from the control bioreactor (Figs. 3B and 4B). Nevertheless, and although a relatively well-replicated community structure was again apparent in the three bioreactors in P7, the succession appeared to have proceeded independently of the applied shocks (Figs. 3B and 4B).

Discussion The stability and reproducibility, as well as the response to operational perturbations, of low-temperature anaerobic digestion are critical considerations for the full-scale implementation of this technology. The potential value of genetic fingerprinting to help monitor bioreactor perfor-

mance has been demonstrated by, for example, McHugh et al. (2006), who reported that process disturbance and the disintegration of sludge bed granules could be predicted well in advance by using molecular bio-monitoring. In that study, a distinct shift in archaeal community structure was observed during a 500-day trial, using terminal restriction fragment length polymorphism (TRFLP) analysis, with a decrease in the relative abundance of Methanosaeta sp. and a proliferation of hydrogenotrophic methanogens (Methanomicrobiales sp.). Although the population change was well underway by day 300, the granulation and performance problems did not manifest until day 425 (McHugh et al., 2006). The possibility of applying nucleic acids-based tools to predict operational problems underscores the potential of a much broader field of environmental diagnostic platform technologies, which are certain to be important tools in waste-to-energy biotechnologies in the future. The period of recovery following the applied OLR perturbation (P4) was the longest observed during the trial. This is, perhaps, not surprising, as previous trials indicate that well-functioning bioreactors, such as R1–R3, may be more susceptible to disturbance due to low levels of key populations (McCarty and Mosey, 1991). On the other hand, McMahon et al. (2004) observed that a history of poor performance promotes the establishment of microbial communities better equipped to deal with extreme organic overloading. The response of anaerobic digestion to organic overloading depends on several factors, such as the duration of the overloading, characteristics of the feed and the activity of the biomass (Stamatelatou et al., 2003). Langenhoff and Stuckey (2000) suggested that a surplus of active biomass in anaerobic bioreactors, such as was the case in this study, likely results in greater stability when environmental perturbations are applied. Sabry (2008) found that temporary OLRs of 5–6 times the average rate could be successfully applied to sewage-degrading UASB bioreactors. Despite the deterioration in COD removal efficiencies after the HRT perturbation, effluent analyses indicated no VFA accumulation to account for the reduced COD removal. This is contrary to the work of Collins et al. (2005), where accumulation of VFAs, principally acetate and propionate, were observed in bioreactor effluent after the implementation of temperature and HRT perturbations. Increased upflow velocity and elevated agitation associated with reduced HRT may serve to erode anaerobic granules and result in the washout of disseminated particles. Washout was observed after the HRT perturbation, but only for one day after the resumption of the 24 h HRT. Rebac et al. (1995) also witnessed this washout of disseminated matter in the initial stages of EGSB bioreactor start-up with a 2.5 h HRT and an operating temperature of 118C. Methanosaeta sp. play a pivotal role in granulation (De Zeeuw, 1987; Dubourgier et al., 1987; McLeod et al., 1990; Wiegant, 1987), as either precursors for granule development (Wiegant, 1987) or as a network that stabilizes biofilm structure (Dubourgier et al., 1987). Although altered microbial community structure was observed in response

