Bioresource Technology 170 (2014) 385–394

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

Transition of municipal sludge anaerobic digestion from mesophilicto thermophilic and long-term performance evaluation

http://dx.doi.org/10.1016/j.biortech.2014.08.0070960-8524/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Address: School of Civil and Environmental Engineering,Georgia Institute of Technology, 311 Ferst Drive, Atlanta, GA 30332-0512, USA. Tel.:+1 404 894 9367; fax: +1 404 894 8266.

E-mail address: spyros.pavlostathis@ce.gatech.edu (S.G. Pavlostathis).1 Present address: The Institute of Environmental Sciences, Bogazici University,

Istanbul 34342, Turkey.2 Present address: North American Höganäs, Johnstown, PA 15902-2904, USA.3 Present address: Civil Engineering Department, Tafila Technical University, Tafila

66110, Jordan.

Ulas Tezel 1, Madan Tandukar 2, Malek G. Hajaya 3, Spyros G. Pavlostathis ⇑School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0512, USA

h i g h l i g h t s

� Transition from 36 to 53.3 �Cdigestion at a rate of 3 �C/day wassuccessful, stable.� Operation at 60 �C led to relatively

stable gas production but high levelsof VFAs.� Methane production at 60 �C was

lower than at mesophilic conditions(36 �C).� For high performance of municipal

CSTR digesters, temperature shouldbe below 60 �C.

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

a r t i c l e i n f o

Article history:Received 3 June 2014Received in revised form 30 July 2014Accepted 2 August 2014Available online 8 August 2014

Keywords:AcidogenesisAnaerobic digestionMethanogenesisMunicipal sludgeTemperature transition

a b s t r a c t

Strategies for the transition of municipal sludge anaerobic digestion from mesophilic to thermophilic wereassessed and the long-term stability and performance of thermophilic digesters operated at a solids reten-tion time of 30 days were evaluated. Transition from 36 �C to 53.3 �C at a rate of 3 �C/day resulted in fluc-tuation of the daily gas and volatile fatty acids (VFAs) production. Steady-state was reached within 35 daysfrom the onset of temperature increase. Transitions from either 36 or 53.3 �C to 60 �C resulted in relativelystable daily gas production, but VFAs remained at very high levels (in excess of 5000 mg COD/L) and meth-ane production was lower than that of the mesophilic reactor. It was concluded that in order to achievehigh VS and COD destruction and methane production, the temperature of continuous-flow, suspendedgrowth digesters fed with mixed municipal sludge should be kept below 60 �C.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Anaerobic digestion has long been used as the main sludge sta-bilization process in municipal treatment plants (Tezel et al., 2011;

Tchobanoglous et al., 2014). Most municipal digesters operate atmesophilic temperatures (35 to 37 �C). However, thermophilicdigestion above 50 �C has been gaining popularity, primarilybecause it achieves a higher extent of pathogen reduction, result-ing in Class A biosolids, and secondarily because it exhibits fasterkinetics. The latter results in higher solids destruction and biogasproduction compared to mesophilic digestion for the same solidsretention time (SRT), or the flexibility to achieve a desired extentof solids destruction while operating at a lower SRT value. Possibledisadvantages of thermophilic digestion may include odor, reduceddewaterability of the digested sludge, and higher heating require-ment, which may not be compensated for by higher biogas produc-tion as compared to mesophilic digestion.

386 U. Tezel et al. / Bioresource Technology 170 (2014) 385–394

Conversion from mesophilic to thermophilic digestion has beenpracticed at both laboratory- and full-scale levels. Two strategiesfor the transition from mesophilic to thermophilic digestion havebeen tested: (a) one-step temperature increase in laboratory scaleupflow anaerobic sludge blanket (UASB) reactors (Fang and Lau,1996; Syutsubo et al., 1997; van Lier et al., 1992) and in continu-ous-flow stirred tank reactors (CSTR) (Bouskova et al., 2005);(b) step-wise temperature increase in CSTR (Zabranska et al.,2002; Bouskova et al., 2005; Palatsi et al., 2009). The results ofthe Bouskova et al. (2005) study, which used a SRT of 20 daysand an organic loading rate of 1.38 g VS/L-day while the reactorswere fed during the temperature transition period, showed thatthe one-step temperature increase from 37 to 55 �C resulted in sta-ble operation in 30 days, as opposed to 70 days for the step-wisetemperature increase (37, 42, 47, 51, and 55 �C). Palatsi et al.(2009) evaluated two strategies for the transition from mesophilicto thermophilic of a mixture of primary and secondary municipalsludge in CSTR laboratory reactors with a SRT and organic loadingrate ranging from 19.6 to 23.4 days and from 1.29 to 1.73 g VS/L-day, respectively. A single temperature change from 35 to 55 �Crequired about 20 days for the reactor efficiency to fully recover,but resulted in higher transient VFAs production, especially propi-onic acid, compared to a step-wise temperature increase (35, 43,50, and 55 �C), which required a longer time to complete the tem-perature transition. These researchers pointed out that the temper-ature range between 43 and 50 �C was critical in switchingmethanogenic activity from mesophilic to thermophilic. Peceset al. (2013) evaluated the response of a mesophilic anaerobicdigester fed with municipal sludge to short- and long-term tem-perature fluctuations. Transition from mesophilic to thermophilicconditions of a laboratory CSTR and exposure at 55 �C for 24 hresulted in an increase in VFAs and a decrease in gas production;reactor recovery was achieved in 22 days and required a non-feed-ing period.

A one-step temperature increase is not feasible at full-scaledigesters because of heating capacity limits. Full-scale testing forthe conversion of a mesophilic to thermophilic municipal sludgedigester was assessed by Iranpour et al. (2002) using continuousheating at an average temperature rise of 3 �C/day, increasing thedigester temperature from approximately 33 to 55 �C, while usinga variable sludge feeding rate. This study achieved stable digesteroperation in less than 30 days. Further operation of the thermo-philic digester achieved Class A biosolids and increased VS destruc-tion and gas production. Information relative to the transition frommesophilic to thermophilic digestion for municipal sludge stabil-ization is very limited, especially for CSTR digesters, under real,full-scale conditions and constraints. In addition, the potentialimpact of thermophilic digestion on digestate quality has not beensufficiently assessed.

The objective of the work presented here was to assess differentconversion strategies for the transition of municipal sludge anaer-obic digestion from mesophilic to thermophilic operation, andevaluate long-term stability and performance of thermophilicdigesters in terms of solids destruction, gas production/composi-tion, and digestate quality.

2. Methods

2.1. Sludge samples

The study was designed and conducted by taking into accountconditions and constraints at the F. Wayne Hill Water ResourcesCenter (WRC), Gwinnett County, GA, USA. This municipal wastewa-ter treatment plant uses activated sludge technology, achievingboth N and P removal. It uses mesophilic (98 �F or 36.7 �C)

anaerobic digestion for sludge stabilization and biogas production,which is converted to electricity through combined heat andpower technology. Primary and thickened waste activated sludge(TWAS) samples were collected at the F. Wayne Hill WRC. The pri-mary sludge was sequentially passed through a 5 � 5 mm squaremesh screen, a 2-mm sieve (US Standard No. 10), and finally a1.4-mm sieve (US Standard No. 14). The TWAS was not further pro-cessed. Both sludge samples were stored under refrigeration (4 �C).The following analyses were performed for both sludge samples:pH, total and volatile solids (TS, VS), total and soluble chemicaloxygen demand (tCOD and sCOD), VFAs, and ammonia. The watercontent of the two sludge samples was not affected by screening.

