changes in microbial ecology in an anaerobic reactor

Bioresource Technology 97 (2006) 1201–1208

Changes in microbial ecology in an anaerobic reactor

Burak Demirel, Orhan Yenigun *

Institute of Environmental Sciences, Bogazici University, Bebek 34342, Istanbul, Turkey

Received 26 January 2004; received in revised form 8 December 2004; accepted 17 May 2005Available online 1 August 2005

Abstract

This study examined the behaviour of the microbial population in an anaerobic reactor, in terms of changes in numbers of totalbacterial community, autofluorescent methanogens, non-methanogens and morphology of the autofluorescent methanogens, usingepifluorescence microscopy and microbiological enumeration techniques. A laboratory-scale, continuous flow-completely mixedanaerobic reactor, coupled with a conventional gravity settling tank and a continuous recycling system, was operated at anHRT range between 24 and 12 h, using dairy wastewater as the substrate. The numbers of the total bacterial community and auto-fluorescent methanogens both decreased during start-up. Also, the proportion of the number of autofluorescent methanogens in thetotal bacterial community varied from 5% to 16% during operation. In particular, the activity of the methane-forming bacteriadecreased significantly at HRTs of 16 and 12 h. A membrane module, instead of a conventional settling tank, would obviously havebeen a more effective method if recycling were required in the anaerobic treatment system.� 2005 Elsevier Ltd. All rights reserved.

Keywords: Anaerobic digestion; Enumeration; Microbiological; Dairy wastewater; Methanogens

1. Introduction

Two-phase anaerobic digestion processes have beenwidely applied for waste treatment. Pohland and Ghosh(1971) first proposed the concept of a two-phase anaer-obic digestion process. The aim was to separate the acidand methane fermentation phases for providing moreattention directed toward determining and satisfyingthe optimum environmental conditions for each micro-bial community in two separate reactor systems. Two-phase anaerobic digestion processes offer significantadvantages in comparison to single phase anaerobicwaste treatment systems. These advantages primarilyinclude increased process stability and control, a higherspecific activity of methanogens and optimization ofenvironmental conditions required for each separate

0960-8524/$ - see front matter � 2005 Elsevier Ltd. All rights reserved.doi:10.1016/j.biortech.2005.05.009

* Corresponding author. Tel.: +90 212 3596946; fax: +90 2122575033.

E-mail address: [email protected] (O. Yenigun).

reactor system (Massey and Pohland, 1978; Cohenet al., 1979).

A considerable amount of research has been carriedout on the benefits of two-phase anaerobic digestionprocesses. More recently, the microbial ecology ofanaerobic reactor systems has also been investigated indetail (Delbes et al., 2001; Ahring et al., 2001; Gerardi,2003; Collins et al., 2003; McHugh et al., 2004). How-ever, only a limited number of studies cover the micro-bial ecology of acidification reactors (Cha and Noike,1997; Ince and Ince, 2000; Solera et al., 2002; Carboneet al., 2002). In addition, a few other studies havefocused particularly on start-up (Anderson et al., 1994;Liu et al., 2002). It is obvious that the performance ofan anaerobic reactor is primarily determined by theamount of active microorganisms retained within thesystem. Besides, changes in operational and environ-mental conditions of the anaerobic reactor and withinthe microbial populations present in the reactor defi-nitely affect each other mutually. Thus, further evalua-tion of these variations seems extremely useful from a

mailto:[email protected]

1202 B. Demirel, O. Yenigun / Bioresource Technology 97 (2006) 1201–1208

microbiological point of view. The aim of this study wasto investigate the behaviour of the microbial popula-tions within an anaerobic reactor, operated at a hydrau-lic retention time (HRT) range between 24 and 12 h, interms of changes in numbers of total bacterial commu-nity, autofluorescent methanogens, non-methanogensand morphology of the autofluorescent methanogens,using epifluorescence microscopy and microbiologicalenumeration techniques.

