Available online at www.sciencedirect.com

Bioresource Technology 99 (2008) 6105–6111

Ecological clarification of cheese whey prior to anaerobicdigestion in upflow anaerobic filter

H. Gannoun, E. Khelifi, H. Bouallagui, Y. Touhami, M. Hamdi *

Laboratory of Microbial Ecology and Technology, Department of Biological and Chemical Engineering,

National Institute of Applied Sciences and Technology, B.P. 676, 1080 Tunis, Tunisia

Received 9 May 2007; received in revised form 6 December 2007; accepted 12 December 2007Available online 31 January 2008

Abstract

Anaerobic digestion of cheese whey wastewaters (CW) was investigated in a system consisting of an ecological pretreatment followedby upflow anaerobic filter (UAF). The pretreatment was conducted to solve the inhibition problems during anaerobic treatment of CWcaused by the amounts of fats, proteins and carbohydrates and to avoid the major problems of clogging in the reactor. The optimizedecological pretreatment of diluted CW induce removal yields of 50% of chemical oxygen demand (COD) and 60% of total suspendedsolids (TSS) after acidification by Lactobacillus paracasei at 32 �C during 20 h and neutralization with lime. The pretreated CW was usedto feed UAF (35 �C). The effects of organic loading rate (OLR) and hydraulic retention time (HRT) on the pretreated CW anaerobicdegradation were examined. The average total COD removals achieved was 80–90%. The performance of the reactor was depressedby increasing the COD concentration to 20 g/l (OLR = 4 g COD/l d) and the COD removal efficiency was reduced to 72%. Significantmethane yield (280 l/kg COD removal) was obtained at an HRT of 2 days.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Cheese whey; Lactobacillus paracasei; Acidification; Precipitation; Upflow anaerobic filter

1. Introduction

The dairy industry, like most other agro-industries, gen-erates residues of which whey is the most important waste-waters produced with extremely high organic content.Cheese whey (CW), a by-product of the dairy industry, itcontains a significant amount of carbohydrates (4–5%),mainly lactose, proteins not exceeding 1%, fats at about0.4–0.5%, lactic acid less than 1%, salts that may rangefrom 1% up to 3% (Gelegenis et al., 2007). Several possibil-ities have been assayed for whey exploitation over the last50 years. Many useful food products such as protein con-centrates, yeast, lactose, lactic acid, and various feed sup-plements can be produced from whey (Cristiani-Urbinaet al., 2000; Perle et al., 1995). Nevertheless, cheese produc-ing units usually do not proceed with investments for

0960-8524/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.biortech.2007.12.037

* Corresponding author. Tel.: +216 71703627; fax: +216 71704329.E-mail address: moktar.hamdi@insat.rnu.tn (M. Hamdi).

recovery of the valuable constituents contained in wheyand so a large number of dairies in dispose of their waste,especially cheese whey, into the environment in enormousquantities. This causes serious pollution problems sincewhey has a high heavy organic pollutant with high chemi-cal oxygen demands (COD) (60–80 g/l) (Mockaitis et al.,2006). 97.7% of total COD of the wastewater wasaccounted for by lactose, lactate, protein and fat (Hwangand Hansen, 1998). Extremely high organic content ofthe CW wastewater renders the application of aerobictreatment inefficient, mainly due to the cost of oxygen sup-plementation, as well as the generation of higher sludgequantities and odours (Gavala et al., 1999b). In manydeveloping countries, which are confronting rapidlyincreasing water pollutions problems, anaerobic digestionbecomes even more favourable and promising (Agrawalet al., 1997; Shigeki and Hideki, 1999).

In Tunisia, anaerobic digestion of several industrialwastes has been carried out including tuna processing