2.2. Ultimate sludge biodegradability

The test was performed using 160-mL glass serum bottles(120 mL liquid volume), sealed with rubber stoppers and flushedwith helium gas following previously described procedures(Tezel et al., 2006). Sludge obtained from a F. Wayne Hill WRC mes-ophilic anaerobic digester was anaerobically incubated in the lab-oratory and severed as inoculum (seed). Then, an aliquot of 75 mLof pre-digested sludge was anaerobically transferred to each bottleand 18 mL of media were then added. The media contained(in g/L): K2HPO4, 0.9; KH2PO4, 0.5; NH4Cl, 0.5; CaCl2�2H2O, 0.10;MgCl2�6H2O, 0.20; FeCl2�4H2O, 0.10; NaHCO3, 6.7. Also, 10 ml/Leach of vitamin and trace metal stock solutions were added tothe media (Beydilli and Pavlostathis, 2005). A total of five seriesin triplicate (5 � 3 = 15 bottles in total) were set up as follows.One series did not receive any sludge and served as the seed blank.Another series was amended with a mix of dextrin/peptone (800/400 mg/L) and served as a check of seed activity (reference series).Three more series were prepared with primary sludge, TWAS, anda mixture of primary sludge/TWAS, respectively. Primary sludgeand TWAS were tested at a sample VS loading equal to 3 g/L. Com-bined primary and TWAS were tested at a total VS loading of 3 g/Land a primary/TWAS TS ratio of 20/80% as practiced at the F.Wayne Hill WRC. Incubation was carried out in the dark at 35 �Cand the bottles were shaken manually once a day. Throughoutthe incubation period, total gas volume and composition (CH4

and CO2) were measured frequently. At the end of the incubation,pH, TS, VS, tCOD, sCOD, VFAs and ammonia were measured. Thebiodegradability test was carried out for 121 days, at which timeall gas production had leveled off.

2.3. Digesters set up and operation

All digesters used in this study were made of wide-mouth Pyrexreactors with a water jacket, and their temperature was controlledwith water recirculation using heated water-circulating baths. Thedigesters were housed in a 22–24 �C room and their contents weremechanically mixed at 90 rpm using a shaft magnetically coupledto an external, variable-speed electric drive. Gas produced was col-lected in graduated burettes by displacement of an acidified brinesolution (10% NaCl w/v and 2% H2SO4 v/v) and measured afterequilibration to atmospheric pressure. Gas data reported here areeither at 22 �C and 1 atm or at standard temperature and pressure(STP; 0 �C and 1 atm).

2.3.1. Mesophilic operation (36 �C)Two digesters (R1 and R2) were set up and operated at 36 �C

and a SRT of 30 days (close to the SRT at the plant). The total diges-ter volume was equal to 4 L with a liquid working volume equal to3 L. Both digesters were started with 3 L mixed liquor obtainedfrom a F. Wayne Hill WRC mesophilic anaerobic digester. The feedfor both digesters was primary/TWAS sludge mixture (20/80% onTS basis), which was kept under refrigeration (4 �C). The combined

Table 1Results of primary sludge and TWAS sample analysis.

Parameter Primary sludge TWAS

pH 4.70 6.13TS, g/kg wet sample 30.3 ± 0.1a 58.8 ± 0.2VS, g/kg wet sample 23.5 ± 0.1 41.2 ± 0.2VS/TS,% 77.6 70.1Total COD, mg/L 41970 ± 1370 59750 ± 1260Soluble COD, mg/L 3450 ± 170 2350 ± 15Ammonia, mg N/L 84 ± 5 224 ± 5Total VFAs, mg COD/L 2300 ± 43 1080 ± 24Acetic 1400 ± 17 567 ± 7Propionic 622 ± 7 297 ± 5iso-Butyric 18 ± 10 34 ± 1n-Butyric 203 ± 1 81 ± 2iso-Valeric NDb 75 ± 4n-Valeric 62 ± 1 28 ± 7iso-Caproic ND NDn-Caproic 7 ± 5 NDHeptanoic ND ND

a Mean ± standard deviation (n P 3).b ND, not detected.

U. Tezel et al. / Bioresource Technology 170 (2014) 385–394 387

sludge (30 mL) was fed to each digester once a day manually usinga plastic syringe. Throughout this phase (98 days), pH, total gasvolume and composition (CH4, CO2), and VFAs were measured fre-quently. At the end of this phase, the digesters mixed liquor wasanalyzed for pH, TS, VS, tCOD, sCOD, VFAs, ammonia, andphosphate.

2.3.2. Transition to moderate thermophilic operation (53 �C)One digester (R1) was maintained at 36 �C and operated as

described above. The temperature of digester R2 was graduallyincreased at a rate of 3 �C/day and then maintained at 53.3 �C(128 �F) (transition 1). For this temperature transition, the follow-ing considerations were taken into account: a) the heating capacityat the F. Wayne Hill WRC is such that the digesters can be heatedfrom 36.7 �C (98 �F) to 53.3 �C (128 �F) in about 5 days (i.e., at atemperature increase of about 3.3 �C/day); b) previous assessmentof possible strategies for the transition from mesophilic to thermo-philic digestion favored fast temperature increases while thedigesters were fed. Based on the above, and to be conservative,for digester R2 the temperature increase from 36 �C to 53.3 �Coccurred in 6 days (i.e., a temperature increase of about3 �C/day). During the period of temperature increase of digesterR2, sludge wasting and feeding was the same in both digesters(i.e., both R1 and R2 digesters received the same sludge loadingas described above and were maintained at a SRT of 30 days).Throughout the temperature transition of digester R2 and untilits performance was stable, pH, total gas volume and composition(CH4, CO2), and VFAs were measured daily. At the end of this phase,the digesters mixed liquor was analyzed for pH, TS, VS, tCOD, sCOD,VFAs, ammonia, and phosphate.

2.3.3. Transition to high thermophilic operation (60 �C)In order to evaluate the possibility of operating the thermophilic

digester at 60 �C, three strategies were applied while the SRT waskept at 30 days: (a) transition from 36 to 60 �C at 3 �C/day (transi-tion 2); (b) transition from 53.3 to 60 �C at 3 �C/day (transition 3);and (c) transition from 53.3 to 60 �C in four steps (55, 57, 58.5,and 60 �C) (transition 4). Corresponding reactors used for the threetemperature transitions to 60 �C are referred to as R3, R4, and R5,respectively. Throughout this phase of the study, pH, total gasvolume and composition (CH4, CO2), and VFAs were measured fre-quently. The mesophilic (36 �C) reactor was operated at 30 days andwas used as control. At the end of this phase, the digesters mixedliquor was analyzed for pH, TS, VS, tCOD, sCOD, VFAs, ammonia,and phosphate.

