ARTICLE IN PRESS

Available at www.sciencedirect.com

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 6

0043-1354/$ - see frodoi:10.1016/j.watres

�Corresponding aE-mail address: sa

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

Treatment of olive oil mill wastewater by combined processelectro-Fenton reaction and anaerobic digestion

Sonia Khoufi, Fathi Aloui, