lable at ScienceDirect

Journal of Environmental Management 92 (2011) 1867e1873

Contents lists avai

Journal of Environmental Management

journal homepage: www.elsevier .com/locate/ jenvman

Performance of sequential anaerobic/aerobic digestion applied to municipalsewage sludge

M. Concetta Tomei*, Sara Rita, Giuseppe MininniWater Research Institute, C.N.R., Via Salaria km 29.300, C.P. 10, 00015 Monterotondo Stazione (RM), Italy

a r t i c l e i n f o

Article history:Received 7 July 2010Received in revised form7 February 2011Accepted 10 March 2011Available online 7 April 2011

Keywords:Sludge stabilizationSequential anaerobic-aerobic digestionVolatile solid reductionBiogas productionProteins and polysaccharides

* Corresponding author. Tel.: þ39 06 90672800; faE-mail address: tomei@irsa.cnr.it (M.C. Tomei).

0301-4797/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.jenvman.2011.03.016

a b s t r a c t

A promising alternative to conventional single phase processing, the use of sequential anaerobic-aerobicdigestion, was extensively investigated onmunicipal sewage sludge from a full scalewastewater treatmentplant. The objective of the work was to evaluate sequential digestion performance by testing the charac-teristics of the digested sludge in termsof volatile solids (VS), ChemicalOxygenDemand (COD) andnitrogenreduction, biogas production, dewaterability and the content of proteins and polysaccharides. VS removalefficienciesof 32% in theanaerobicphaseand17% in theaerobic onewereobtained, and similarCODremovalefficiencies (29% anaerobic and 21% aerobic) were also observed. The aerobic stage was also efficient innitrogen removal providing a decrease of the nitrogen content in the supernatant attributable to nitrifi-cation and simultaneous denitrification. Moreover, in the aerobic phase an additional marked removal ofproteins andpolysaccharides produced in the anaerobic phasewas achieved. The sludge dewaterabilitywasevaluated by determining the Optimal Polymer Dose (OPD) and the Capillary Suction Time (CST) anda significant positive effect due to the aerobic stagewas observed. Biogas productionwas close to the upperlimit of the range of values reported in the literature in spite of the lowanaerobic sludge retention time of 15days. Fromapreliminaryanalysis itwas found that theenergydemandof theaerobic phasewas significantlylower than the recovered energy in the anaerobic phase and the associated additional costwas negligible incomparison to the saving derived from the reduced amount of sludge to be disposed.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Improving sewage sludge management is a key objective for thedevelopment of an integrated strategy for treating domesticwastewater. In fact, treatment and disposal of sewage sludge fromwastewater treatment plants (WWTPs) accounts for up to 60%, ofthe total cost of wastewater treatment (Wei et al., 2003).Decreasing available land space, coupled with increasingly strin-gent regulations governing the design and operation of new land-fills (i.e. EU Landfill Directive 99/31), have caused the cost of siting,building, and operating new landfills to rise sharply. The increas-ingly restrictive targets for the continuous reduction of biode-gradable waste sent to landfills make land application a disposalalternative to be considered for the final destination of sewagesludge. Sewage sludge has been utilized in agricultural applicationsfor several years as it represents an alternative source of nutrientsfor plant growth and is an efficient soil conditioner (Logan andChaney, 1983). However, land application of sewage sludge is

x: þ39 06 90672787.

All rights reserved.

restricted to prevent health risks to humans and livestock due topotentially toxic components, i.e. heavy metals, pathogens, andpersistent organic pollutants and to the high amounts of solublesalts, which may negatively affect the soil properties.

Regardless of which sludge disposal alternative is employed, allcan take advantage of more effective stabilization processes, andthis justifies the increased attention paid to sludge stabilizationprocesses aimed at increasing their efficiency and reducing costs.

A promising strategy to improve the digestibility of sewagesludge is the application of a sequential process: anaerobic-aerobicdigestion. The basicmotivation behind this approach arises from thedifferent reaction environments (anaerobic and aerobic) provided toattain optimal biodegradability conditions for the different volatilesolid (VS) sludge fractions. This point is well addressed in Kumaret al. (2006a, 2006b) reporting a simple classification for thesludge constituent fractions in terms of VS degradability:

- a fraction degradable only under anaerobic conditions- a fraction degradable only under aerobic conditions- a fraction degradable both under anaerobic and aerobicconditions

- a non degradable fraction

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e18731868

It follows that conventional one stage digestion (aerobic oranaerobic) does not always provide efficient VS biodegradation.Moreover, the combination of the two digestion phases can beadvantageous in that they can alleviate some of the drawbackscharacterizing the two individual digestion methods. Aerobicdigestion is relatively simple to manage and produces biosolidsuseful for agriculture application but is characterized by highenergy consumption. In contrast, anaerobic digestion allows forenergy recovery in the form of biogas but is less stable and morecomplex to manage. Moreover, the efficiency of the digestionprocess strongly affects the next sludge treatment steps in that thedemand for polymer conditioning agents results from colloidalproteins and polysaccharides released from sludge particlesduring digestion (Novak et al., 2003). Consequently a moreeffective VS degradation process should reduce the costs of sludgeconditioning. Sequential digestion can potentially accomplish theabove mentioned weak points of conventional processing, withpotential advantages including a positive energy balance withpartial energy recovery in the anaerobic stage, a reduction inconditioning costs, and the use of the specific metabolic pathways,aerobic and/or anaerobic, able to degrade the different VS frac-tions in the sludge.