2.4. Microbial community analysis

Microbial community analysis related to temperature transition2, digester R3, was conducted to quantitatively assess the changein bacterial and archeal communities associated with the changein the reactor’s temperature from 36 �C to 60 �C and its perfor-mance after was maintained at 60 �C for a relatively long period(about 10 SRTs). Coprothermobacter spp., Archaea, and the aceticlas-tic methanogenic families Methanosarcinaceae and Methanosaeta-ceae were selected as targets. Both Methanosarcinaceae andMethanosaetaceae belong to the order Methanosarcinales. Methano-saetaceae are strict aceticlastic methanogens, whereas Methanosar-cinaceae utilize both acetate and H2/CO2 for methanogenesis.Coprothermobacter spp. is an important bacterial group in theanaerobic digestion of municipal sludge, as is proteolytic and reg-ulates the release of amino acids and ammonium. Quantitative PCRanalysis was performed to detect and quantify the 16S rRNA genesof the above mentioned target microorganisms. Three duplicatebiomass samples were collected from R3 at the following day ofoperation and temperature: sample A at day 0, 36 �C; sample B

at day 8 when the temperature had just reached 60 �C during thetemperature transition; and sample C at day 286 after the reactorhad been maintained at 60 �C for 10 months and its performancewas relatively stable in terms of biogas production and residualVFAs (see Section 3.3.1, below). Details on the procedures followedfor DNA extraction, preparation of qPCR standards, and qPCRassays are included in Supplementary data (Text S1).

2.5. Analytical methods

TS, VS, pH, COD, and ammonia measurements were conductedaccording to procedures described in Standard Methods (APHA,2012). Total phosphorus in the sludge feed and digester effluentwas measured following the molybdovanadate/acid persulfatedigestion method (HACH procedure 10127; HACH, Loveland, CO,USA). Phosphate measurements were conducted using ionchromatography/conductivity detection with samples filteredthrough a 0.2 lm syringe filters (Tugtas and Pavlostathis, 2007).Total gas production was measured by displacement of an acidifiedbrine solution (10% NaCl w/v and 2% H2SO4 v/v) in graduatedburettes. The gas composition and VFAs were determined by gaschromatography with thermal conductivity and flame ionizationdetection, respectively, as previously reported (Okutman Tas andPavlostathis, 2005; Misiti et al., 2013).

3. Results and discussion

3.1. Sludge characteristics and ultimate anaerobic biodegradability

The results of the analysis of the primary sludge and TWASsamples are shown in Table 1. Both samples were acidic and theVS/TS ratio was higher for the primary sludge. Significant levelsof soluble COD and VFAs were found in both samples, especiallyin the primary sludge, indicating that a degree of sludge solubiliza-tion and preacidification had taken place. Acetate and propionatewere the major VFAs.

The anaerobic biodegradability test was carried out for121 days. At the end of the incubation, all samples tested had pHvalues between 7.04 and 7.15 and VFAs were not detected. Gasproduction in all five series started without any lag and the pri-mary sludge sample had the highest gas production rate (Fig. S1;Supplementary data). After about 30 days of incubation, the gasproduction rate was relatively similar in all five series indicatingthat the gas production from this point forward was the result of

388 U. Tezel et al. / Bioresource Technology 170 (2014) 385–394

slow seed and/or sludge destruction. Table 2 summarizes theresults of the batch test for the two individual samples, as wellas the combined primary/TWAS sample. A good COD balance wasachieved indicating that methanogenesis was the main terminalelectron transfer process. The anaerobic biodegradability of TWASwas significantly lower than the primary sludge. Based on the ulti-mate biodegradability (i.e., VS destruction) of each component (i.e.,primary sludge and TWAS), and taking into account the VS concen-tration of each component in the combined primary/TWAS sampletested, the calculated VS destruction in the combined sample wasequal to 43.6%, which is comparable to the measured VS destruc-tion of 40.0% (Table 2).

3.2. Mesophilic operation and transition to moderate thermophilicoperation

Both digesters R1 and R2 were operated identically for 25 daysat a SRT of 30 days and an organic loading rate of 1.1 g VS/L-day.The performance of both R1 and R2 based on gas production, pHand VFAs concentration was identical (Fig. 1). On day 25, the tem-perature of R2 was increased at a rate of 3 �C/day and reached53.3 �C on day 31 (transition 1), whereas that of R1 remained at36 �C. The daily gas production of R2 increased almost linearlywith the increase in temperature, reached a high of 2100 mL/dayon day 29 and then decreased precipitously and remained between340 and 450 mL/day for about 8 days (Fig. 1). It is noteworthy that,while the reactor temperature was increased, a precipitousdecrease in gas production occurred when the R2 temperaturewas 48 �C. Similar to our results, Palatsi et al. (2009) observed adrastic decrease in methane production when the reactor temper-ature was between 43 and 50 �C and considered this temperaturerange as the most critical for the transition from mesophilic tothermophilic digestion of municipal sludge. The decrease in gasproduction was accompanied by a sharp increase in VFAs concen-tration (Fig. 1). The performance of R2 was stabilized at day 60, i.e.,after 35 days of operation past the initiation of temperatureincrease.

The digesters characteristics and steady-state data are shown inTable 3. Although the mean feed pH was 5.8, the R1 and R2 meanpH values were 7.1 and 7.4, respectively, and the reactors wereoperated without any alkalinity addition even during the temper-ature transition of R2 when the VFAs concentration increased over3 g COD/L. The soluble COD, VFAs, ammonia, and phosphate in R2were 5.1-, 5.4-, 1.3-, and 1.2-fold higher compared to R1. The gasproduction was 335 and 344 mL at STP/g VS added for R1 andR2, respectively. The COD destruction in both digesters was alsosimilar. In contrast to comparable total gas production and CODdestruction, VS destruction and soluble COD in R2 were higherthan in R1. Thermophilic conditions in R2 enhanced sludge

Table 2Results of the batch anaerobic ultimate biodegradability test.

Parameter Primary sludge (P

Initial primary sludge, g VS/L 3Initial TWAS, g VS/L –Initial total VS, g/L 3VS destruction,b % 56.7COD destruction,b % 68.2Methane,% of total gas 67.3Total gas produced,b mL @ STP/g VS added 598.3Methane, mL @ STP/g VS added 399.8Methane, mL @ STP/g VS destroyed 705.6COD balancec 5.4

a PS/TWAS mix, 20/80% TS basis.b Seed-corrected, corresponding to the individual component or mix of components.c (CODinitial � CODfinal � CODmethane) � 100/CODinitial.

disintegration and hydrolysis but products were partially recalci-trant and thus not fully processed to methane. The thermophilicR2 reactor achieved only 5% higher methane production comparedto R1. The specific methane production was 305 and 313 mL atSTP/g COD destroyed in R1 and R2, respectively, which are belowthe theoretical value of 350 mL methane at STP/g COD destroyed.Reduction of alternative electron acceptors such as nitrate and sul-fate results in a lower specific methane production due to electronchanneling away from methane formation. However, neithernitrate nor sulfate was detected in the feed used in this study.The overall COD balance for R1 and R2 was 4.9 and 4.1, respectively(Table 3).