6106 H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111

liquid effluent (Achour et al., 2000), the abattoir wastewa-ter (Mabrouk and Hamdi, 2001; Gannoun et al., 2007),olive mill wastewater (Hamdi, 1996) and fruit and vegeta-ble waste (Bouallagui et al., 2003). Currently, the whey pro-duction in Tunisia is estimated at 35.000 ton/year. Duringthe last few decades, this production has increased veryrapidly with the development of the dairy industry. Anaer-obic digestion is essentially viable method for treating highorganic content wastewater coming from cheese-producingplants (Erguder et al., 2001). However, in spite of wastereduction and energy potential, anaerobic digestion is notwidespread in the dairy industry. This is largely due tothe problem of slow reaction rates which require longhydraulic retention time (HRT) and poor process stability.Malaspina et al. (1995) stated that CW a difficult substrateto treat (especially in highly loaded reactors) due to itshigh organic content and its low bicarbonate alkalinity(50 meq/l). Various types of anaerobic reactors have beenused in laboratories to treat anaerobically CW. Patelet al., 1995 examined the application of upflow fixed filmanaerobic reactor with high COD in the influent (70 g/l)and achieved a removal of up to 81%. A laboratory UASBreactor was reported for influent wastewater concentra-tions between 12 and 60 g COD/l. COD removal rates var-ied between 85% and 99%, at an HRT of 6 days and in anOLR range of from 2 to 7.3 g COD/l d. In addition toanaerobic filters and upflow anaerobic sludge blanket(UASB) reactors, hybrid digesters and anaerobic sequenc-ing batch reactors (ASBR) are also employed for treatingdairy effluents. An ASBR system was reported to providesoluble COD and BOD5 removal rates of 62% and 75%,respectively, at an HRT of 6 h (Banik and Dague, 1997).High COD removal was obtained up to organic load of19.78 g COD/l d with anaerobic digestion of cheese wheyin two-phase reactor followed by a membrane filtrationsystem. However, the flux decline because the formationand compaction of a cake layer on the membrane surfacecaused by the particulate matter inside the porous mem-brane (Saddoud et al., 2007).

The dilution of cheese whey or the combination of twotreatment methods, biological and chemical, is anotherpossible option for reducing the instability and low effi-ciency problems caused by its high organic content. Physi-cal–chemical treatments allow the partial removal of theorganic load by protein and fat precipitations with differentchemical compounds such as aluminium sulphate, ferricchloride and ferrous sulphide (Karpati et al., 1995; Ruston,1993). However, since the reagent costs are high and thesoluble COD removal is poor in physical–chemical treat-ment processes, biological processes are usually preferred(Vidal et al., 2000). Therefore, one way of improving theperformance of digesters treating wastewaters with highcontent especially the high-rate anaerobic systems, suchas UASB reactors or upflow anaerobic filters is to promotean adequate pretreatment of the substrate.

The study was conducted to optimize particulate matterremoval by an ecological clarification and to assess the

feasibility of anaerobic digestion of clarified cheese wheyusing an upflow anaerobic filter fed with various OLR atdifferent COD inlet and HRT.

2. Methods

2.1. Cheese whey clarification by lactic bacteria acidification

and lime addition

The fresh CW was collected from a local ‘‘TunisianCheese Factory’’ (EL Jedaida, Tunisia) which used tradi-tional technologies for cheese manufacture. It was analyzedand stored at �20 �C to avoid the acidification and thechemical composition modification of cheese whey. Lacto-

bacillus paracasei ssp. paracasei1, which was isolated previ-ously from fresh CW, was maintained on MRS agar (ManRogosa and Shapman).

Acidification of CW with L. paracasei was carried inErlenmeyer flasks with a total volume of 150 ml under sta-tic conditions at 32 �C. At the end of the fermentation, thepH was adjusted to neutral value with lime addition whichinduces precipitation of some compounds especially pro-teins and grease. A factorial design experiment (Deshayes,1980) was used to study the effect of three factors (temper-ature of CW fermentation, lime dosage and initial CODconcentration) on the COD and TSS removals efficiency.In Table 1, the eight rows correspond to the eight experi-ments that should be carried and the three columns corre-spond to the three factors studied. For each factor high (+)and low (�) levels were tested. The effect of each factor onresponse (COD and TSS removal) was evaluated by thedetermination of Ci coefficients which, are calculated asfollows:

Ci ¼RX i � Y i

8ð1Þ

where Yi is the experimental value obtained (COD and TSSremoval), and Xi the level of the factor (+ or �) in the i

experiment (1–8). The chemical composition of cheesewhey before and after ecological pretreatment is shown inTable 2.

2.2. Combined system for CW treatment

Acidification of CW with L. paracasei was carried out in2 l batch bioreactor at 32 �C. At the end of the fermentation,the pH was adjusted to the neutral value with lime addition,used as a coagulant. A continuously stirred tank reactor(CSTR) with variable working volume was used to feedthe UAF with a biologically pretreated CW. The suspendedsolids were removed by sedimentation in settler. Mixing wasassured by the continuous rotation of the magnetic stirrer.The anaerobic digestion of the pretreated CW was con-ducted in an UAF reactor consisting of glass column of60 cm in height and 10 cm in diameter. The total volumeof the reactor is 2 l. The digester was packed with the Flo-cor (U3L3, porosity 95%, specific surface 230 m2 m�3) as a

Table 1Factorial design experiment (k = 3, n = 8)

Temperature of lacticfermentation (�C)

COD inlet (g/l) Lime dosage after lacticfermentation (g/l)

COD removal (%) TSS removal (%)

Level Xi

– 25 55 0.6+ 32 5.5 (diluted) 1.8

Exp. no.