Recently multiestep digestion has received increased interestand previous studies have demonstrated the potential of thisapproach to improve the performance of conventional digestionprocessing (Kumar et al., 2006a, 2006b; Parravicini et al., 2008;Zupancic and Ros, 2008). Kumar et al. (2006a) observed that anaerobic digestion stage downstream from the anaerobic digestionunit improved the overall process performance both in terms of VSremoval efficiency and dewatering properties of the biosolids. Theyoperated the aerobic stage at SRT values in the range of 3e9 daysand found that three days of additional aerobic digestion resulted in20% additional VS removal beyond that achieved by the anaerobicstage while the dewaterability of the digested sludge increased withthe aerobic SRT. Parravicini et al. (2008), in a full scale plant, ach-ieved additional VS removal in the range of 16% with post-aeration(SRT w 6d) of anaerobically digested sludge and found the relativeincrease of operational costs associated to the additional aerobicdigestion phase to be negligible. Another positive effect of thesequential anaerobic-aerobic digestion is the ammonia nitrogenremoval in the supernatant stream that can significantly reduce thenitrogen load recycled to the WWTP as observed by Zupancic andRos (2008) with efficiency of ammonium nitrogen removal up to85% in the aerobic stage.

In this paper the results of an extended investigation intosequential digestion applied tomunicipal sewage sludge of a full scaleurban wastewater treatment plant are reported. The objective of thepaper is to verify the applicability of sequential digestion to a realsludgeby testing the characteristics of theproduced sludge in termsofVS, Chemical Oxygen Demand (COD) reduction, biogas production,dewaterability properties (by measuring Optimal Polymer Dose(OPD), Capillary Suction Time (CST), proteins and carbohydratescontent), and coliform abatement. Moreover, to verify the contribu-tion of the supernatant from the digestion to the nitrogen load to theWWTP, TKN was followed in the sequential digestion process. Finallya preliminary energy and cost analysis was performed by comparingthe sequential digestion to conventional anaerobic digestion.

2. Materials and methods

2.1. Sludge

Primary and secondary sludges utilized in this study wereprovided by the Rome North wastewater treatment plant. The plantis a conventional activated sludge system including screening,

primary clarification and secondary treatment, and serves about700,000 P.E. It is operated with a relatively high sludge age (20 d)and it is characterized by an unusual dilution of incoming sewage.The influent COD average value is 200 mg/L that is quite low incomparison to the range of COD values of 210e740mg/L reported inHenze et al. (1987) for settledwastewater in EU countries and to theCOD values of 500 and 650 mg/L reported in Gray (1992) for rawwastewater in USA and UK respectively.

Secondary sludge was obtained for each feed step from therecycle stream, then thickened for 18 h, while the anaerobic inoc-ulum was taken from the full scale digester of the plant fed withprimary and secondary sludge. Primary sludge, due to practicalreasons of accessibility, was obtained once a week, and was storedat 4 �C until used. The mixed sludge sample utilized in the exper-iments was prepared by mixing primary and secondary sludge inthe ratio 3/2 on a VS basis. This ratio has been chosen on the basis ofthe data reported in Vesilind and Spinosa (2001) and is represen-tative of the primary and secondary sludge production observed inItalian WWTPs.

2.2. Reactors

Lab scale reactors utilized in this study were 7.4 L cylindricalglass vessels. The reactors were operated in series, with the sludgebeing fed to the anaerobic reactor once per day and an equivalentvolume of digested sludge extracted from the anaerobic reactor andfed to the following aerobic reactor. Both of the reactors wereequipped with mechanical stirrers fitted with helicoidal blades.

The first reactor operated under anaerobic conditions and wasequipped with a thermostatic jacket and a control device keepingthe temperature at 37 � 0.5 �C. At the beginning of the start upphase it was filled with the inoculum and sewage sludge (ratioinoculum/sludge 1:1). The working volume was 7 L and the SludgeRetention Time (SRT) was controlled at 15 d.

The second reactor, whichwas operated under aerobic conditions,has a working volume of 4.5 L, and air was supplied by a compressorable to maintain the concentration of dissolved oxygen (DO) at levelsof w3 mg/L. The air flow rate, in order to reduce additional energyconsumption, was kept at the minimum value necessary to maintainthe dissolved oxygen in the reactor at the prefixed value. Such anoperating strategy prevented ammonia stripping. The reactor wasoperated at room temperature with an SRT of 12 days.

2.3. Analysis

Regular sample collection and analysis were initiated one weekafter start up. Feed, anaerobic digested and aerobic digested sludgewere analyzed for VS, COD, CST, OPD, total soluble nitrogen,proteins and polysaccharides, and total and fecal coliforms.Analytical methods and devices are reported in the following.

Volatile solids - Volatile solid concentration was measuredaccording to Standard Methods (APHA, 1998 2540E).

Chemical Oxygen Demand - COD Cell Tests (MERCK- referring toEPA 410.4 method), based on potassium dichromate oxidation andspectrophotometric determination (Spectroquant Nova30), wereemployed.

Total soluble nitrogen - Total soluble nitrogen was measured onfiltered samples (0.45 mm). Nitrogen Cell Test (MERCK) was used.This test is based on the transformation of organic and inorganicnitrogen compounds into nitrate according to Koroleff’s method bytreatment with an oxidizing agent in a thermoreactor. In a solutionacidified with sulphuric and phosphoric acid, this nitrate reactswith 2,6-dimethylphenol to form 4-nitro-2,6-dimethylphenol thatis determined photometrically.

0.00

0.50

1.00

1.50

2.00

2.50

30 d 60 d 90 d 120 d 150 d 180 d 210 d 240 d

VS

%

feed anaerobically digested aerobically digested

Fig. 1. VS concentration during the entire period of operation.

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e1873 1869

Capillary Suction Time and Optimal Polymer Dose - CST wasdetermined by a Triton Electronics apparatus, according to theStandardMethod procedure (APHA,1998 2710G). OPD is the dosageof polymer for sludge conditioning, which corresponds to thelowest value of CST. The sludgewas conditioned with Praestol 2540(Stockhausen GmbH, Germany); stock solution of this polymer wasprepared at a concentration of 1 g/L. Increasing doses of this solu-tion (from 0.5 to 2.5 mL) were added to 30 mL of the sample sludgeand mixed for 5 min. The CST was then determined as described inthe previous paragraph.

Proteins and Polysaccharides - Samples were centrifuged for10 min at 4000 rpm and the supernatant was passed througha 0.45 mm filter and analysed for biopolymer concentration. Proteinconcentration was determined according to the spectroscopicBradford method (Bradford, 1976), using a wavelength of 595 nm.Polysaccharide concentration was evaluated by the Dubois method(Dubois et al., 1956), based on the reaction of the sample withphenol and sulphuric acid. Absorbance of the treated sample wasmeasured at a wavelength of 490 nm.

Total and fecal coliforms - The number of total coliforms and fecalcoliforms in each sludge sample was determined according to themultiple tube fermentation procedure, using Hach-Lange kits.

Methane - Methane in the biogas was determined by a gascro-matograph PERKIN ELMER AutoSystem equipped with a Carboxen1000 (Supelco) column and a TCD detector. Sample volume was50 mL, transport gas was Helium (3 bar) and operating Tempera-tures were 180 �C for the oven,150 �C for the injector and 250 �C forthe detector.

2.4. Biogas detection device

The flow rate of biogas produced by the anaerobic reactor wasmeasured by a volumetric counter using a closedwater displacementsystemwith electrical contacts and with an electromagnetic valve todischarge the produced biogas to the atmosphere (Mata-Alvarezet al., 1986). The measurement device was controlled by a Program-mable Logic Controller that also provides the recording of signals.

3. Results and discussion

The SRT is a key parameter in determining the performance ofthe anaerobic digestion process and different criteria can beconsidered in its choice depending on the process objectives. Theoptimum SRT range of values suggested by Dohányos andZábranská (2001) is 12e18 d while in previous studies onsequential digestion (Kumar et al., 2006a; Novak et al., 2011) theSRT of the anaerobic digester was in the range of 10e15 d formesophilic and 15e20 d for thermophilic digestion. The interme-diate 15 d SRT value utilized in this study was chosen taking intoaccount these data with the objectives of ensuring good perfor-mance of anaerobic digestion with a reduced reactor volume.

Concerning the aerobic digester, previous studies on sequentialanaerobic-aerobic digestion showed that SRTs values of 3e6 days inthe aerobic stagewere enough to have a significant improvement inVS degradation (Kumar et al., 2006a) and biosolids dewatering(Subramanian et al., 2007) but these low SRT values are not suitablefor nitrification so a more conservative value of 12 days wasconsidered in this study.

Reactors were operated in semi-continuous mode for approxi-mately 8 months in order to have an extended work period, whichcould represent real plant operation. The start up phase, until stableperformance in the two reactors was observed, was quite rapid(�15 d), and this may be consistent with the fact that in both anaer-obic and aerobic reactors the biomass inoculum was already accli-matised. After this start up period, the performance of the system

varied as a consequence of the variability of the fed sludge, but theimprovement of the digestion performance, gained with the addi-tional aerobic stage, was maintained for the entire operation period.

3.1. VS removal

Fig. 1 shows the VS concentration in the influent and in theeffluent from anaerobic and aerobic digestion. A marked variationof the VS load is observed over the experimental period.

For better legibility the VS data are reported in the form ofa histogramwith a reference interval of 30 days. Average values andrelated standard deviations are reported in the graph. The removalefficiencies, as discussed below, refer to the entire operation periodof 240 days.

Reasonably stable performance was achieved in the anaerobicreactor with an average VS removal efficiency, of 32 � 5%. Theobserved variation can be attributed to the variability of the feed VSconcentration, which was 1.6 � 0.26 (expressed as %w/w). In thesubsequent aerobic stage an additional VS removal of 17 � 5% wasachieved, and therefore the global VS removal efficiencywas of 44%.

The observed efficiency of the anaerobic stage is in the range ofvalues (27e71%) reported in a survey of 30 anaerobic sludge digestersof municipalWWTPs (Speece, 1988) and in Bhattacharya et al. (1996)who found percent VS reductions in the range of 26e50% inconventional anaerobic digestion of mixed sludge of three WWTPs.

In any case the anaerobic removal of VS in our experiments is not“high” as an absolute value but this digestion efficiency is good inrelation to the operating conditions of the WWTP. The low influentCOD combined with the high sludge age (w20d) give a secondarysludge partially digested from thewater line. In addition the primarysludge was sampled from the bottom of the thickener that wasoperated at a sludge volume ratio (volume of the sludge blanket/volume of the thickened sludge removed daily) > 20 d. Also in thiscase the high retention time of the sludge in the thickener couldcause a partial pre-digestion of the sludge.

Additional VS removal in the aerobic stage is comparable tovalues reported in Parravicini et al. (2008) (16%) and Kumar et al.(2006a) (20%).

Considering the poor digestibility characteristics of the fedsludge, the additional removal efficiency gained in the aerobic stagecan be considered as a significant improvement of the entiredigestion performance.

3.2. COD removal

The same form of representation was chosen for total COD;concentration profiles are shown in Fig. 2 for a period of 210 days.The average removal efficiency in the anaerobic phase was 29 � 6%while an additional average removal in the aerobic phase of 21�3%

05

10152025303540

40 d 100 d 150 d 210 d

gC

OD

/L

feed anaerobically digested aerobically digested

Fig. 2. COD concentration during the entire period of operation.

Table 1Concentration of biopolymers in raw and digested sludge samples.

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e18731870

can also be observed. The parallel data analysis of VS and COD isuseful to define the evolution and the mechanisms of biodegra-dation of the soluble and particulate organic matter in the subse-quent digestion phases.

A comparison between the COD and VS patterns shows a lowerCOD removal efficiency in the anaerobic phase relative to the VS.The opposite is the case for the aerobic stage inwhich the efficiencyof total COD removal is higher than for VS removal. This findingcould be explained by considering the role of soluble COD. Total CODconsists of a particulate fraction (the VS) and a soluble fraction; inthe anaerobic phase there is net soluble COD production due to thehydrolysis of the particulate matter, that can account for theincrease of soluble COD in the effluent and the lower COD removalefficiency. The soluble COD produced in the anaerobic phase ispartially removed in the subsequent aerobic phase thus resulting inhigher total COD removal when compared to VS removal. SolubleCOD pattern values over a period of 20 days confirmed thishypothesis with an average increase of 35% in the anaerobic phasefollowed by a 42% average decrease in the aerobic one.

3.3. Nitrogen removal

According to Zupancic and Ros (2008) the supernatant fromanaerobic sludge digestion is characterized by a high nitrogencontent and can represent a significant fraction (up to 50%) of thenitrogen load of a wastewater treatment plant. The presence of anaerobic phase in which nitrification and/or (depending on theoxygen concentration level) simultaneous nitrification-denitrifica-tion can take place, is potentially able to reduce this load. Duringthe experimental campaign, the fate of total soluble nitrogen in thesupernatant from the digesters was followed over a period of 2weeks (6 samples) and the results are reported in Fig. 3.