The major VFAs in the feed, expressed as COD, were acetic(52%), propionic (23%), and n-butyric (12%) acids. Acetic acid wasthe major VFA in both the R1 and R2 reactors (54 and 79%, respec-tively) just before the temperature transition in R2. When the VFAsconcentration was the highest in R2 after its temperature hadreached 53.3 �C (day 36), the major VFAs component was aceticacid (72%), followed by propionic (13%). When the performanceof R2 stabilized (day 60), acetic acid was the major VFAs compo-nent (82%), followed by propionic (9%) and heptanoic acid (9%).Based on these results, although the total VFAs concentrationwas higher in R2 compared to R1, acetic acid was the predominantVFA in both reactors. As mentioned above, although the solubleCOD concentration in R2 was more than 5-fold higher comparedto R1, the VFAs in both reactors accounted for about 7% of the sol-uble COD.

3.3. Transition and operation at 60 �C

3.3.1. Transition 2The second temperature transition was from 36 to 60 �C at a

rate of 3 �C/day while the digester, designated as R3, was operatedwith a SRT of 30 days. R3 was started with 1.5 L of mixed liquorfrom the mesophilic reactor R1 (36 �C, 30 days SRT) gradually col-lected from the daily waste over 15 days, kept unfed. Then, R3 waswasted and fed while maintained at 36 �C for 14 days before thetemperature transition to 60 �C. Fig. 2 shows the performance ofR3 before and during the temperature increase, as well as whileit was maintained at 60 �C. As soon as the reactor temperatureincreased, the gas production increased from a mean value of690 mL to a maximum of 1085 mL on day 17 when the reactortemperature was at 45 �C and then precipitously decreased to370 mL by day 22 when the reactor temperature had reached60 �C. From day 35 to 62, the gas production stabilized to about200 mL, which is about one third of the gas produced by the mes-ophilic digester (R1, 36 �C) for the same reactor volume (1.5 L). TheVFAs concentration in R3 increased linearly from 77 to 3700 mgCOD/L on day 30 when the reactor temperature had been at

S) TWAS Primary sludge + TWASa

– 0.653 2.353 340.0 40.046.4 40.767.7 68.6363.7 404.8244.7 278.5611.8 696.2�0.5 �3.8

TEM

P. (o C

)

30

40

50

60

VFAs

(mg

CO

D/L

)

0

1000

2000

3000

4000

TIME (Days)

0 10 20 30 40 50 60 70 80 90 100

pH

5.56.06.57.07.58.0

GAS

(mL/

d)0

50010001500200025003000 Mesophilic (R1)

Meso/Thermophilic (R2)

C

B

Feed Mesophilic (R1)Meso/Thermophilic (R2)

A

Meso/Thermophilic (R2)Mesophilic (R1)

Meso/Thermophilic (R2)

D

Fig. 1. Temperature (A), daily gas production at 22 �C and 1 atm (B), VFAs (C) and pH (D) in the two digesters operated at 30 days SRT. On day 25, the R2 digester temperaturewas gradually increased to 53.3 �C at 3 �C/day (transition 1), whereas that of R1 remained at 36 �C. VFAs in the mesophilic digester R1 were below 25 mg COD/L at all times(error bars represent mean values ± one standard deviation; n = 3).

Table 3Feed, effluent characteristics, and performance of digesters R1 and R2 (SRT 30 d; reactor volume 3 L)a.

Parameter Feed Mesophilic (R1) Meso/thermophilic (R2)

Temperature – 36 �C 36–53.3 �CpH 5.75 ± 0.08b 7.09 ± 0.01 7.38 ± 0.02TS, g/L 46.9 ± 0.1 36.4 ± 0.1 35.4 ± 0.1VS, g/L 33.0 ± 0.1 22.3 ± 0.1 21.3 ± 01Total COD, g/L 59.7 ± 0.7 37.0 ± 1.5 36.5 ± 1.9Soluble COD, g/L 4.6 ± 0.2 0.62 ± 0.03 3.19 ± 0.10VFAs, mg COD/L 2466 ± 49 43 ± 18 231 ± 40Ammonia, mg N/L 336 ± 5 980 ± 5 1232 ± 5Phosphate, mg P/L 438 ± 5 413 ± 5 505 ± 5Total gas, mL at 22 �C/day – 1196 ± 41 1226 ± 40Methane,% – 62.5 ± 0.7 64.1 ± 1.3Methane, mL at 22 �C/day – 748 ± 29 786 ± 52VS destruction,% – 32.4 35.5COD destruction,% – 38.0 38.9Total gas, mL at STP/g VS added – 335 344Methane, mL at STP/g VS destroyed – 647 622Methane, mL at STP/g COD destroyed – 305 313COD balance, %c – 4.9 4.1

a Steady-state data (day 70 to 91).b Mean ± standard deviation (n P 5).c COD balance = (CODin � CODout � CODCH4) � 100/CODin.

U. Tezel et al. / Bioresource Technology 170 (2014) 385–394 389

60 �C for 8 days, and remained at this level for another 6 daysbefore it increased again and reached 6610 mg COD/L on day 62(Fig. 2).

Due to low gas production and VFAs accumulation at a highconcentration, R3 was kept at 60 �C unfed for a period of 44 daysand then was intermittently fed while its volume was alsoincreased first to 2 and then to 3 L. Frequent, daily feeding resumedon day 146 and continued till the end of this phase of the study(309 days of operation). During the latter period of operation whenthe reactor volume was 3 L, the mean gas production was 1125 mL/day, but varied from a low 780 to a high 1450 mL/day. Althoughthe mean gas production was close to that of the mesophilic reac-tor (R1, 36 �C, 30 days SRT) for the same reactor volume (3 L), the

VFAs in R3 remained at relatively high levels. As shown in Fig. 2,the VFAs decreased to a low value of 3900 mg COD/L during theperiod that the reactor was kept unfed, and then during the latterphase when feeding was resumed, the VFAs concentration variedbetween 3000 and 5500 mg COD/L. During the entire period ofoperation of R3, its pH remained above 7 without any alkalinityaddition with the exception of a single addition of 2 g NaHCO3

on day 86 when the pH was 6.7.Just before the reactor temperature was increased on day 14,

acetic acid was the major VFA component (62% of total VFAsCOD), followed by propionic (11%), and heptanoic acid (9%), andthe remaining VFAs were below 6.6%. On day 62, when the dailywasting and feeding was stopped converting R3 to a batch reactor,

VFAs

(mg

CO

D/L

)

02000400060008000

VOLU

ME

(L)

0

1

2

3

TEM

P. (o C

)0

1530456075

GAS

(mL)

0400800

12001600

TIME (Days)0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320

pH

5.56.06.57.07.58.08.5

Daily feedingBatch

A

B

D

E

C

Fig. 2. Reactor volume (A), temperature profile (B), gas production at 22 �C and 1 atm (C), VFAs (D), and pH (E) in digester R3 operated at 30 days SRT and transitioned from 36to 60 �C at 3 �C/day (transition 2) (error bars represent mean values ± one standard deviation; n = 3).