2. Methods

2.1. Experimental set-up

The laboratory-scale experimental set-up consisted ofa feed tank, a feed pump, an anaerobic reactor, a gravitysettling tank, a recycle pump, an effluent tank and a wettest meter. The schematic configuration of the experi-mental set-up is shown in Fig. 1. The feed tank was con-structed of plexiglass, and it was continuously stirred bya Welp magnetic stirrer. The continuous flow completelymixed anaerobic reactor was also constructed of plexi-glass, and had a working volume of 5.2 l. The influentfeeding port and gas outlet were located on the top coverof the reactor. The gas outlet was connected to a wet testmeter. Biogas production was monitored using thewet test meter (GCA/Precision Scientific Chicago/IL,USA). There was a water jacket around the reactorand an aquarium heater (Visi Therm, model numberVTE 300) was placed in this water jacket, in order toprovide the required temperature range. Temperaturewas maintained within a range of 35 ± 1 �C during the

Influent Feed

Feeding Pump

Recycle

Feeding TankReac

Rec

MaS

MagneticStirrer

Fig. 1. The configuration of t

entire study. No pH control was exerted on the system.Feeding was carried out using Masterflex C/L Cole-Parmer and Watson Marlow pumps. HRT was adjustedvolumetrically, through controlling the flow rate of theinfluent feed. Mixing was provided using a Welp mag-netic stirrer located underneath the reactor. The reactoreffluent overflowed to the settling tank. The settledsludge was continuously recycled to the reactor using aWelp pump. The supernatant overflowed to the effluenttank. Mixed samples were taken from the reactor understeady-state conditions for further analyses.

2.2. Feed and seed

Industrial wastewater was obtained from a dairyplant manufacturing yoghurt. The general characteris-tics of the dairy wastewater are given in Table 1. Waste-water was always kept at 4 �C prior to use. Feed wasprepared daily, using dairy wastewater, micro andmacro nutrients and distilled water. Lactose was alsoadded to the feed frequently, in order to increase theinfluent feed strength. Alkalinity was provided usingsodium hydrogen carbonate (NaHCO3) in the feed.Nitrogen was seldom added to the feed, in the form ofammonium chloride (NH4Cl), and phosphorus wasnot used, since dairy wastewater contained the requiredamounts of both nutrients. Stock solutions of 100 mg/lNiSO4 Æ 6H2O, 10 mg/l (NH4)6Mo7O24 Æ 4H2O and100 mg/l CoCl2 Æ 6H2O were also used, in order toprovide required amounts of nickel (Ni), molybdenum(Mo) and cobalt (Co) to the feed as micro nutrients.All of the chemicals were reagent grade, obtained fromcommercial sources.

BiogasTo Wet Test Meter

Gas Outlet

Effluent

tor SettlingEffluent Tank

Tank

ycle Pump

gnetictirrer

he experimental set-up.

Table 1Characteristics of dairy wastewater

Parameter Unit Concentration

pH 5.8–11.4Total solids mg/l 2705–3715Total volatile solids mg/l 1196–1804Suspended solids mg/l 340–1730Volatile suspended solids mg/l 255–830Total COD mg/l 1155–9185Soluble COD mg/l 550–5960TKN mg/l 14–272Total phosphorus mg/l 8–68Alkalinity mg CaCO3/l 316–972Lipid mg/l 7–62Chloride mg/l 233–993Calcium mg/l 12–120Magnesium mg/l 2–97Sodium mg/l 123–2324Potassium mg/l 8–160Copper mg/l 0.01–0.1Manganese mg/l 0.03–0.4Nickel mg/l 0–0.1Iron mg/l 0.5–7Zinc mg/l 0.05–0.4

B. Demirel, O. Yenigun / Bioresource Technology 97 (2006) 1201–1208 1203

The anaerobic reactor was seeded with a granularsludge obtained from the upflow anaerobic sludge blan-ket (UASB) reactor of the wastewater treatment plant ofa local alcohol distillery facility. Total solids (TS) andtotal volatile solids (TVS) concentrations of the seedsludge were determined to be 90,225 and 82,550 mg/l,respectively, prior to inoculation.