The marked decrease detected in the aerobic phase (51 � 8%) ispresumably attributable to the effect of a simultaneous nitrifica-tion-denitrification process. In fact, the DO concentration of 3 mg/Lmaintained in the aerobic reactor is lower than the limit (�4 mg/L)

050

100150200250300350400

80th d 83rd d 86th d 89th d 92nd d 94th d

mg

N/L

feed anaerobic digester aerobic digester

Fig. 3. Total soluble nitrogen concentration in the feed and in the supernatant fromthe digesters over a period of two weeks (from the 80th to the 94th day).

required to ensure favourable conditions for nitrification (Zupancicand Ros, 2008). The consequence may be the incomplete oxygen-ation of the biomass with a consequent presence of anoxic zones atmicroscopic level due to the partial penetration of the oxygeninside the bioflocs. In such anoxic conditions nitrate instead ofoxygen can be utilized as an electron acceptor.

For practical application of sequential digestion, this point isquite important in relation to the design criteria of the aerobicstage. In the literature (Kumar et al., 2006a, 2006b) low SRTs aresuggested for the aerobic phase that can be considered to be a fin-ishing step to complete the removal process initiated in theanaerobic one. Instead, if it is desired to reduce the nitrate loadrecycled to the WWTP with the supernatant from the digester, theoptimal SRT should be high enough to ensure efficient nitrificationand simultaneous denitrification. This condition is critical also forthe dissolved oxygen level considering that DO has to be highenough to ensure the required nitrification efficiency and at thesame time be compatible with simultaneous denitrification.

3.4. Biopolymers (proteins and polysaccharides)

The protein and polysaccharide concentrations in the raw anddigested sludge were periodically measured, and the data areshown in Table 1.

In both cases, an increase in the anaerobic phase followed bya decrease in the subsequent aerobic one is observed. Proteins andcarbohydrates are associated with the soluble and colloidal fractionof COD that increases in the anaerobic stage as a consequence of thehydrolytic process of organic matter and this COD is efficientlyremoved during aerobic digestion.

The observed protein pattern is in agreement with the results ofthe experiments on anaerobic and aerobic digestion of Novak et al.(2003) and Kumar et al. (2006a): in anaerobic digestion thereduction of iron resulted in the release of large quantities ofproteins into solution, that are available for subsequent aerobicdegradation.

With regard to the polysaccharides, as reported in Novak et al.(2003) in an aerobic environment the degradation of calcium andmagnesium proteins results in a consequent polysaccharidesrelease into solution that will be removed depending on the aerobicretention time. The effect of aerobic SRT on polysaccharide degra-dation was investigated by Kumar et al. (2006a): in sequentialanaerobic mesophilic digestion (15 d SRT) followed by aerobicdigestion, with a 3 day aerobic SRT, they observed that poly-saccharides increase in the effluent while a significant removal(37%) was achieved with 9 days aerobic SRT. We worked with

Day Proteins (mg/L) Polysaccharides (mg/L)

Feed Anaerobicallydigested

Aerobicallydigested

Feed Anaerobicallydigested

Aerobicallydigested

12 30.0 55.7 31.1 59.2 148.3 71.720 30.0 57.2 20.0 59.2 135.8 51.123 24.3 34.3 14.3 62.5 153.3 77.035 28.3 37.8 12.1 73.3 85.9 75.068 23.8 54.4 21.0 73.3 85.8 71.284 23.8 31.2 13.4 85.8 134.2 91.0100 23.0 32.7 26.8 58.3 137.5 83.3112 19.1 36.7 20.8 74.2 101.7 85.8131 21.7 38.7 11.9 52.5 100.8 78.4154 28.28 44.6 15.0 52.6 124.2 93.3159 35.72 50.0 20.5 74.2 140.0 39.2Mean 26.2 43.0 18.8 65.9 122.5 74.3S.D. 4.7 9.7 6.2 10.8 24.6 16.3

Table 3Optimal polymer dose in three tests referring to three different periods of theexperimental campaign (TS ¼ Total Solids).

Test Feed (kgpol/tonTS) Anaerobically digested(kgpol/tonTS)

Aerobically digested(kgpol/tonTS)

1 1.10 2.28 0.572 2.22 3.13 2.693 2.01 3.24 1.72

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e1873 1871

anaerobic mesophilic digestion (15 d SRT) followed by an aerobicphase at a higher SRT of 12 days, that should be high enough toprovide degradation of the released polysaccharides, which wasobserved.

It is worth noting that the content of colloidal proteins andpolysaccharides is a key parameter determining the demand forpolymer conditioning agents so must be considered in the routinecharacterization of sludge in order to optimize the sequence oftreatment operations.

35

40a

3.5. Dewaterability assessment

In order to provide a first estimation of dewatering character-istics, the CST was periodically measured in samples from the feed,anaerobically digested sludge and aerobically digested sludge, andthe results are reported in Table 2.

The sequential digestion causes an increase of the CST in the firstanaerobic stage which is appreciably reduced in the aerobic stage.Sludge dewaterability was also assessed by determining optimalpolymer dose, and the results are shown in Table 3, while in Fig. 4a typical CST profile detected during the OPD test (3) is reported.