390 U. Tezel et al. / Bioresource Technology 170 (2014) 385–394

the major VFAs components were acetic (45%), propionic (22%),and iso-valeric acid (13%). On day 124, when intermittent wastingand feeding resumed, the total VFAs concentration decreased from6610 to about 3900 mg COD/L. The major VFAs components wereacetic (29%), propionic (44%), followed by iso-valeric acid (17%).Comparing the distribution of VFAs between day 62 and 124 showsthat during the batch operation of the R3 reactor, the acetic andiso-valeric acid concentration decreased, whereas the concentra-tion of propionic acid increased. Throughout this study, this wasthe first time that the propionic acid concentration exceeded thatof the acetic acid. By the end of this phase of the study, the aceticacid concentration had decreased while that of the propionic andiso-valeric acid had increased, each representing 12%, 46%, and21% of total VFAs COD, respectively. At the end of this phase ofthe study, the R3 effluent had the following characteristics: VS23.2 ± 1.7 g/L, soluble COD 12330 ± 180 mg/L, total VFAs4903 ± 472 mg COD/L, ammonia 1650 ± 7 mg N/L, and phosphate325 ± 10 mg P/L. The soluble COD was about 26% of the totalCOD. The VS destruction was 38.7%, which is higher than thatachieved by the mesophilic (36 �C; 32.4%) and thermophilic(53.3 �C; 35.5%) digesters, all operated at a SRT of 30 days (Table 3).It is noteworthy that Gray (Gabb) et al. (2006) reported a signifi-cant decrease in VS destruction with an increase in digestion tem-perature from 49 to 52 and 62 �C, a finding that has been observedin other previous studies according to these authors. As a result ofhigh VS destruction and taking into account that the digestersludge feed was high in TWAS, a total ammonia concentration of2184 mg N/L was recorded at 220 d of operation. Based on thereactor conditions (60 �C and pH 7.8), the un-ionized ammoniaconcentration was calculated as 540 mg N/L, which is well above100 mg ammonia-N/L, a concentration that may be inhibitory tomethanogens (Rittmann and McCarty, 2001), though ammonia

inhibition is a complex process affected by the total ammonia con-centration, pH, temperature, and C/N ratio (Angelidaki and Ahring,1994; Rajagopal et al., 2013; Borowski and Weatherley, 2013). Thehigh un-ionized ammonia concentration may be responsible forthe observed periodic increase and decrease of gas productiontowards the latter part of this phase of the study (Fig. 2). If thereactor pH was lowered to 7, for the same total ammonia concen-tration the un-ionized ammonia concentration would havedecreased to 108 mg N/L and could have resulted in a more stablereactor operation. Given the unstable gas production and accumu-lation of VFAs at a high concentration, this reactor was abandoned.The inability to establish reactor operation at 60 �C with a rela-tively low VFAs level was attributed to lack of microbial acclima-tion to this temperature in spite the relatively long SRT valueused and a relatively long period of batch operation (Fig. 2), as wellas to possible ammonia inhibition.

3.3.2. Transition 3The third temperature transition was from 53.3 to 60 �C at a

rate of 3 �C/day while the digester, designated as R4, was operatedwith a SRT of 30 days. R4 was started with 2 L of mixed liquor col-lected from the thermophilic digester R2 (53.3 �C, 30 days SRT).Fig. 3 shows the performance of R4 before and during the temper-ature increase, as well as at 60 �C over a relatively long time.

The gas production, after an initial small increase during thetemperature increase, precipitously decreased to 270 mL by day8 when the reactor temperature had reached 60 �C, increased againreaching the highest value of 720 mL by day 21, and thendecreased and fluctuated between 230 and 440 mL/day (meangas production, 325 mL/day) (Fig. 3C). The R4 mean gas productionwas about 40% of that produced by the mesophilic digester (R1,36 �C) for the same reactor volume (2 L). The VFAs in R4 increased

TEM

P. (o C

)

40

50

60

70

VFAs

(mg

CO

D/L

)

02000400060008000

10000G

AS (m

L/d)

0400800

12001600

TIME (Days)0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

pH

6.06.57.07.58.08.5

A

B

C

D

Fig. 3. Temperature (A), gas production at 22 �C and 1 atm (B), VFAs (C), and pH (D) in digester R4 operated at 30 days SRT and transitioned from 53.3 to 60 �C at 3 �C/day(reactor volume, 2 L; transition 3) (error bars represent mean values ± one standard deviation; n = 3).

U. Tezel et al. / Bioresource Technology 170 (2014) 385–394 391

linearly from about 340 to 3800 mg COD/L on day 18 when thereactor temperature had been at 60 �C for 14 days. Then, the VFAsdecreased, coinciding with the observed increase in gas productionon day 21. Past day 24, the VFAs increased gradually and then fluc-tuated between 5500 and 7900 mg COD/L till the end of this phaseof the study (Fig. 3B). During the entire period of operation of R4,its pH remained above 7 without any alkalinity addition with theexception of a single addition of 2 g NaHCO3 on day 140 whenthe pH had dropped to 6.9.

Before the reactor temperature was increased on day 4, bothacetic and propionic acid were the major VFA components (43and 46% of total VFAs COD, respectively), followed by heptanoicacid (8%), and the remaining VFAs were all very low. However,on day 8, which marked the lowest gas production after the reactortemperature had reached 60 �C, the acetic and propionic acid rep-resented 40% and 35% of the total VFAs COD, respectively. For theremaining operation period of R4, acetic acid increased at a higherrate and was always higher than propionic acid. On day 62, whenthe highest VFAs concentration was observed (7910 mg COD/L),the acetic and propionic acid represented 53% and 26% of the totalVFAs COD, showing a disproportional increase of acetic acid con-centration over that of propionic acid. By the end of this phase ofthe study, both the acetic and propionic acid concentration haddecreased, while that of the iso-valeric acid had increased, eachrepresenting 38%, 27%, and 15% of total VFAs COD, respectively.Compared to the performance of reactor R3, which was developedat the same SRT of 30 days, but starting at 36 �C, the performanceof R4 was more stable, i.e., its gas production and VFAs levels fluc-tuated less, but its COD destruction and gas production was muchlower as discussed above.

3.3.3. Transition 4In view of the fact that none of the previous temperature transi-

tions was successful in establishing reactor operation at 60 �C witha relatively low VFAs level, a fourth temperature transition from53.3 to 60 �C was tested, but instead of a continuous temperaturerise at a rate of 3 �C/day used in all previous temperature transi-tions, the temperature rise in this case was done in four steps (55,

57, 58.5, and 60 �C), while the reactor was operated with a SRT of30 days. The digester, designated as R5, was started with 2 L ofmixed liquor collected from the thermophilic reactor R2 (53.3 �C,30 days SRT). Fig. 4 shows the performance of R5 throughout itsoperation. R5 was operated at 53.3 �C for 7 days before the firsttemperature transition to 55 �C and achieved a mean gas produc-tion of 1040 mL/day during this period (Fig. 4D). With the increaseof temperature to 55 and then to 57 �C, the gas productionincreased initially and then decreased and for 50 days (day 7 today 57) fluctuated between 670 and 1060 mL/day. With furtherincrease of the temperature to 58.5 �C, a gradual decrease of gasproduction was observed which continued after the reactor tem-perature was increased to 60 �C. After the reactor liquid volumehad reached 3 L, the gas production stabilized first at about285 mL/day and towards the latter part of operation at 415 mL/day (Fig. 4C). The R5 mean gas production towards the end of itsoperation was about 35% of that produced by the mesophilic diges-ter (R1, 36 �C) for the same reactor volume (3 L). With the exceptionof a few sudden increases and decreases, the VFAs concentration inR5 increased linearly from 415 to 3900 mg COD/L corresponding toreactor temperature 53.3 and 58.5 �C, respectively. Further increaseof the reactor temperature to 60 �C resulted in an initial smalldecrease followed by an increase of the VFAs concentration whichfluctuated between 3900 and 8150 mg COD/L, stabilizing to about7000 mg COD/L towards the latter part of the R5 operation(Fig. 4D). During the entire period of operation of R5, its pHremained above 7 without any alkalinity addition with the excep-tion of two additions of 3 g NaHCO3 on day 140 and 145 whenthe pH had dropped to 6.7.