2.3. Analytical methods

Hach COD Reactor Model 45600 (digestion at150 �C for 2 h) and a Hach DR/2010 spectrophotometerwere used for chemical oxygen demand (COD) analyses.Monitoring of pH was carried out using a WTW pH 330pH meter with a WTW SenTix probe. Influent and efflu-ent volatile fatty acid (VFA) concentrations were deter-mined using a HP 5890 Series II Gas Chromatographwith a flame ionization detector (FID) and an HP-FFAP column. The lipid contents of the samples weredetermined using the partition-gravimetric method asoutlined in Standard Methods (APHA, 1992). TKN,total phosphorus, COD, alkalinity, total solids, totalvolatile solids, suspended solids, volatile suspendedsolids and chloride analyses were also performedaccording to the methods outlined in Standard Methods(APHA, 1992). Analyses of calcium, magnesium, potas-sium, sodium, iron, nickel, copper and zinc wereperformed using a Perkin–Elmer Analyst 300 AAS.

Samples for microbiological analyses were takenfrom the reactor at each steady-state condition. Enu-meration studies were carried out immediately aftersampling. A method based on Pike et al. (1972) was usedfor homogenization of sludge samples. Enumeration of

the total bacterial community and autofluorescentmethanogens in homogenized and diluted samples wascarried out using an OLYMPUS BX-50 Model Epifluo-rescence Microscope fitted with a 100 W high-pressuremercury lamp. A magnification of 600 was used withOlympus ·60 water immersion lenses having a ·10 eye-piece. Differentiation between the methanogenic andnon-methanogenic bacteria was achieved by determin-ing the number of autofluorescent cells when irradiatedwith UV light. Morphological changes within the fluo-rescent methanogenic population were determined bysubdividing the total population into six morphologi-cally distinct groups; cocci, short rods (0.2–0.5 · 3 lm),medium rods (0.3–0.6 · 6 lm), long rods (0.3–0.6 ·10 lm), filaments and sarcina (Morgan et al., 1991).The samples were counted with a Neubauer Chamberwhich had a depth of 0.1 mm and an area of 1 mm2.There were 5 · 5 squares, and each square contained16 small squares (4 · 4) in the chamber. One drop ofsample was placed onto the Improved Neubauer Cham-ber. Among 25 squares, each 4 · 4, 5 squares were cho-sen randomly for counting and an average was taken.Bacteria were counted only on to the top and the leftgraduation lines of each small cube in order to preventduplication errors (Ince and Ince, 2000). Volatile sus-pended solids (VSS) content of samples were determinedjust before sample preparation. The following formulawas used to calculate the numbers of the total bacterialcommunity and autofluorescent methanogens per unitvolume (Ince and Ince, 2000):

N ¼ ðC � DF Þ=V

N, number of organisms per unit volume; C, mean countper square; DF, dilution factor; V, volume of the area(4 · 10�6 ml).

Standard deviations of the measurements for totalbacterial community, total autofluorescent methano-gens, non-methanogens, Methanococcus-like species,small rods, medium rods and long rods were computedto be 0.34, 0.43, 0.49, 0.46, 3.73, 2.56 and 3.44, respec-tively, during the entire work. The reactor was operatedfor about three weeks at each HRT in order to attainsteady-state conditions. After this stage, operation wascontinued for another week. All experimental resultsare reported for steady-state conditions.

3. Results and discussion

3.1. Reactor operation and performance

The reactor was operated at an HRT of 24 h and atan organic loading rate (OLR) range varying between1 and 5 kg COD/m3 d during start-up. After the reactorreached steady-state conditions at 24 h HRT and anOLR of about 4.6 kg COD/m3 d at the end of the

0

400

800

1200

1600

2000

4 6 8 10Organic loading rate

(kg COD/m3/d)

Tota

l vol

atile

fatty

aci

d (V

FA) c

once

ntra

tion

(mg/

l)

10

14

18

22

26

Hyd

raul

ic re

tent

ion

time

(hou

r)

Total volatile fatty acidconcentrationHydraulic retention time

Fig. 2. Total VFA production (mg/l) versus OLR (kg COD/m3 d) andHRT (h).