1 � � � 22 322 + � � 37 403 � + � 26 294 + + � 32 385 � � + 19.25 276 + � + 28 317 � + + 37.5 458 + + + 50 62

Ci (COD) +5.28 +4.91 +2.22Ci (TSS) +4.75 +5.5 +3.25

Table 2Cheese whey wastewater characteristics before and after ecologicalpretreatment

Parameters Raw CW Pretreated CW

pH 4.46 ± 0.3 7.2 ± 0.25Conductivity (ms/cm) 7.6 ± 0.15 5.6 ± 0.12BOD5 (g O2/l) 40 ± 2.55 20 ± 1.87COD (g/l) 60 ± 10 25 ± 5TS (g/l) 59 ± 0.5 42 ± 0.5TSS (g/l) 1.5 ± 0.23 0.8 ± 0.1Proteins (g/kg) 125 ± 2 80 ± 1Fats (%) 0.9 ± 0.5 0.6 ± 0.2

H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111 6107

support for the growth of microorganisms. The inoculumwas obtained from an active biogas digester of fruit and veg-etable waste treatment (Bouallagui et al., 2003). The diges-ter was loaded with pretreated CW and operated at theoptimal mesophilic temperature range (35 ± 1 �C) underdifferent ORL, obtained by varying the COD inlet and theHRT. The system was fed by a peristaltic pump connectedto a programmable timer.

0

2

4

6

8

0 6 10 12 14 16 18 20Time (hours)

pH

0

1

2

3

4

OD

600

42 8

Fig. 1. Effect of COD concentration on the bacterial growth (OD600): 5 g/l(N), 10 g/l (j), 15 g/l (d); and on the pH evolution: 5 g/l (D), 10 g/l (s)and 15 g/l (h) during lactic acidification of CW by L. paracasei 1 at 32 �C.

2.3. Technical analysis

The analysis of the samples taken from the effluents ofthe digesters was carried out when steady-state conditionswere established. COD was estimated using the methoddescribed by Knechtel (1978). Total solids (TS), total sus-pended solids (TSS), fats, biological oxygen demand(BOD), total nitrogen, fats and volatile fatty acids (VFA)were determined according to the procedure listed in stan-dards methods for the examination of water and wastewa-ter (American Public Health Association, 1992). Theproteins content were determined according to the methodof Bradford (1976). The biogas produced was collecteddaily in plastic bags at room temperature. The total volumewas later determined with a wet gas meter and time to time

the methane content was estimated using an ORSATapparatus.

3. Results and discussion

3.1. Optimisation of treatment of CW by lactic acidification

and lime neutralization

The pretreatment step was based on L. paracasei growthon CW, lactose fermentation into lactic acid and precipita-tion of organic matter after lime addition. Fig. 1 shows thatthe kinetics of L. paracasei growth and the decrease of pHdepend on the initial COD. pH is a crucial indicator of fer-mentation progress and its drops occurs due to the metab-olism of sugars by lactic acid bacteria mainly to lactic acid(Adams, 1990). Hydrophobic and electrostatic interactionsbetween caseins and divalent ions especially Ca2+ is pro-nounced at low pH (Walstra, 1990). Tango and Ghaly(1999) demonstrated the positive effect of temperatureand pH on lactic acid production from CW using a Lacto-

bacillus strain under batch conditions. The lowest TSS andCOD removal were obtained with experiment 1 while thehighest TSS and COD removal were obtained with experi-ment 8. Indeed, the calculated coefficient based on theexperimental design, showed that all factors have a positive

5

6

7

8

9

0 10 20 30 40 50 70

pH

COD1inlet COD2 inlet COD3inlet

60

70

80

90

100

0 10 20 30 40 50 60 70 80

CO

D re

mov

al (%

)

0

0,5

1

1,5

0 10 20 30 40 50 60 70 80

P (l/

l.d)

0

100

200

300

400

0 10 20 30 40 50 60 70Time (days)

Met

hane

yie

ld (l

/Kg

CO

D

rem

oval

)

60 80

COD4inlet

80

a

b

c

d

Fig. 2. Effect of loading rates by varying the inlet COD on pH variation ofthe effluent (h) (a), the COD removal (j) (b), methane productivity (�)(c) and methane yield (d) (d) during anaerobic digestion of pretreated CWin UAF (HRT = 5 d).

6108 H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111

effect on COD and TSS removal. The temperature anddilution factor exhibited positive effect on the kinetics ofacidification by lactic fermentation. The precipitation ofproteins by lime addition is efficient when lactic fermenta-tion given the lowest pH. In fact, the comparison betweenthe runs 2 and 8, and then between the runs 7 and 8 showedthat the addition of lime improved COD and TSS removalsonly when lactic fermentation was carried out with dilutedCW at 32 �C. High concentrations of organic material andTSS in the raw CW (50–60 g/l and 1.3 g/l, respectively)were reduced by precipitation of proteins and grease afterlactic acidification and lime neutralization (Table 2). Theprotein fraction of the TSS obtained after clarification varybetween 40% and 60%. In the light of these results, theoptimized conditions established for efficient ecologicalpretreatment are: lactic acidification of diluted CW usingL. paracasei at 32 �C and neutralization with lime.

The ecological pretreatment generates a clarified CWsuitable for anaerobic digestion and biosolid residues.The rich protein and grease solid by-products can be pro-cessed into valuable products by the dairy industry ordirectly used as animal feed. Recent studies have discussedseveral serious drawbacks of using the alum salts, forexample Alzheimer’s disease and other related problemsassociated with residual aluminium in treated waters(Ozacar and Ayhan, 2003).

3.2. Upflow anaerobic filter performances

The start-up of a digester is always a critical stepbecause of the risk of strong acidification. According tothe study of Rajeshwari et al. (2000), the treatment ofcheese whey wastewaters by anaerobic degradation is con-strained by the drop in pH that inhibits further conversionof acids to methane. This makes the entire region acidic,ultimately resulting in the failure of the reactor. This is acommon problem encountered with cheese whey. With eco-logical pretreatment of CW, the start-up occurred in 4weeks without any specific problems. The ecological pre-treatment reduced the pollution content and improvedthe BOD/COD ratio from 0.5 to 0.7 making CW wastewa-ter more suitable for anaerobic treatment. The steady-stateof the UAF reactor is mainly due to the liberation ofammonia resulting from degradation of residual proteinin pretreated CW and a synergic interaction between theacidogenic and the methanogenic bacteria. The same phe-nomenon was observed with the anaerobic co-digestionof abattoir and olive mill wastewater (Gannoun et al.,2007).

The effects of OLR on the UAF performances were eval-uated in the first time by changing the feed concentration(5, 10, 15 and 20 g COD/l) and in the second by varyingthe HRT from 4 to 1 day. The effects of varying feed con-centration on the UAF reactor performance are reported inFig. 2. The pH was monitored continuously in the reactorbecause it is a sensitive parameter used to determine fer-menter stability (Fig. 2a). Under optimal conditions, pH

remained at its neutral value, which is favourable to theactivity of methanogenic bacteria. However, under thehighest OLR, the pH decreased rapidly and the conversionof substrate to biogas was reduced due to the inhibition ofmethanogenic bacteria, caused by VFA accumulation andpH decrease. The maximum COD removal efficiency of95% was reached at 15 g/l of COD corresponding to anOLR of 3 g COD/l d (Fig. 2b). When increasing theCOD concentration to 20 g/l (OLR = 4 g COD/l d), theCOD removal efficiency was reduced to 63% after 80 days.The UAF performance is highly sensitive to the quality ofthe feed of dairy wastes especially, the COD value. Indeed,the yield and kinetics of the biological reactions involved inanaerobic digestion are strongly dependent upon wastecomposition (Archana et al., 1999). The complete degrada-tion of the wastewaters clearly depends on the hydrolysisrate of each different compound. According to Vidal

H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111 6109

et al. (2000), the most easily biodegradable substrates, aremainly sugars and some proteins, whereas the second onecorresponded to fat degradation. In the case of fats, theconversion rate is limited either by the conversion of thelong chain fatty acids, or by the physical processes of dis-solution and mass transfer of these acids.