to the HRT perturbation, and was likely due to sludge washout, Methanosaeta-like species were detected in R1, R2, and R3 granules throughout the entire trial, indicating that those organisms were keystone members of the community, irrespective of the applied perturbations. The effect of the aeration perturbation on COD removal efficiency in R2–3 was negligible. We posit that the granular structure of the biofilms promoted oxygen tolerance. This seems to be the case for Methanosarcina, for example, which was more oxygen tolerant as cell aggregates than as dispersed cells (Kiener and Leisinger, 1983). Kato et al. (1997) formulated a hypothesis to describe the oxygen tolerance of methanogens in granular sludges. They proposed that facultative bacteria rapidly consume oxygen, creating anaerobic microniches inside the granules where methanogens are well protected against contact with oxygen. Indeed, this theory can be endorsed by literature demonstrating the poor penetration of oxygen into biofilms. Oxygen penetration into actively respiring aerobic biofilms has been reported to reach a depth of only 100–300 mm (de Beer, 1990; Hooijmans, 1990). Moreover, the controlled and regulated addition of oxygen to anaerobic digesters could certainly be used to enhance the degradation of a number of recalcitrant pollutants (Kato et al., 1997) but would likely lead to decreased methane yield. The reduced physiological activity of the microbial consortia in R2–3 in response to the temperature shock was exemplified by poor COD removal efficiencies. It is likely that the carbon supply to the granules was also negatively affected by operation at 48C, due to the increased viscosity associated with lower temperatures. Thus, more energy is required for sludge bed mixing, which is problematic, particularly at low biogas production rates (Lettinga et al., 2001; Nachaiyasit and Stuckey, 1997). Moreover, lower temperatures result in reduced diffusion of soluble compounds (Lettinga et al., 2001), thus leading to decreased substrate diffusion into anaerobic granules. In addition, the Arrhenius equation (Levenspiel, 1972) indicates that reaction rates will half with each 108C reduction in incubation temperature. In this experiment, the temperature was decreased by 118C (for 48 h or 2 HRTs) but COD removal efficiency was reduced by just 37.5% (R2) and 26.5% (R3) (Fig. 1). This was likely due to surplus active biomass R2 and R3, which contributed to the stability and robustness of the bioreactors. Typically, bioreactors are thought to possess a metabolic overcapacity (i.e., capacity to catabolize considerably more feed at maximum growth rates than is fed) (Langenhoff and Stuckey, 2000). The response of R2 and R3 COD removal to this perturbation was remarkably similar, and the rapid recovery is in agreement with the findings of Rebac et al. (1995), who described a 2day recovery period in VFA-degrading EGSB bioreactors after the reduction of the operating temperature to 5–78C (from 10 to 128C; Rebac et al., 1995). Collins et al. (2005) compared the performance, and temporal bacterial and archaeal communities, of two identical EGSB bioreactors, to which a series of replicated


85

perturbations, including of hydraulic loading, oxygenation, temperature and upflow velocity, was applied. The study indicated that the replicated response of anaerobic bioreactors, in terms of COD removal, VFA degradation and accumulation, development of methanogenic activity and population dynamics, with respect to the operational pressures applied, may be predictable. However, a control bioreactor to monitor dynamics in an unperturbed system was not employed. In this study, our data from R1 and from the two perturbed bioreactors support the findings of Collins et al. (2005) insofar as we describe replicated, robust treatment in R2 and R3 with respect to the periods (P3–P7). However, and importantly, R1 microbial community succession closely follows the temporal community structure of R2 and R3 granules. In fact, NMDS analyses of the DGGE data indicate that community development in R2 and R3 was largely independent of the applied perturbations (Fig. 3A and B). Most of the temporal development of the archaeal community in R2 and R3 appeared independent of perturbations; however, a distinct R2–R3 structure was apparent by the end of P7. Rather than a specific response to the temperature perturbation, however, we suggest that this may be due to a cumulative response over time to the lowtemperature conditions. Although a Methanosaeta-like ribotype (band 1) was present in all of the DGGE samples analyzed, and appeared to be a keystone member of the archaeal community, each of the remaining excised bands indicated the emergence during the trial of hydrogenotrophic methanogens. This phenomenon was observed across each of the reactors, and has been observed previously in low-temperature bioreactors (e.g., Enright et al., 2009). In fact, a Methanocorpusculum-like methanogen appears dominant in R1–R3 by the conclusion of the trial (Fig. 4A). Those observations are supported by the SMA data, which also indicate the temporal development of hydrogenotrophic methanogenic activity in each of the reactors. In that case, one possible explanation for the low effluent acetate concentrations is the activity of syntrophic acetate oxidizing bacteria. However, this is an energetically unfavorable reaction (DG80 ¼ þ 104.6 kJ/mol; Hattori, 2008) and the role of syntrophic acetate oxidizers in methanogenic bioreactors is still not clear, but will likely attract increasing interest in the future. Similarly, for the Bacteria, community succession only once appeared to be a response to perturbation (P6). Otherwise, bacterial community structure developed with time, rather than with perturbations.

Conclusions The data indicate that the effect of loading and HRT perturbations on microbial community structure cannot be measured, or indeed, that operating conditions did not significantly alter the temporal community development. Microbial succession independent of the applied shocks

86


indicates the robustness of the reactor biomass, which is relevant for successful wastewater treatment under the lowtemperature conditions applied. The receipt of financial support from the Irish Research Council for Science and Engineering Technology (IRCSET) under the Embark initiative, and from Science Foundation Ireland, is gratefully acknowledged.