Table 2Operating conditions and the performance of the reactora

HRT (h) OLR (kg COD/m3 d) pH Reactor MLVSS (mg/l)

24 4.6 7.5 570522 4.9 7.1 946520 5.3 7.1 12,36518 5.5 6.8 839516 6.4 6.6 505012 9.3 7.0 4775

a The reactor was operated for a period of 7 HRTs under steady-state conditions for each HRT level.


start-up phase, HRT was decreased, to 22, 20, 18, 16and 12 h, respectively, in order to promote VFA produc-tion in the system. VFA production was negligible at anOLR of 4.6 kg COD/m3 d and 24 h HRT. Furthermore,VFA production seemed to be affected slightly by varia-tions in the OLR range between 4.9 and 5.5 kg COD/m3 d (between 22 and 18 h HRT). However, VFA pro-duction increased significantly at OLRs of 6.4 and9.3 kg COD/m3 d (at 16 and 12 h HRT), respectively.Total VFA production versus OLR and HRT is givenin Fig. 2. Generally, VFA production increased pro-portionally to the OLR. Acetic, propionic, butyric andvaleric acids were commonly produced VFAs duringmesophilic acidogenesis of dairy wastewater. High pro-pionic acid production was encountered in the systemthroughout the study. This could have resulted fromthe breakdown of lactate to propionate. Lactate isknown to be the preferred substrate for propionate-forming bacteria, which dominate at low HRTs. Forcomplex types of wastes, high propionic acid productioncan be encountered. Besides, the changes in VFA pro-duction could also be explained by a population selec-tion according to the type of substrate (Dinopoulouet al., 1988). Distribution of net VFA production be-tween 24 and 12 h HRT is given in Fig. 3. Net VFA pro-duction is the difference between influent and reactoreffluent VFA samples at steady-state conditions.

The operating conditions and the performance of thereactor for the entire operation are outlined in Table 2.

0100200300400500600700

10 12 14 16 18 20 22 24 26Hydraulic retention time

(hour)

VFA

con

cent

ratio

n (m

g/l)

HAC

HPRO

HISOBUT

HBUT

HVAL

HCAP

Fig. 3. Net VFA (mg/l) distribution versus HRT (h).

As HRT was decreased, the OLR increased, and the rateof acid production increased proportionally to the OLR(Fig. 2). The maximum rate of acid production wasachieved at the highest OLR and the lowest HRT levelsapplied. Poor rates of acid production between 24 and18 h could be attributed to recycling of biomass andhigh pH. Changes in the reactor mixed liquor volatilesuspended solids (MLVSS) concentrations are alsoshown in Table 2. In spite of the conventional gravitysettling tank and continuous recycling system, it was stilldifficult to maintain a steady-state MLVSS concentra-tion in the reactor throughout the operation. Since grav-ity settling for solids separation is strongly dependent onthe settling properties of the anaerobic flocs, solids settleability can often be problematic, because active an-aerobic flocs usually have biogas associated with them.Ghosh et al. (1975), Massey and Pohland (1978), Ghosh(1981), and Elefsiniotis and Oldham (1994) operatedacid-phase digesters with conventional settling systems.According to Ghosh (1981), the beneficial effects of con-centrating the cell mass and substrate in the digestereffluent and recycling them to the digester increase sig-nificantly at shorter HRTs. However, Banerjee et al.(1998) reported that more biomass was recycled intothe reactors, due to good settling characteristics inthe clarifier, at high HRTs (18 and 30 h). Decrease inreactor MLVSS concentrations towards the end of thisstudy, particularly at 16 and 12 h HRT, could be attrib-uted to higher flow rates and less time for settling of thebiomass, as reported previously by Banerjee et al.(1999). Furthermore, Banerjee et al. (1999) also statedthat the increased temperature levels made the settlingcharacteristics of the biomass in the clarifier poorer.The temperature range of 35 ± 1 �C maintained in thiswork might also have affected the settling characteristicsof the biomass adversely. A membrane module, insteadof a conventional settling tank, would obviously havebeen a more effective method if recycling were requiredin an anaerobic treatment system.