It is evident from both CST and OPD data the positive effect ofthe aerobic phase on the sludge dewaterability overcomes thenegative effect (increase of CST and OPD) observed in the anaerobicphase.

The beneficial effect of the aerobic stage on sludge dewater-ability was also referred to by Kumar et al. (2006a) for mixed sludge(1:1 by weight) with a substantial reduction in the CST, and OPDafter sequential mesophilic- aerobic digestion at aerobic SRTs aslow as 3 days.

47.7

25

30

0 20 40 60 80 100

Praestol (mg/L)

CS

T (

s)

62.6

100

120

140

160

180

200

220

0 20 40 60 80 100

Praestol (mg/L)

CS

T (s)

200

b

c

3.6. Biogas

Produced biogas was continuously monitored in all the experi-ments. The results in terms of daily and specific production(referred to the unit of destroyed VS) are reported in Fig. 5 andFig. 6. After a short lag phase (about 10 days), a regular increase isobserved. The average specific biogas production is 0.84 � 0.09 m3/(kg/VS destroyed), a value that is within the range of 0.19e1.6 m3/(kg/VS destroyed) reported in the literature (Speece, 1988;Bolzonella et al., 2005) for mesophilic digestion of sewage sludge.Thus, the 15 days SRT time adopted in our experiments is sufficientto have satisfactory biogas production that suggests acceptableperformance of the anaerobic digestion process.

Biogas data were also correlated with the added VS givinga specific production of 0.27 � 0.03 m3/(kg VS added). Also in thiscase the observed values are in the reported literature range formixed sludge 0.14e0.91 m3/(kg VS added) (Speece, 1988). It has tobe pointed out that the specific biogas production referred to theadded VS, even if it is a significant parameter to express the processefficiency, is affected by the intrinsic variability characterizing the

Table 2CST values (sec) in raw and digested sludge samples.

Day Feed Anaerobically digested Aerobically digested

30 72.3 176.8 81.554 48.6 163.7 84.068 48.6 124.5 75.293 43.8 103.3 89.8103 54.1 119.2 69.5112 48.7 121.6 66.5121 29.5 105.7 87.2Mean 49.4 130.7 79.1S. D. 12.7 28.4 8.9

sludge from different plants so it is less significant as generalparameter to quantify the process kinetics.

The methane fraction of the produced biogas was measuredover a one week period and was found to be equal to 66 � 3%,which is consistent with literature data (Bou�sková et al., 2005).

3.7. Total and fecal coliforms

Preliminary tests were also performed to monitor the fate in thedigestion stages of total and fecal coliforms, which are conventionalparameters used as indicators of pathogens. Although this param-eter is of interest for sludge application in agriculture, the presentEU regulations (Directive 86/278/EEC) on the use of sewage sludgein agriculture do not recommend a specific limit value of patho-genic microorganisms however, the “Working document on

32.3

100

120

140

160

180

0 20 40 60 80 100

Praestol (mg/L)

CS

T (s)

Fig. 4. CST profile vs Praestol dose (Test 3). (a) feed, (b) anaerobically digested sludge,(c) aerobically digested sludge. The optimal Praestol dose (referring to the samplevolume) is reported in the graph.

Table 4Total and fecal coliforms expressed as MPN/g suspended solids in the subsequentdigestion stages.

Parameter Feed (MPN/gSS) Anaerobically digested(MPN/gSS)

Aerobically digested(MPN/gSS)

Total coliforms 1.1$106 3.1$105 1.1 105

Fecal coliforms 4.2$105 9.4$104 2.1$104

0

500

1000

1500

2000

2500

3000

3500

0 50 100 150

time (days)

mL

/d

Fig. 5. Specific biogas production on daily base.

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e18731872

sludge” (that is the basis of the new regulation) excludes landapplication for sewage sludge not hygienized by specific advancedtreatments.

Results of these preliminary tests are summarized in Table 4. Asignificant abatement is observed in the anaerobic stage of aboutone order of magnitude for both total and fecal coliforms. Anadditional abatement, 64% for the total and 78% for the fecal coli-forms, is gained in the subsequent aerobic stage. These resultsprovide the first positive evaluation of sequential digestion in termsof coliform removal even if they need to be integrated with morespecific microbial tests (beyond the scope of this paper) to havea more complete picture of the potential achievable degree ofhygienization.

3.8. Energy balance and cost analysis

To evaluate the proposed digestion scheme also in terms ofenergy consumption, a preliminary evaluation of the energybalance was performed comparing energy production in theanaerobic stage (conversion of methane into electric energy) to theenergy demand in the aerobic phase.