References APHA. 1998. Standard methods for the examination of water and wastewater, 20th edn. Washington, DC: APHA. Baker GC, Smith JJ, Cowan DA. 2003. Review and re-analysis of domainspecific 16S primers. J Microbiol Methods 55:541–555. Bouallagui H, Touhami Y, Ben Cheikh R, Handi M. 2005. Bioreactor performance in anaerobic digestion of fruit and vegetable wastes. Process Biochem 40:989–995. Coates JD, Coughlan MF, Colleran E. 1996. Simple method for the measurement of the hydrogenotrophic methanogenic activity of anaerobic sludges. J Microbiol Methods 26:237–246. Colleran E, Concannon F, Golden T, Geoghegan F, Crumlish B, Killilea E, Henry M, Coates J. 1992. Use of methanogenic activity tests to characterise anaerobic sludges, screen for anaerobic biodegradability and determine toxicity thresholds against individual anaerobic trophic groups and species. Water Sci Technol 25:31–40. Collins G, Woods A, McHugh S, Carton MW, O’Flaherty V. 2003. Microbial community structure and methanogenic activity during start-up of psychrophilic anaerobic digesters treating synthetic industrial wastewaters. FEMS Microbiol Ecol 46:159–170. Collins G, Mahony T, O’Flaherty V. 2005. Stability and reproducibility of low-temperature anaerobic biological wastewater treatment. FEMS Microbiol Ecol 55:449–458. de Beer D. 1990. Microelectrode studies in biofilms and sediments. Ph.D. Dissertation, Universiteit van Amsterdam, Amsterdam, The Netherlands. De Zeeuw WJ. 1987. Granular sludge in UASB reactors. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW, editors. Granular anaerobic sludge: Microbiology and technology. The Netherlands: Pudoc. Wageningen. p 132–145. Dubourgier HC, Prensier G, Albagnac G. 1987. Structure and microbial activities of granular anaerobic sludge. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW, editors. Granular anaerobic sludge: Microbiology and technology. The Netherlands: Pudoc. Wageningen. p 18–33. Elmitwalli TA, Soellner J, De Keizer A, Bruning H, Zeeman G, Lettinga G. 2001. Biodegradability and change of physical characteristics of particles during anaerobic digestion of domestic sewage. Water Res 35: 1311–1317. Enright AM, McHugh S, Collins G, O’Flaherty V. 2005. Low-temperature anaerobic biological treatment of solvent-containing pharmaceutical wastewater. Water Res 39:4587–4596. Enright AM, McGrath V, Gill D, Collins G, O’Flaherty V. 2009. Effects of seed sludge and operation conditions on performance and archaeal community structure of low-temperature anaerobic solvent-degrading bioreactors. Syst Appl Microbiol 32:65–79. Griffiths RI, Whiteley AS, O’Donnell AG, Bailey MJ. 2000. Rapid method for coextraction of DNA and RNA from natural environments for analysis of ribosomal DNA and rRNA-based microbial community composition. Appl Environ Microbiol 66:5488–5491. Hattori S. 2008. Syntrophic acetate-oxidizing microbes in methanogenic environments. Microbes Environ 23(2):118–127. Hooijmans CM. 1990. Diffusion coupled with bioconversion in immobilised systems. Ph.D. Dissertation, Technische Universiteit Delft, The Netherlands.