3.2. Bacterial enumeration studies

Medium rods and Methanococcus-like species wereobserved to be the dominant methanogens in the seed


sludge prior to inoculation. No presence of filamentsand sarcina was encountered in the seed sludge. Theproportion of the number of autofluorescent methano-gens in the total bacterial community was 18% in theseed. The proportion of the number of autofluorescentmethanogens in the total bacterial community decreasedfrom 18% to 11% by the end of the start-up period. Thisproportion was reported to be between 0.01% and 1% byAnderson et al. (1994), at a pH between 5.0 and 5.5, andan HRT of 12 h, during start-up in pre-acidification ofdairy wastewater. Lower pH and HRT levels, in com-parison to this study, certainly provided this lower pro-portion. The numbers of total bacterial community andautofluorescent methanogens both decreased duringstart-up. Methanococcus-like species were still the mostdominant methanogens, with an increase in number,at the end of start-up. However, medium rods, whichconstituted another dominant methanogen in the seedsludge, seemed to decrease in number. Decreases innumbers of both total bacterial community and auto-fluorescent methanogens could be attributed to reactoroperation at a relatively low HRT of 24 h. Decrease inthe number of the autofluorescent methanogens seemedsignificant; however, they were active, because no acidformation could be observed at 24 h HRT. Accordingto Horiuchi et al. (2002) and Liu et al. (2002), the micro-bial population in the acid reactor depended on the pH.A high pH of about 7.5 seemed to provide high amountsof active methanogens in the system during this run, inspite of a lower HRT. The composition of wastewaterwas reported to be a very significant factor affectingthe bacterial composition of the anaerobic biomass,both during start-up and routine operations (Morganet al., 1991; McHugh et al., 2003). Thus, wastewatercomposition was another imperative variable that mightbe accounted for as well, since the seed sludge, whichwas obtained from a single-phase UASB reactor treatingalcohol distillery effluent, was acclimatized to dairywastewater. Variations in the numbers of autofluores-cent methanogens at the end of the start-up phase are

Fig. 4. Changes in the numbers of autofluores

shown in Fig. 4. Changes within the morphology ofthe autofluorescent methanogens during start-up couldhave been due to the type of substrate and/or to varia-tions in the OLR. Long rods could not be retained with-in the system at the end of the start-up phase. This couldhave resulted from a lower HRT (increased influent flowrate) and/or type of the substrate. Long rods might nothave acclimatized to dairy wastewater.

The granulated structure of the seed sludge disinte-grated soon after the reactor started operation. Liuet al. (2002) also reported disintegration of granulatedseed sludge during start-up of acidogenic reactors. Thiscould be an important reason for high MLSS andMLVSS concentrations in the reactor effluent duringthe rest of this study (data not shown), since biogranuleswere reported to settle better, due to their larger sizes,than the suspended sludge in the reactor and, thus,had less of a tendency to be washed out (Fang, 2000).Disintegration of granules might have resulted due tothe degree of stirring in the reactor. Stirring was moder-ate during the entire operation.

The changes in the numbers of the total bacterialcommunity, non-methanogens and methanogens in theanaerobic reactor for the entire operation are shown inFig. 5. The number of non-methanogenic bacteria wasdetermined by subtracting the number of autofluores-cent methanogens from the total bacterial community(Anderson et al., 1994). The number of the total bacte-rial community firstly decreased between 24 and 20 hHRT, and then increased between 18 and 12 h. Thenumber of autofluorescent methanogens decreased sig-nificantly between 24 and 22 h HRT, then increased be-tween 20 and 16 h, and finally decreased again in thefinal run (at 12 h). The proportion of the number ofautofluorescent methanogens in the total bacterialcommunity is also displayed in Fig. 6. This proportionvaried between 5% and 16% during the operation. Ahigh proportion observed at 20 h HRT could be attrib-uted to a high pH of about 7.1 (Table 2). However,at 16 h HRT, in spite of a low pH of about 6.6, a high

cent methanogens at the end of start-up.