Fig. S2 shows the trend in total VFAs as well as that of the pre-dominant VFAs (i.e., acetic, propionic, and iso-valeric) as a functionof reactor R5 temperature. On day 21 when the reactor tempera-ture had been at 55 �C for 14 days, the observed increase in thetotal VFAs was predominantly due to an increase in propionic acid,which accounted for 72% of the total VFAs COD. As the reactor tem-perature increased to 57 and then to 58.5 �C, the total VFAsincreased by 6.3- and 9.4-fold, respectively, compared to the VFAsconcentration at 53.3 �C. The acetic acid concentration at 57 and

VOLU

ME

(L)

0

1

2

3

TEM

P. (o C

)50

55

60

65

VFAs

(mg

CO

D/L

)

02000400060008000

10000

GAS

(mL/

day)

0200400600800

100012001400

TIME (Days)0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

pH

5.56.06.57.07.58.08.5

A

B

C

D

E

Fig. 4. Reactor volume (A), temperature (B), gas production at 22 �C and 1 atm (C), VFAs (D), and pH (E) in digester R5 operated at 30 days SRT and transitioned from 53.3 to60 �C in four steps (55, 57, 58.5, and 60 �C; transition 4) (error bars represent mean values ± one standard deviation; n = 3).

392 U. Tezel et al. / Bioresource Technology 170 (2014) 385–394

58.5 �C remained practically unchanged, but a significant increaseof the propionic acid concentration was observed, which accountedfor between 60% and 64% of the total VFAs COD (Fig. S2). Similar toour results, Wilson et al. (2008) and Palatsi et al. (2009) observed ahigher propionic acid concentration compared to acetic acid as thedigestion temperature increased to and above 50 �C. In the presentstudy, transition and operation at 60 �C resulted in 16.5-foldincrease in the total VFAs concentration compared to that at53.3 �C. At 60 �C, a dramatic change in the VFAs distribution wasobserved, marked by a very high increase in the concentration ofall acids, except iso-caproic and n-caproic, which remained practi-cally at the levels observed at 58.5 �C, and propionic acid, whichdecreased by 36% compared to 58.5 �C. The predominant VFAs at60 �C were acetic (40%), propionic (25%), and iso-valeric (14%) acid.The performance of R5 towards the end of this phase of the studywas relatively stable, i.e., its gas production and VFAs levels had alower fluctuation, but its COD destruction and gas production werelow.

3.4. Comparison of performance at 60 �C

As mentioned above, three reactors were transitioned to andoperated at 60 �C using three different strategies, while the reactorswere operated with a SRT of 30 days. In terms of specific gas pro-duction rate, the performance of the three reactors operated at60 �C is compared to that of the mesophilic (36 �C) reactor operatedwith the same SRT and identical feed sludge and the results areshown in Fig. S3A. The thermophilic reactor R3 (transition 2) hadabout the same specific gas production rate as the mesophilic reac-tor (Fig. S3A), but as discussed above, its gas production fluctuated

over time. The specific gas production rate of reactors R4 and R5(transition 3 and 4, respectively) was significantly lower than thatof the mesophilic (36 �C) by 39% and 35%, respectively, but theirgas production fluctuated less than that of reactor R3. Similar toour results, Wilson et al. (2008) reported 31–54% lower methaneproduction in anaerobic digesters operated at 57.5 �C comparedto a mesophilic (37 �C) digester and the methane production at57.5 �C was about 23% of that achieved at 53 �C. Ahring et al.(2001) reported 18% decrease in the specific methane productionfrom the anaerobic digestion of cattle manure at SRT of 15 days at65 �C as compared to digestion at 55 �C. Digestion at 65 �C alsoresulted in a decrease of VS destruction to 22% from 28% observedat 55 �C and an increase of VFAs from below 0.3 g/L to as high as2.6 g COD/L, mostly as acetate and propionate. Degradation of pro-pionate was completely inhibited at 65 �C (Ahring et al., 2001).

All three thermophilic digesters maintained at 60 �C had highlevels of VFAs (between 4900 and 7600 mg COD/L) (Fig. S1B). Itis noteworthy that the steady-state total VFAs concentration inthe mesophilic (36 �C) and moderate thermophilic (53.3 �C) reac-tors, both operated at a SRT value of 30 days, was 43 ± 18 and231 ± 40 mg COD/L, respectively (Table 3). In terms of VFAsdistribution, reactors R4 and R5 (transition 3 and 4, respectively)had similar relative VFA components as follows (%): acetic(37.6–41.2), propionic (25.1–27.4), butyric (13.9–15.4), valeric(16.6–18.9), caproic (1.1–1.5), and heptanoic (0–0.8) acid. Incontrast, reactor R3 (transition 2) had a relatively lower acetic acid(12.1%) and a higher propionic (46%) and valeric (23.1%) acidfraction (Fig. S3C).

Aceticlastic methanogens are more sensitive than hydrogeno-trophic methanogens at temperatures encountered in thermophilic

U. Tezel et al. / Bioresource Technology 170 (2014) 385–394 393

digestion (Ahring et al., 2001), which may explain the relativelyhigh acetate levels observed in all digesters operated at 60 �C.However, enrichment of syntrophic acetate-oxidizing, non-metha-nogenic thermophilic bacteria under the dual stress of high acetateand relatively high ammonia levels (Lee and Zinder, 1988; Haoet al., 2011, 2013; Fotidis et al., 2013; Lü et al., 2013; Ho et al.,2014) in R3 over its long-term operation, as well as acclimationand proliferation of aceticlastic methanogens at 60 �C while R3was maintained for about 70 days in batch mode, may have con-tributed to the observed lower levels of acetic acid in R3 comparedto all other digesters maintained at 60 �C with daily wasting/feed-ing, operated for relatively shorter times.