The following assumptions were considered:

� energy consumption in the aerobic digestion 1 kWh/kg VSdestroyed (Mininni et al., 1985);

� biogas production in anaerobic digestion 0.84 Nm3/kgVSdestroyed;

� methane presence in biogas: 66% on volumetric basis;� lower calorific value of the methane: 8400 kcal/Nm3;� efficiency in electric energy production using a thermal cycle:30%.

This results in the amount of the recovered energy with thebiogas to be 1.65 kWh/kg VS destroyed.

According to generally adopted criteria, the energy balance isreferred to 1 kg TS fed to the digestion by assuming a ratioVS/TS ¼ 0.75. It was assumed that a VS removal efficiency of 32%

0

0.20.4

0.6

0.8

11.2

1.4

0 50 100 150time (days)

m3

/kg

VS

Fig. 6. Specific biogas production referring to the destroyed VS.

and 17% in the anaerobic and aerobic phase respectively wereachieved according to the average data obtained in this study.Consequently the energy production in the anaerobic phase is of0.39 kWh/kg TS and there is a demand of 0.087 kWh/kg TS in theaerobic phase. This results in an energy requirement in the aerobicphase that is significantly lower (22%) than the energy productionin the anaerobic one.

In addition to evaluate and quantify the advantages derivedfrom the additional VS removal in the aerobic phase, a cost analysiswas performed. The cost analysis also refers to 1 kgTS fed to thedigestion. The cost data represent an average value for the Italiansituation. An assumed cost of electric energy required in the aerobicphase of 0.12 €/kWh and a price of the sold energy of 0.20 €/kWhwere used, typical of the situation in Italy. The price of the soldenergy is derived by the added environmental value of renewableenergy generated by biomass that is “green electricity generation”as recognized in many EU countries.

The calculated value of the produced energy is 0.078 €/kgTS andthe cost associated with the demand in the aerobic phase is0.0104 €/kgTS. Therefore, in terms of costs the difference is evenmore relevant being the value of the produced “green” energyhigher than the value of the energy supplied by the distribution line.

For a more complete cost evaluation an important basis ofcomparison between single anaerobic digestion and the sequentialone is the saving derived from the reduction of the amount of solidsto be disposed of as a consequence of the additional TS removal inthe aerobic phase.

A simplified approach was considered by assuming a disposalcost of 0.1 €/kg sludge and final sludge concentration of 20% and25% TS for the conventional anaerobic digestion and the sequentialone justified by the better dewaterability properties of the digestedsludge observed in our experiments. This results in a net cost saving(evaluated taking into account the energy cost for the aerobicdigestion) of 0.10 €/kgTS that could be relevant considering that, forinstance, in a plant serving 500.000 inhabitants the total saving isw920.000 €/year.

4. Conclusions

An extensive investigation over a period of 8 months, wascarried out to evaluate the performance of the sequential anaer-obic-aerobic digestion of sewage sludge. In addition to the classicalefficiency parameters of VS, COD and nitrogen removal and biogasproduction, CST and OPD for the dewaterability assessment werealso determined. The fate of biopolymers (proteins and poly-saccharides) was also tracked to provide additional information onthe degradation mechanisms of the different sludge fractions.

The main conclusions of the study can be summarized asfollows:

� sequential anaerobic-aerobic digestion provides effective VSremoval efficiency (32� 5% in the anaerobic phase and 17� 5%in the aerobic one) corresponding to an average global removalefficiency referred to the fed sludge of 44%;

� COD removal efficiencies were 29 � 6% and 21 � 3% in theanaerobic and aerobic stage respectively (average globalremoval efficiency referred to the fed sludge of 44%);

M.C. Tomei et al. / Journal of Environmental Management 92 (2011) 1867e1873 1873

� total soluble nitrogen in the supernatant increases in theanaerobic phase and decreases after the aerobic stage (due tosimultaneous nitrification-denitrification) thus reducing thenitrogen load recycled to the WWTP;

� in the aerobic phase a marked removal of proteins and poly-saccharides was observed;

� the final aerobic stage has a beneficial effect on sludge dew-aterability by recovering the negative effect of the anaerobicdigestion as shown by the CST and OPD pattern;

� biogas production (0.84 � 0.09 m3/(kg/VS destroyed)), iswithin the range of literature values reported for mesophilicdigestion of sewage sludge;

� the value of the produced energy (0.078 €/kgTS) is significantlyhigher than the cost associated with the demand in the aerobicphase (0.0104 €/kgTS) while the net cost saving, taking intoaccount the cost for disposal is 0.10 €/kgTS.

Finally it is worth noting that anaerobic-aerobic sequentialdigestion also has an engineering advantage: the additional option,with respect to conventional one-phase digestion, of individuallyoptimizing the SRTs in the anaerobic and aerobic phases dependingon the required characteristics of the stabilized sludge.

Acknowledgements

Thanks are due to Prof. Andrew J. Daugulis (Queen’s University,Kingston, Ontario) for the critical analysis of the paper and for thevaluable suggestions and advices provided.