The methane yield obtained from anaerobic digestion ofpretreated cheese whey wastewater was improved byincreasing the OLR from 1 g COD/l d to 3 g COD/l d.The highest biogas production rate of 3.2 l/d was obtainedwith OLR of 3 g COD/l d. However, there was a decreasein the conversion of the substrate into biogas when theorganic loading rate increased from 3 g COD/l d to4 g COD/l d (Fig. 2c). The VFAs in the effluent at anOLR of 4 g COD/l d (780 mg/l) were found to be aroundtwo times the VFAs in the effluent at OLR of 3 g COD/l d (330 mg/l). Therefore, changes in VFA concentrationcan be in response to variations in organic loading ratesor levels of toxicants. Despite, the fact that the loadingrates used in this work were similar to those tested in theliterature (Gavala et al., 1999a); the methanogenesis wasinhibited due to a pH decrease from 7.2 to 5.9 because ofthe rapid degradation of sugars to volatile fatty acids.More recently, Yu and Fang (2001) used an upflow reactorto treat dairy wastewater, and they reported that, at pH6.5, acetate, propionate and butyrate represented 79% ofthe total VFA. The volume of methane produced per kgof COd removal decreased in proportion as the feed con-centration was increased (Fig. 2). In fact, the highest con-version of wastewater to methane was obtained at lowOLR of 1 COD/l d.

The profiles obtained for pH, COD removal, biogasproductivity and methane yield parameters for each HRTare presented in Table 3. As the retention time wasreduced, the COD removal became lower and a gradualincrease in the amount of biogas productivity rates (l/l d)was observed. Varying the HRT from 4 to 2 days had noeffect on the fermentation stability and pH remained con-stant. However, the methane production yield decreasedfrom 280 to 96 l CH4/kg removal COD by decreasing theHRT to 1 day. With high OLR, UAF is more affected byCOD than by CW dilution factor at low HRT. Erguderet al. (2001) stated that HRT values as low as 2–3 dayscan be used for the anaerobic treatment of cheese whey,with a COD removal efficiency of 95%. In a more recentstudy, the effects of HRT between 12 and 24 h on anaerobic

Table 3pH, COD removal, biogas productivity and methane yield obtained duri(CODinlet = 5 g/l)

Parameter Operating conditions

HRT = 4 days

pH 7.51 ± 0.5COD removal (%) 90.2 ± 0.5Biogas productivity (1 l/d) 0.47 ± 0.05Methane yield (l CH4/kg COD removed) 89 ± 0.3

acidogenesis of dairy wastewater was investigated, using alaboratory-scale continuous flow completely mixed anaero-bic reactor with solids recycle (Demirel and Yenigun,2004). The acid production gradually increased propor-tionally to the OLR, with decrease in HRT.

There are many laboratory- and pilot-scale studies in theliterature on the anaerobic treatment of CW (Table 4). Ingeneral, it is difficult to compare systems operated in differ-ent laboratories, meaning that the anaerobic sludge charac-teristics might be quite different. Patel et al. (1995)investigated anaerobic digestion of high-strength cheesewhey with COD of 70 g/l, using an upflow fixed film reac-tor with various support media, obtaining a maximumCOD removal of 81%. A pilot scale upflow anaerobic filterreactor treating dairy wastewaters provided more than 85%COD and 90% BOD removal, at an OLR of 6 kg COD/m3 d (Ince, 1998). More recently, an upflow anaerobic filterreactor treating dairy and cheese wastewater yielded anaverage of 80% COD removal in an OLR range up to21 kg COD/(m3 d) (Ince et al., 2000). Biological treatmentof a cheese-producing wastewater by a laboratory-scaleUASB reactor was also reported for influent wastewaterconcentrations between 12 and 60 g COD/l (Gavala et al.,1999a). COD removal varied between 85% and 99%, atan HRT of 6 days and in an OLR range from 2 to7.3 g COD/l d), while removal rates were around 81% inan HRT range between 30 and 40 days. Dairy wastewaterscontaining high concentrations of fat and grease were trea-ted by an UASB reactor (Cammarota et al., 2001). CODremoval was reported to be about 90%. Recently, feasibil-ity of using UASB reactors for dairy wastewater treatmentwas explored by operating two types of UASB reactors(Ramasamy et al., 2004). The reactors were operated atan HRT range between 3 and 12 h, and on loadings rang-ing from 2.4 to 13.5 kg COD/m3 d. At 3 h, maximum CODreduction ranged between 95.6% and 96.3%, while at 12HRT reductions were around 92–90%, for both reactors.Anaerobic sequencing batch reactor have been investigatedfor treating cheese whey, results have been promisingshowing the real potentials of this system as an alternativeto continuous flow (Zaiat et al., 2001; Ratusznei et al.,2003).