3.5. Microbial community response to temperature increase

The gene numbers of the microbial community in the threesamples collected from R3 operated at temperature values from36 to 60 �C (transition 2) are presented in Fig. 5. The PCR and gelelectrophoresis results revealed that the methanogenic archaeaassociated with the family Methanosaetaceae were not detectedin any of the three samples analyzed, agreeing with the qPCR datawhich shows that the gene copy numbers of the family Methano-sarcinaceae were very close to that of the domain Archaea. Theseobservations suggest that the aceticlastic archeal community wasdominated by Methanosarcinaceae. Mladenovska and Ahring(2000) and Demirel and Scherer (2008) have reported thedominance of Methanosarcinaceae in thermophilic digesters andunstable anaerobic digesters with high VFA concentrations, respec-tively. A previous study reported that Methanosarcinaceae domi-nate in high-rate anaerobic thermophilic digesters fed withwaste activated sludge (Ho et al., 2013).

Fig. 5 shows the changes in 16S rRNA gene concentrations ofCoprothermobacter, Archaea, and Methanosarcinaceae in the threesamples collected at different operational temperatures and dura-tion of digester R3. At 36 �C, when the reactor performance wasstable in terms of gas production and had negligible residual VFAs(sample A), the 16S rRNA gene concentrations of Coprothermobact-er spp., Archaea, and Methanosarcinaceae, were 1.6 ± 0.46 � 106,

SAMPLEA B C

Gen

e C

opie

s/m

L ex

tract

ed D

NA

100

101

102

103

104

105

106

107

108

Coprothermobacter spp. Archaea Methanosarcinaceae

Fig. 5. 16S rRNA gene concentrations of Coprothermobacter spp., Archaea, andMethanosarcinaceae in three samples collected from digester R3 transitioned from36 to 60 �C (sample A and B; transition 2) and maintained at 60 �C for over10 months (sample C) (Error bars represent mean values ± one standard deviation;n = 3).

8.4 ± 0.46 � 105, and 1.3 ± 0.0017 � 105 copies/mL of extractedDNA, respectively. When the digester R3 temperature was transi-tioned from 36 �C and had just reached 60 �C (sample B), the 16SrRNA gene concentration of both Coprothermobacter spp. and Met-hanosarcinaceae decreased significantly. In contrast, the archeal16S rRNA gene concentrations remained almost unchanged(Fig. 5). This observation indicates that hydrolytic bacteria and ace-ticlastic Methanosarcinaceae were negatively affected by the tem-perature increase. Zinder et al. (1984) reported that an anaerobicdigester, fed with the organic fraction of municipal refuse at10 days retention time, maintained at 58 �C, when was dominatedby Methanosarcina sp., aceticlastic methanogenesis was maximal at58 �C and completely inhibited at 65 �C. Ho et al. (2014) reportedthe dominance of Methanosarcinaceae in anaerobic digesters at55–60 �C, but increase of temperature to 65 �C resulted in loss ofMethanosarcinaceae, accumulation of VFAs, and a decrease inmethane production. The change in bacterial and archeal popula-tions observed in the present study, as shown by the 16S rRNAgene copies, could be due to a dynamic population shift to otherbacterial species and methanogenic archaea, such as hydrogeno-trophic methanogens, as well as an increase in non-methanogenicarchaea (e.g., Crenarchaeota). Zinder et al. (1984) reported thatmethanogenesis from CO2 with a culture from an anaerobic diges-ter maintained at 58 �C was optimal at 65 �C. Ahring et al. (2001)reported that the hydrogenotrophic methanogens were more tol-erant and predominated in anaerobic CSTRs treating cattle manurewhen the operational temperature was changed from 55 to 65 �C.At relatively high pH (i.e., at or above 8.0), free ammonia couldbecome inhibitory, particularly to Methanosaeta (Demirel andScherer, 2008).

In the latter phase of the thermophilic operation when R3 wasmaintained at 60 �C for 10 months (sample C), even lower numbersof Coprothermobacter spp., Archaea and Methanosarcinaceae wereobserved (Fig. 5). Although the gas production during this latterphase was stable, the reactor had a relatively high VFAs andammonia concentration (see Sections 3.3.1 and 3.6, respectively).At this phase, the archeal community was vastly dominated byMethanosarcinaceae. Based on the qPCR data, it appears that thehydrogenotrophic methanogenic population also declined drasti-cally at 60 �C. Thus, the increase in the reactor temperature had asignificant effect on the microbial community and populationdynamics, which eventually affected performance.

3.6. Digestate quality

The phosphorus and ammonia concentrations in sludge filtratesfrom mesophilic and thermophilic digesters, all operated at SRT of30 days, are shown in Table S2. The orthophosphate phosphorus(PO4

3�-P) concentration in filtered digesters effluent samples ran-ged from 285 to 505 mg P/L, which corresponds to 15.6 and26.9% of the total sludge P, and increased as the reactor tempera-ture increased from 36 to 53.3 �C. The total ammonia concentra-tion in the digesters filtered effluent ranged from 980 to2184 mg N/L and the highest concentration was observed at60 �C, which is consistent with previous observations accordingto which the ammonia concentration increases with increasedSRT and temperature (Bivins and Novak, 2001; Gray (Gabb) et al.,2006). The un-ionized free ammonia concentration calculated bytaking into account the digester pH and temperature is also shownin Table S2. With the exception of digester R3, the un-ionizedammonia concentration values are well below 100 mg ammonia-N/L, a concentration that may be inhibitory to methanogens(Rittmann and McCarty, 2001), particularly to Methanosaeta spp.(Demirel and Scherer, 2008). In contrast, R3, operated at 60 �Cand pH 7.8, exhibited high VS destruction which resulted in thehighest total ammonia level, but also in a high un-ionized

394 U. Tezel et al. / Bioresource Technology 170 (2014) 385–394

ammonia concentration (data at 220 days of operation), whichmay have contributed to the observed periodic increase anddecrease of gas production towards the latter part of operation ofR3 (Fig. 2).

4. Conclusions

Transition of mixed municipal sludge anaerobic digestion frommesophilic to thermophilic conditions is feasible with continuousheating at a constant rate and sludge feeding of suspended growth,well mixed digesters. The strategy by which final temperature wasachieved did not make any significant difference. However, inorder to achieve stable, high performance (i.e., high VS and CODdestruction and methane production), the digester temperatureshould be kept below 60 �C. Taking into account full-scale munici-pal plant constraints (e.g., heating capacity), gradual transitionfrom mesophilic to thermophilic conditions can result in the suc-cessful conversion to a stable thermophilic sludge digestionprocess.

Acknowledgements

This work was supported by a contract from the GwinnettCounty, Department of Water Resources, Lawrenceville, GA, USAadministered through Hazen and Sawyer, P.C., Atlanta, GA, USA.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.biortech.2014.08.007.

References

Ahring, B.K., Ibrahim, A.A., Mladenovska, Z., 2001. Effect of temperature increasefrom 55 to 65 �C on performance and microbial population dynamics of ananaerobic reactor treating cattle manure. Water Res. 35 (10), 2446–2452.

American Public Health Association (APHA), 2012. Standard Methods for theExamination of Water and Wastewater, 22nd ed. American Public HealthAssociation/American Water Works Association/Water EnvironmentFederation, Washington DC, USA.

Angelidaki, I., Ahring, B.K., 1994. Anaerobic thermophilic digestion of manure atdifferent ammonia loads: effect of temperature. Water Res. 28 (3), 727–731.