References

APHA, 1998. Standard Methods for the Examination of Water and Wastewater,twentieth ed. (Washington DC, USA).

Bhattacharya, S.K., Madura, R.L., Walling, D.A., Farrel, G.B., 1996. Volatile solidsreduction in two-phase and conventional anaerobic sludge digestion. WaterResearch 30, 1041e1048.

Bolzonella, D., Pavan, P., Battistoni, P., Cecchi, F., 2005. Mesophilic anaerobicdigestion of waste activated sludge: influence of the solid retention time in thewastewater treatment process. Process Biochemistry 40, 1453e1460.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of micro-gram quantities of protein utilizing the principle of protein-dye binding.Analytical Biochemistry 72, 248e254.

Bou�sková, A., Dohányos, M., Schmidt, J.E., Angelidaki, I., 2005. Strategies forchanging temperature from mesophilic tothermophilic conditions in anaerobicCSTR reactors treating sewage sludge. Water Research 39, 1481e1488.

Dohányos, M., Zábranská, J., 2001. Anaerobic digestion. In: Spinosa, L., Vesilind, P.A.(Eds.), Sludge into Biosolids e Processing, Disposal, Utilization. IWA Publishing,London, U.K., pp. 223e241.

Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetricmethods for determination of sugars and related substances. AnalyticalChemistry 28, 350e356.

Gray, N.F., 1992. How nature deals with waste. Chap. 1. In: Biology of WastewaterTreatment. Oxford University Press, U.K., p. 828.

Henze, M., Grady Jr., C.P.L., Gujer, W., Marais, G.v.R., Matsuo, T., 1987. Activatedsludge model No 1. In: IAWPRC Scientific and Technical Reports No 1, IAWPRC.IWA Publishing, London, p. 38.

Kumar, N., Novak, J.T., Murthy, S., 2006a. Effect of secondary aerobic digestion onproperties of anaerobic digested biosolids. In: Water Environmental Federation79th Annual Technical Exhibition and Conference, Dallas, pp. 6806e6829.

Kumar, N., Novak, J.T., Murthy S., 2006b. Sequential anaerobiceaerobic digestion forenhanced volatile solids reduction and nitrogen removal. In: Water Environ-mental Federation Residuals and Biosolids Management Conference, Cincinnati,OH, March 12e14.

Logan, T.J., Chaney, R.L., 1983. Utilization of municipal waste water and sludge onland metals. In: Proceedings of the Workshop on Utilization of Municipal WasteWater and Sludge on Land. University of California, Riverside, pp. 235e326.

Mata-Alvarez, J., Mtz-Viturtia, A., Torres, R., 1986. A simple device to measure biogasproduction in laboratory scale digesters. Biotechnology Letters 8, 719e720.

Mininni, G., Di Pinto, A.C., Mendicelli, A., Santori, M., 1985. Analisi tecnico-eco-nomica del trattamento dei fanghi derivanti dalla depurazione delle acque discarico urbane. Quaderno dell’Istituto di Ricerca Sulle Acque N. 70, IRSA CNR,Roma, pp. 89.

Novak, J.T., Banjade, S., Murthy, S.N., 2011. Combined anaerobic aerobic digestion forincreased solids reduction and nitrogen removal. Water Research 45, 618e624.

Novak, J.T., Sadler, M.S., Murthy, S.N., 2003. Mechanism of floc destruction duringanaerobic and aerobic digestion and the effect on conditioning and dewateringof biosolids. Water Research 37, 3136e3144.

Parravicini, V., Svardal, K., Hornek, R., Kroiss, H., 2008. Aeration of anaerobicallydigested sewage sludge for COD and nitrogen removal: optimization at large-scale. Water Science and Technology 57, 257e264.

Speece, R.E., 1988. A survey of municipal sludge digesters and diagnostic activityassays. Water Research 22, 365e372.

Subramanian, S., Kumar, N., Murthy, S.N., Novak, J.T., 2007. Effect of anaerobicdigestion and anaerobic/aerobic digestion processes on sludge dewatering.Journal of Residual Science and Technology 4, 17e23.

Vesilind, P.A., Spinosa, L., 2001. Production and regulations. In: Spinosa, L.,Vesilind, P.A. (Eds.), Sludge into Biosolids e Processing, Disposal, Utilization.IWA Publishing, London, pp. 3e18.

Wei, Y., Van Houten, R.T., Borger, A.R., Eikelboom, D.H., Fan, Y., 2003. Minimizationof excess sludge production for biological wastewater treatment. WaterResearch 37, 4453e4467.

Zupancic, G.D., Ros, M., 2008. Aerobic and two-stage anaerobiceaerobic sludgedigestion with pure oxygen and air aeration. Bioresource Technology 99,100e109.