Furthermore, new generations of more advanced anaer-obic reactor systems have also been developed; two-phaseanaerobic treatment systems are becomes essential forimproving biogas production rate and methane yield.

ng anaerobic digestion of pretreated CW at different HRT in UAF

HRT = 3 days HRT = 2 days HRT = 1 day

7.46 ± 0.45 7.33 ± 0.6 5.7 ± 0.885.7 ± 1 77.2 ± 3 72 ± 50.72 ± 0.12 1.15 ± 0.28 0.52 ± 0.08143 ± 12 280 ± 30 110 ± 0.5

Table 4Performances of different types of anaerobic reactors used in the treatment of CW

Reactor type Cheese whey HRT (d) COD inlet (g/l) OLR (g COD/ l d) COD removal References

DSFFR Raw CW 5 13 2.6 88 Malaspina et al. (1995)UAF Raw CW 2 70 – 81 Patel et al. (1995)TSUAD Raw CW 20 69.6 3.5 39.5 Ghaly (1996)Hybrid Raw CW – – 2.82 97 Strydom et al. (1997)CSTR + UAF Synthetic CW 5 7–23 95 Ince (1998)UASB Deproteinated CW 6 – 2–7.3 85–99 Gavala et al. (1999a)SFR and MFR (UAF) Raw CW – – 20 90–92 Punal et al. (1999)UAF Dairy wastewater 0.5 – Up to 21 80 Ince et al. (2000)Two-stage UASB

reactorsRaw CW 2–3 42.7–55.1 – 95–97 Erguder et al. (2001)

Hybrid Acidic CW 2 – Up to 11 95 Calli and Yukselen(2002)

Hybrid Synthetic dairywastewater

4.7–1.7 0.82–6.11 90–97 Ramasamy et al. (2004)

ASBR Raw CW – 0.5–4 – 90 Mockaitis et al. (2006)TSMAMD Raw CW 4 68.6 19.7 98.8 Saddoud et al. (2007)UAF Clarified CW 2–5 5–20 4 98 This work

DSFFR: downflow stationary fixed-bed reactor; TSUAD: two-stage unmixed anaerobic digester; UAF: anaerobic upflow filter. UASB: upflow anaerobicsludge blanket; CSTR: continuous stirred tank reactor; SFR: single-fed reactor; MFR multi-fed reactor; ASBR: anaerobic sequencing batch reactor;TSMAMD: two-stage mixed anaerobic membrane digester.

6110 H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111

Laboratory-scale single-fed (SFR) and multi-fed (MFR)upflow anaerobic filters treating cheese whey were operatedat OLRs above 20 kg COD/m3 d (Punal et al., 1999).Anaerobic treatment of a high-strength acidic cheese wheyby a laboratory-scale upflow hybrid reactor resulted inremoval efficiencies of more than 95%, at 2 days of HRTand up to an OLR of about 11 g COD/l d (Calli andYukselen, 2002). Using a laboratory-scale mesophilictwo-phase system for the cheese factory effluent, 97%COD removal was achieved at an OLR of 2.82 kg COD/m3 d (Strydom et al., 1997).

4. Conclusions

The pretreatment step was based on L. paracasei growthon CW, lactose fermentation into lactic acid and precipita-tion of organic matter after lime addition. The optimizedconditions established for efficient ecological pretreatmentare: lactic acidification of diluted CW using L. paracasei

at 32 �C and neutralization with lime. During the ecologi-cal pretreatment step, maximal removal yields of 50% ofCOD and 60% of TSS were obtained after acidificationof diluted CW by L. paracasei at 32 �C during 20 h andneutralization with lime. The protein faction of the TSSobtained after pretreatment vary between 40% and 60%.Operation of the reactor with clarified CW at COD concen-tration of 15 g COD/l, corresponding to an organic load of3 g COD/l d is recommended, because these conditionsensure the highest levels of COD removal (95%) and meth-anisation (280–380 l/kg COD removal). Hence, increasingOLR to 4 g COD/l d promotes increased accumulation ofVFA and decreased methane production. This indicatesthe need to reduce the OLR by reducing the feed or dilutingthe influent. At constant COD concentration of 5 g COD/land at HRT varying from 4 to 2 days, the levels of COD

removal were 90.2% and 77.2%, respectively. However,decreasing the HRT to 1 day causes a partial inhibitionof the process by VFA accumulation and the methane pro-duction yield decreased to 110 l CH4/kg removal COD.The combined system achieved up to 90% of COD removaland had stable operations throughout the whole experi-mental period. The results of this work demonstrate thatUAF can be used for clarified cheese whey wastewatertreatment and energy recovery with rapid start-up times.However, the stability and the efficiency of the UAF canbe increased if longer HRT values were used.

References

Achour, M., Khelifi, O., Bouazizi, I., Hamdi, M., 2000. Design of anintegrated bioprocess for the treatment of tuna processing liquideffluents. Process Biochem. 35, 1013–1017.