Fig. 5. Changes in numbers of the total bacterial community, non-methanogens and methanogens.

Fig. 6. Proportion of the number of autofluorescent methanogens inthe total bacterial community (%) versus HRT (h).


proportion was observed again. Despite a high numberof methanogens, their activity was not substantial,because the rate of acid formation increased substan-tially in this run, in comparison to 18 h HRT. The low-est proportion of 5%, which was attained at 12 h HRT,was also accompanied by the maximum rate of acid pro-duction. Ince and Ince (2000) reported the proportion ofautofluorescent methanogens in the total bacterial com-munity to be between 0.01% and 3%, in a continuousstirred tank reactor (CSTR), at an HRT and a pH of12 h and 5.5 to 6.0, respectively, during pre-acidificationof dairy wastewater. This ratio was much lower thanthat obtained in this work. The lower number of auto-fluorescent methanogens in the former study obviouslyresulted from a different reactor configuration (a CSTRwith no recycle) and a much lower operating pH.

The increase in the numbers of the total bacterialcommunity between 18 and 12 h HRT could have re-sulted from acclimation of the total bacterial communityto operation at lower HRTs and also from variations inthe composition of influent feed (Fig. 5). High fluctua-tions were observed in industrial wastewater characteris-

tics during the experimental study (Table 1). Thenumber of autofluorescent methanogens also decreasedsignificantly as HRT was decreased from 24 to 22 h(Fig. 5). However, between 20 and 12 h, the number ofautofluorescent methanogens increased. Methanogenscould also have acclimatized to reactor operating condi-tions and characteristics of the particular substrate. De-creases in numbers of both total bacterial communityand autofluorescent methanogens after 24 h HRT re-sulted from reactor operating conditions, and mainlyfrom gradual decrease in HRT. Cha and Noike (1997)also found that the bacterial populations in a mesophilicacidogenic reactor decreased with decrease in HRT from48 to 6 h. Since there was a continuous recycle systemand no pH control, complete wash-out of methanogensdid not occur in this study. However, the activity ofmethane formers seems to have been suppressed bydecreasing HRT, particularly to lower values, withinthe range investigated. Methanococcus-like species andmedium rods could generally be observed in the systemover the entire operation, despite variations in opera-tional parameters such as HRT and pH.

A key feature of the two-phase anaerobic process isthat it allows enrichment for different bacteria in eachdigester by independently controlling the operationalconditions (Cha and Noike, 1997). Employing ananaerobic reactor coupled with a recycle system as theacid-phase digester may decrease enrichment of the aci-dogenic bacteria in the digester, particularly if a mixedanaerobic culture is used as seed without pH control,which is indeed one of the most important advantagesof two-phase anaerobic digestion systems. However, thismay be overcome by proper adjustment of operationaland environmental parameters, such as HRT and/orpH. Since HRT was adjusted volumetrically in thisstudy, lower HRTs were provided by working at higherinfluent flow rates, thus, there was not sufficient time for


the methane-forming bacteria to consume food. Themethane formers could be retained within the reactordue to the continuous recycle system, but could not beactive, especially at lower HRT values.

Until now, few studies have investigated the effects ofreactor configuration on the microbial populations inanaerobic acidogenic reactors. In an earlier and exten-sive study, Morgan et al. (1991) investigated the effectsof different single-phase reactor systems on microbialcommunities, and concluded that the reactor configura-tion had no significant influence on the number of eachbacterial population. On the contrary, Liu et al. (2002)later reported that the reactor type was one of the majordeterminants of the microbial community shift. How-ever, both studies covered the start-up period. McHughet al. (2003) recently pointed out that variations in reac-tor design would result in changes within the microbialpopulations present in the system, due to the complexmicrobiological nature of the anaerobic digestion pro-cess. Since a continuous recycle system accompaniedthe anaerobic acidogenic reactor in this work, its effectson the behaviour of the microbial populations in thereactor should be taken into consideration.