Kato MT, Field JA, Lettinga G. 1997. Anaerobe tolerance to oxygen and the potentials of anaerobic and aerobic cocultures for wastewater treatment. Braz J Chem Eng 14(4):395–407. Kiener A, Leisinger T. 1983. Oxygen sensitivity of methanogenic bacteria. Syst Appl Microbiol 4:305–312. Kovach WL. 1999. MVSP—A Multivariate Statistical Package for Windows ver. 3.1. Pentraeth, Wales, UK: Kovach Computing Services. Langenhoff AAM, Stuckey DC. 2000. Treatment of dilute wastewater using anaerobic baffled reactor: Effect of low temperature. Water Res 34(15):3867–3875. Lettinga G, Rebac S, Parshina S, Nozhevnikova A, van Lier J, Stams AJM. 1999. High-rate anaerobic treatment of wastewater at low temperatures. Appl Environ Microbiol 65:1696–1702. Lettinga G, Rebac S, Zeeman G. 2001. Challenge of psychrophilic anaerobic wastewater treatment. Trends Biotechnol 19:363–370. Levenspiel O. 1972. Chemical reaction engineering, 2nd edn. New York: John Wiley. Macarie H. 2000. Overview of the application of anaerobic treatment to chemical and petrochemical wastewaters. Water Sci Technol 42(5–6): 201–214. McCarty PL, Mosey FE. 1991. Modelling of anaerobic digestion processes (A discussion of concepts). Water Sci Technol 24:17–33. McHugh S, Carton MW, Collins G, O’Flaherty V. 2004. Reactor performance and microbial community dynamics during anaerobic biological treatment of wastewaters at 16–378C. FEMS Microbiol Ecol 48: 369–378. McHugh S, Collins G, O’Flaherty V. 2006. Long-term, high-rate anaerobic biological treatment of whey wastewaters at psychrophilic temperatures. Biores Technol 97:1669–1678. McLeod FA, Guiot SR, Costerton JW. 1990. Layered structure of bacterial aggregates produced in an upflow anaerobic sludge bed and filter reactor. Appl Environ Microbiol 56(6):1598–1607. McMahon KD, Zheng D, Stams AJM, Mackie RI, Raskin L. 2004. Microbial population dynamics during start-up and overload conditions of anaerobic digesters treating municipal solid waste and sewage sludge. Biotechnol Bioeng 87(7):823–834. Muyzer G, Waal ECD, Uitterlinden AG. 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl Environ Micorbiol 59:695–700. Nachaiyasit S, Stuckey DC. 1997. The effect of shock loads on the performance of an anaerobic baffled reactor (ABR): 2. Step and transient hydraulic shocks at constant feed strength. Water Res 31:2747–2754. O’Flaherty V, Lens P. 2003. Biofilms in wastewater treatment. In: Lens P, Moran AP, Mahony T, Stoodley P, O’Flaherty V, editors. Biofilms in

industry, medicine and environmental technology. London: IWA Press. p 385–418. Rebac S, Ruskova J, Gerbens S, van Lier JB, Stams AJM, Lettinga G. 1995. High-rate anaerobic treatment of wastewater under psychrophilic conditions. J Ferm Bioeng 5:499–506. Rincon B, Raposo F, Borja R, Gonzalez JM, Portillo MC, Saiz-Jimenez C. 2006. Performance and microbial communities of a continuous stirred tank reactor treating two-phases olive mill solid wastes at low organic loading rates. J Biotechnol 121(4):534–543. Ryan PD, Harper DAT, Whalley JS. 1995. PALSTAT, statistics for palaeontologists. London, UK: Chapman & Hall (now Kluwer Academic Publishers). Sabry T. 2008. Application of the UASB inoculated with flocculent and granular sludge in treating sewage at different hydraulic shock loads. Biores Technol 99:4073–4077. Sambrook J, Fritsch E, Maniatis T. 1989. Molecular cloning: A laboratory manual. New York: Cold Spring Harbor Laboratory. Shelton DR, Tiedje JM. 1984. General method for determining anaerobic biodegradation potential. Appl Environ Microbiol 47:850– 857. Stamatelatou K, Vavilin V, Lyberatos G. 2003. Performance of a glucose fed periodic anaerobic baffled reactor under increasing organic loading conditions: 1. Experimental results. Biores Technol 88:131– 136. Stephenson RJ, Patoine A, Guiot SR. 1999. Effects of oxygenation and upflow liquid velocity on a coupled anaerobic/aerobic reactor system. Water Res 33(12):2855–2863. Stroot PG, McMahon KD, Mackie RI, Raskin L. 2001. Anaerobic codigestion of municipal solid waste and biosolids under various mixing conditions. Performance data. Water Res 35(7):1804–1816. van der Last ARM, Lettinga G. 1991. Anaerobic treatment of domestic sewage under moderate climatic (Dutch) conditions using upflow reactors at increased superficial velocities. In: Proc of the 6th International Symposium on Anaerobic Digestion. Sao Paulo, Brazil. p 153– 154. Whitman WB, Bowen TL, Boone DR. 1992. The methanogenic bacteria. In: Balows A, Tramper HG, Dworkin M, Harder W, Schleifer KH, editors. The Prokaryotes, a handbook on the biology of Bacteria: Ecophysiology, isolation, identification, applications, Vol. 1, 2nd edn. New York: Springer Verlag. p 719–767. Wiegant WM. 1987. The ‘spaghetti theory’ on anaerobic sludge formation, or the inevitability of granulation. In: Lettinga G, Zehnder AJB, Grotenhuis JTC, Hulshoff Pol LW, editors. Granular anaerobic sludge: Microbiology and technology. The Netherlands: Pudoc. Wageningen. p 146–152.


87