Adams, M.R., 1990. Topical aspects of fermented foods. Trends Food Sci.Technol. 1, 140–144.

Agrawal, L.K., Harada, H., Okui, H., 1997. Treatment of dilute in UASBreactor at a moderate temperature: performance aspects. J. Ferment.Bioeng. 83, 179–184.

American Public Health Association (APHA), American Water WorksAssociation (AWWA), Water Environment Federation (WEF),1992. Standard Methods for the Examination of Water andWastewater, 18th ed. American Public Health Association, Wash-ington, DC, USA.

Archana, S., Unni, B.G., Devendra, S.H., 1999. A novel fed-batchdigestion system for biomethanation of plant biomasses. J. Biosci.Bioeng. 87, 678–682.

Banik, G.C., Dague, R.R., 1997. ASBR treatment of low strengthindustrial wastewater at psychrophilic temperatures. Water Sci.Technol. 36, 337–344.

Bouallagui, H., BenCheikh, R., Marouani, L., Hamdi, M., 2003.Mesophilic biogas production from fruit and vegetable waste intubular digester. Bioresour. Technol. 86, 85–89.

Bradford, M.M., 1976. A rapid and sensitive method for quantitation ofmicrogram quantities of protein utilizing the principle of protein–dyebinding. Anal. Biochem. 72, 248–254.

H. Gannoun et al. / Bioresource Technology 99 (2008) 6105–6111 6111

Calli, B., Yukselen, M.A., 2002. Anaerobic treatment by a hybrid reactor.Environ. Eng. Sci. 19, 143–150.

Cammarota, M.C., Teixeira, G.A., Freire, D.M.G., 2001. Enzymatic pre-hydrolysis and anaerobic degradation of wastewaters with high fatcontents. Biotechnol. Lett. 23, 1591–1595.

Cristiani-Urbina, E., Netzahuatl-Munoz, A.R., Manriquez-Rojas, F.J.,Juarez-Ramirez, C., Ruiz-Ordaz, N., Galindez-Mayer, J., 2000. Batchand fed-batch cultures for the treatment of whey with mixed yeastcultures. Process Biochem. 35, 649–657.

Demirel, B., Yenigun, O., 2004. Anaerobic acidogenesis of dairy waste-water: the effects of variations in hydraulic retention time with no pHcontrol. J. Chem. Technol. Biotechnol. 79, 755–760.

Deshayes, M.P.C., 1980. Utilisation de modeles mathematiques pourl’optimisation en fermentation. Bull. Soc. Chim. Franc�aise 1 (2), 24–34.

Erguder, T.H., Tezel, U., Guven, E., Demirer, G.N., 2001. Anaerobicbiotransformation and methane generation potential of cheese whey inbatch and UASB reactors. Waste Manage. 21, 643–650.

Gannoun, H., Ben Othman, N., Bouallagui, H., Hamdi, M., 2007.Mesophilic and thermophilic anaerobic co-digestion of olive millwastewaters and abattoir wastewaters in an upflow anaerobic filter.Ind. Eng. Chem. Res. 46, 6737–6743.

Gavala, H.N., Kopsinis, H., Skiadas, I.V., Stamatelatou, K., Lyberatos,G., 1999a. Treatment of dairy wastewater using an upflow anaerobicsludge blanket reactor. J. Agric. Eng. Res. 73, 59–63.

Gavala, H.N., Skiadas, I.V., Lyberatos, G., 1999b. On the performance ofa centralised digestion facility receiving seasonal agroindustrial waste-waters. Water Sci. Technol. 40, 339–346.

Gelegenis, J., Georgakakis, D., Angelidaki, I., Mavris, V., 2007. Optimi-zation of biogas production by co-digesting whey with diluted poultrymanure. Renew. Energy 32, 2147–2160.

Ghaly, A.E., 1996. A comparative study of anaerobic digestion of acidcheese whey and dairy manure in a two-stage reactor. Bioresour.Technol. 55, 61–72.

Hamdi, M., 1996. Anaerobic digestion of olive mill wastewater, a review.Process Biochem. 31, 105–110.

Hwang, S., Hansen, C.L., 1998. Characterization of and bioproduction ofshort-chain organic acids from mixed dairy-processing wastewater.Trans. Am. Soc. Agric. Eng. 41, 795–802.

Ince, O., 1998. Performance of a two-phase anaerobic digestion systemwhen treating dairy wastewater. Water Res. 32, 2707–2713.