Direct microscopic count is a rapid method for enu-meration of bacteria, despite the fact that it does notpermit distinction between viable and non-viable meth-anogens. This technique counts the number of methano-gens, but unfortunately does not completely reflect themethanogenic activity in the digester (Anderson et al.,1994). Moreover, Dolfing et al. (1985) also found that20–30% of methanogenic bacteria, such as Methanosa-

eta, do not exhibit fluorescence. Another disadvantageof the method was clumping of cells, which was commonin anaerobic digester sludge. Fading of the autofluores-cent cells and the presence of weakly or non-fluorescentcells might create inaccuracy, as reported previously byMorgan et al. (1991). Slight variations might also occur,due to errors in the sampling procedures and the enu-meration method.

4. Conclusions

Total VFA production in the anaerobic reactor in-creased proportionally to the OLR, with the decreasein HRT from 24 to 12 h. The maximum rate of acid pro-duction was achieved at the highest OLR and the lowestHRT levels applied. Poor rates of acid production be-tween 24 and 18 h could be attributed to recycling ofbiomass and high pH. Acetic, propionic, butyric andvaleric acids were commonly produced during meso-philic acidogenesis of dairy wastewater. High propionicacid production encountered during the entire operationcould have resulted from breakdown of lactate to propi-onate. In addition, the changes in VFA productioncould also be explained by a population selection

according to the type of substrate. The numbers of totalbacterial community and autofluorescent methanogensboth decreased during start-up. Changes within themorphology of the autofluorescent methanogens duringstart-up could be attributed to the type of substrate usedand/or to variations in the OLR. Although high num-bers of autofluorescent methanogens were present inthe seed sludge, their numbers decreased significantlyin the reactor during the operation. The ratio betweenthe number of autofluorescent methanogens and thenumber of the total bacterial community varied from5% to 16%. Since there was a continuous recycle systemand no pH control, complete wash-out of methanogensdid not occur, however, the activity of methane formersseemed to be suppressed by maintaining HRT at lowerlevels, within the HRT range investigated in this work.Decreases in reactor MLVSS concentrations towardsthe end of the study could be attributed to higher flowrates and less time for settling of the biomass. A mem-brane module, instead of a conventional settling tank,would have been a more effective method if recyclingwere required in the anaerobic treatment system.

Acknowledgement

The authors wish to express their gratitude to Boga-zici University Research Fund for the financial supportof this research with project numbers 99Y01D and01Y101D.

References

Ahring, B.K., Ibrahim, A.A., Mladenovska, Z., 2001. Effect oftemperature increase from 55 to 65 �C on performance andmicrobial population dynamics of an anaerobic reactor treatingcattle manure. Water Research 35, 2446–2452.

Anderson, G.K., Kasapgil, B., Ince, O., 1994. Microbiological study oftwo-stage anaerobic digestion during start-up. Water Research 28,2383–2392.

APHA, AWWA, WPCF, 1992. Standard Methods for the Examina-tion of Water and Wastewater, 18th ed. American Public HealthAssociation, Washington.

Banerjee, A., Elefsiniotis, P., Tuhtar, D., 1998. Effect of HRT andtemperature on the acidogenesis of municipal primary sludge andindustrial wastewater. Water Science and Technology 38 (8–9),417–423.

Banerjee, A., Elefsiniotis, P., Tuhtar, D., 1999. The effect of additionof potato-processing wastewater on the acidogenesis of primarysludge under varied hydraulic retention time and temperature.Journal of Biotechnology 72, 203–212.

Carbone, S.R., Da Silva, F.M., Tavares, C.R.G., Dias Filho, B.P.,2002. Bacterial population of a two-phase anaerobic digestionprocess treating effluent of cassava starch factory. EnvironmentalTechnology 23, 591–597.

Cha, G.C., Noike, T., 1997. Effect of rapid temperature change andHRT on anaerobic acidogenesis. Water Science and Technology 36(6–7), 247–253.