Beydilli, M.I., Pavlostathis, S.G., 2005. Decolorization kinetics of the azo dye ReactiveRed 2 under methanogenic conditions: effect of long-term culture acclimation.Biodegradation 16 (2), 135–146.

Bivins, J.L., Novak, J.T., 2001. Changes in dewatering properties between thethermophilic and mesophilic stages in temperature-phased anaerobic digestionsystems. Water Environ. Res. 73 (4), 444–449.

Borowski, S., Weatherley, L., 2013. Co-digestion of solid poultry manure withmunicipal sewage sludge. Biores. Technol. 142, 345–352.

Bouskova, A., Dohanyos, M., Schmidt, J.E., Angelidaki, I., 2005. Strategies forchanging temperature from mesophilic to thermophilic conditions in anaerobicCSTR reactors treating sewage sludge. Water Res. 39 (8), 1481–1488.

Demirel, B., Scherer, P., 2008. The roles of acetotrophic and hydrogenotrophicmethanogens during anaerobic conversion of biomass to methane: a review.Rev. Environ. Sci. Biotechnol. 7 (2), 173–190.

Fang, H.H.P., Lau, I.W.C., 1996. Startup of thermophilic (55 �C) UASB reactors usingdifferent mesophilic seed sludges. Water Sci. Technol. 34 (5–6), 445–452.

Fotidis, I.A., Karakashev, D., Kotsopoulos, T.A., Martzopoulos, G.G., Angelidaki, I.,2013. Effect of ammonium and acetate on methanogenic pathway andmethanogenic community composition. FEMS Microbiol. Ecol. 83 (1), 38–48.

Gray (Gabb), D.M.D., Hake, J.M., Ghosh, S., 2006. Influence of staging, mean cellresidence time, and thermophilic temperature on the thermophilic anaerobicdigestion process. Water Environ. Res. 78 (5), 497–509.

Hao, L.-P., Lui, F., He, P.-J., Li, L., Shao, L.-M., 2011. Predominant contribution ofsyntrophic acetate oxidation to thermophilic methane formation at high acetateconcentrations. Environ. Sci. Technol. 45 (2), 508–513.

Ho, D.P., Jensen, P.D., Batstone, D.J., 2013. Methanosarcinaceae and acetate-oxidizingpathways dominate in high-rate thermophilic anaerobic digestion of waste-activated sludge. Appl. Environ. Microbiol. 79 (20), 6491–6500.

Ho, D.P., Jensen, P.D., Batstone, D.J., 2014. Effects of temperature and hydraulicretention time on acetotrophic pathways and performance in high-rate sludgedigestion. Environ. Sci. Technol. 48 (11), 6468–6476.

Iranpour, R., Oh, S., Cox, H.H.J., Shao, Y.J., Moghaddam, O., Kearney, R.J., Deshusses,M.A., Stenstrom, M.K., Ahring, B.K., 2002. Changing mesophilic wastewatersludge digestion into thermophilic operation at Terminal Island TreatmentPlant. Water Environ. Res. 74 (5), 494–507.

Lee, M.J., Zinder, S.H., 1988. Isolation and characterization of a thermophilicbacterium which oxidizes acetate in syntrophic association with a methanogenand which grows acetogenically on H2–CO2. Appl. Environ. Microbiol. 54 (1),124–129.

Lü, F., Hao, L., Guan, D., Qi, Y., Shao, L., Pinjing He, P., 2013. Synergetic stress of acidsand ammonium on the shift in the methanogenic pathways duringthermophilic anaerobic digestion of organics. Water Res. 47 (7), 2297–2306.

Misiti, T., Tezel, U., Pavlostathis, S.G., 2013. Fate and effect of naphthenic acids on oilrefinery activated sludge wastewater treatment systems. Water Res. 47 (1),449–460.

Mladenovska, Z., Ahring, B.K., 2000. Growth kinetics of Methanosarcina spp. isolatedfrom full-scale biogas plants treating animal manures. FEMS Microbiol. Ecol. 31(3), 225–229.

Okutman Tas, D., Pavlostathis, S.G., 2005. Microbial reductive transformation ofpentachloronitrobenzene under methanogenic conditions. Environ. Sci.Technol. 39 (21), 8264–8272.

Palatsi, J., Gimenez-Lorang, A., Ferrer, I., Flotats, X., 2009. Start-up strategies ofthermophilic anaerobic digestion of sewage sludge. Water Sci. Technol. 59 (9),1777–1784.

Peces, M., Astals, S., Mata-Alvarez, J., 2013. Response of a sewage sludge mesophilicanaerobic digester to short and long-term thermophilic temperaturefluctuations. Chem. Eng. J. 233, 109–116.

Rajagopal, R., Daniel, I., Massé, D.I., Singh, G., 2013. A critical review on inhibition ofanaerobic digestion process by excess ammonia. Biores. Technol. 143, 632–641.

Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology: Principles andApplications. McGraw-Hill, New York, NY, USA.

Syutsubo, K., Harada, H., Ohashi, A., Suzuki, H., 1997. An effective start-up ofthermophilic UASB reactor by seeding mesophilically-grown granular sludge.Water Sci. Technol. 36 (6–7), 391–398.

Tchobanoglous, G., Stensel, H.D., Tsuchihashi, R., Burton, F., 2014. WastewaterEngineering – Treatment and Resource Recovery. McGraw-Hill, New York, NY,USA.

Tezel, U., Pierson, J.A., Pavlostathis, S.G., 2006. Fate and effect of quaternaryammonium compounds on a mixed methanogenic culture. Water Res. 40 (19),3660–3668.

Tezel, U., Tandukar, M., Pavlostathis, S.G., 2011. Anaerobic Biotreatment ofMunicipal Sewage Sludge. In: Agathos, S. (Ed.), Environmental Biotechnologyand Safety. In: Moo-Young, M. (Editor-in-Chief), Comprehensive Biotechnology,2nd ed., vol. 6 Elsevier, Amsterdam, The Netherlands.

Tugtas, A.E., Pavlostathis, S.G., 2007. Effect of sulfide on nitrate reduction in mixedmethanogenic cultures. Biotechol. Bioeng. 97 (6), 1448–1459.

van Lier, J.B., Grolle, K.F.C., Stams, A.J.M., de Maccario, E.C., Lettinga, G., 1992. Start-up of a thermophilic upflow anaerobic sludge bed (UASB) reactor withmesophilic granular sludge. Appl. Microbiol. Biotechnol. 37 (1), 130–135.

Wilson, C.A., Murthy, S.M., Fang, Y., Novak, J.T., 2008. The effect of temperature onthe performance and stability of thermophilic anaerobic digestion. Water Sci.Technol. 57 (2), 297–304.

Zabranska, J., Dohanyos, M., Jenicek, P., Zaplatilkova, P., Kutil, J., 2002. Thecontribution of thermophilic anaerobic digestion to the stable operation ofwastewater sludge treatment. Water Sci. Technol. 46 (4–5), 447–453.

Zinder, S.H., Anguish, T., Cardwell, S.C., 1984. Effects of temperature onmethanogenesis in a thermophilic (58�C) anaerobic digestor. Appl. Environ.Microbiol. 47 (4), 808–813.