Ince, O., Ince, B.K., Donnelly, T., 2000. Attachment, strength andperformance of a porous media in an upflow anaerobic filter treatingdairy wastewater. Water Sci. Technol. 41, 261–270.

Karpati, I., Bencze, L., Borszerki, J., 1995. New process for physico-chemical pretreatment of dairy effluents with agricultural use of sludgeproduced. Water Sci. Technol. 22 (9), 93–100.

Knechtel, R.J., 1978. A more economical method for the determination ofchemical oxygen demand. Water Pollut. Control Fed. 50, 25–29.

Mabrouk, N., Hamdi, M., 2001. Anaerobic digestion of poultry abattoirwastewater. In: Proceedings of the Ninth World Congress Anaerobicwastewater, IWA Antwerpen, Belgium, pp. 2–6.

Malaspina, S.F.L., Cellamare, C.M., Tilche, A., 1995. Cheese whey andcheese factory wastewater treatment with a biological anaerobic–aerobic process. Water Sci. Technol. 32, 59–72.

Mockaitis, G., Ratusznei, S.M., Rodrigues, Jose A.D., Zaiat, M., Foresti,E., 2006. Anaerobic whey treatment by a stirred sequencing batchreactor (ASBR): effects of organic loading and supplemented alkalin-ity. J. Environ. Manage. 79, 198–206.

Ozacar, M., Ayhan, S.I., 2003. Evaluation of tannin biopolymer as acoagulant aid for coagulation of colloidal particles. Colloids Surf. A229, 85–96.

Patel, P., Desai, M., Madamwar, D., 1995. Biomethanation of cheesewhey using anaerobic upflow fixed film reactor. J. Ferment. Bioeng. 79(4), 398–399.

Perle, M., Kimchie, Sh., Shelef, G., 1995. Some biochemical aspects of theanaerobic degradation of dairy wastewater. Water Res. 29 (6), 1549–1554.

Punal, A., Mendez-Pampin, R.J., Lema, J.M., 1999. Characterization andcomparison of biomasses from single and multi fed upflow anaerobicfilters. Bioresour. Technol. 68, 293–300.

Rajeshwari, K.V., Balakrishnan, M., Kansal, A., Lata, K., Kishore,V.V.N., 2000. State-of the art of anaerobic digestion technology forindustrial wastewater treatment. Renew. Sustain. Energy Rev. 4, 135–156.

Ramasamy, E.V., Gajalakshmi, S., Sanjeevi, R., Jithesh, M.N., Abbasi,S.A., 2004. Feasibility studies on the treatment of dairy wastewaterswith upflow anaerobic sludge blanket reactors. Bioresour. Technol. 93,209–212.

Ratusznei, S.M., Rodrigues, J.A.D., Zaiat, M., 2003. Operating feasi-bility of anaerobic whey treatment in a stirred sequencing batchreactor containing immobilized biomass. Water Sci. Technol. 48 (6),179–186.

Ruston, B., 1993. Chemical pretreatment of dairy wastewater. Water Sci.Technol. 28 (2), 67–72.

Saddoud, A., Hassaıri, I., Sayadi, S., 2007. Anaerobic membrane reactorwith phase separation for the treatment of cheese whey. Bioresour.Technol. 98 (11), 2102–2108.

Shigeki, U., Hideki, H., 1999. Treatment of sewage by a UASB reactorunder moderate to low temperature conditions. Bioresour. Technol.72, 275–282.

Strydom, J.P., Britz, T.J., Mostert, J.F., 1997. Two-phase anaerobicdigestion of three different effluents using a hybrid bioreactor. WaterSalinat. 23, 151–156.

Tango, M.S.A., Ghaly, A.E., 1999. Effect of temperature on lactic acidproduction from cheese whey using Lactobacillus helveticus underbatch conditions. Biomass Bioenerg. 16, 61–78.

Vidal, G., Carvalho, A., Mendez, R., Lema, J.M., 2000. Influence of thecontent in fats and proteins on the anaerobic biodegradability of dairywastewaters. Bioresour. Technol. 74, 231–239.

Walstra, P., 1990. On the stability of casein micelles. J. Dairy Sci. 73,1965–1979.

Yu, H.Q., Fang, H.H., 2001. Acidification of mid- and high-strength dairywastewaters. Water Res. 35, 3697–3705.

Zaiat, M., Rodrigues, J.A.D., Ratusznei, S.M., Camargo, E.F.M.,Borzani, W., 2001. Anaerobic sequencing batch reactors for wastewa-ter treatment: a developing technology. Appl. Microbiol. Biotechnol.55 (1), 29–35.