Cohen, A., Zoetemeyer, R.J., Van Deursen, A., Van Andel, J.G., 1979.Anaerobic digestion of glucose with separated acid production andmethane formation. Water Research 13, 571–580.

Collins, G., Woods, A., McHugh, S., Carton, M.W., O�Flaherty, V.,2003. Microbial community structure and methanogenic activityduring start-up of psychrophilic anaerobic digesters treatingsynthetic industrial wastewaters. FEMS Microbiology Letters1578, 1–12.

Delbes, C., Moletta, R., Godon, J.J., 2001. Bacterial and archaeal 16SrDNA dynamics during an acetate crisis in an anaerobic digestorecosystem. FEMS Microbiology Ecology 35, 19–26.

Dinopoulou, G., Sterritt, R.M., Lester, J.N., 1988. Anaerobic acido-genesis of a complex wastewater: 1. The influence of operationalparameters on reactor performance. Biotechnology and Bioengi-neering 31, 958–968.

Dolfing, J., Griffioen, A., van Neerven, A.R.W., Zevenhuizen,L.P.T.M., 1985. Chemical and bacteriological composition ofgranular methanogenic sludge. Canadian Journal of Microbiology31, 744–750.

Elefsiniotis, P., Oldham, W.K., 1994. Effect of HRT on acidogenicdigestion of primary sludge. Journal of Environmental Engineering120 (3), 645–660.

Fang, H.H.P., 2000. Microbial distribution in UASB granules and itsresulting effects. Water Science and Technology 42 (12), 201–208.

Gerardi, M.H., 2003. The Microbiology of Anaerobic Digesters. JohnWiley & Sons, Inc., USA.

Ghosh, S., Conrad, J.R., Klass, D.L., 1975. Anaerobic acidogenesisof wastewater sludge. Journal WPCF 47 (1), 30–45.

Ghosh, S., 1981. Kinetics of acid-phase fermentation in anaerobicdigestion. Biotechnology and Bioengineering 11, 301–313.

Horiuchi, J.I., Shimizu, T., Tada, K., Kanno, T., Kobayashi, M., 2002.Selective production of organic acids in anaerobic acid reactor.Bioresource Technology 82, 209–213.

Ince, B.K., Ince, O., 2000. Changes to bacterial community make-up ina two-phase anaerobic digestion system. Journal of ChemicalTechnology and Biotechnology 75, 500–508.

Liu, W.T., Chan, O.C., Fang, H.H.P., 2002. Microbial communitydynamics during start-up of acidogenic anaerobic reactors. WaterResearch 36, 3203–3210.

Massey, M.L., Pohland, F.G., 1978. Phase separation of anaerobicstabilization by kinetic controls. Journal WPCF 50 (9), 2204–2222.

McHugh, S., Carton, M., Mahony, T., O�Flaherty, 2003. Methano-genic population structure in a variety of anaerobic bioreactors.FEMS Microbiology Letters 219, 297–304.

McHugh, S., Carton, M., Collins, G., O�Flaherty, V., 2004. Reactorperformance and microbial community dynamics during anaerobicbiological treatment of wastewaters at 16–37 �C. FEMS Micro-biology Letters 48, 369–378.

Morgan, J.W., Evison, L.M., Forster, C.F., 1991. Changes to themicrobial ecology in anaerobic digesters treating ice cream waste-water during start-up. Water Research 25, 639–653.

Pike, E.B., Carrington, E.G., Ashburner, P.A., 1972. An evaluation ofprocedures for enumeration of bacteria in activated sludge. Journalof Applied Bacteriology 35, 309–321.

Pohland, F.G., Ghosh, S., 1971. Developments in anaerobic stabiliza-tion of organic wastes: The two-phase concept. EnvironmentalLetters 1, 255–266.

Solera, R., Romero, L.I., Sales, D., 2002. The evolution of biomass ina two-phase anaerobic treatment process during start-up. Chemicaland Biochemical Engineering Quarterly 16 (1), 25–29.

changes in microbial ecology in an anaerobic reactor

Documents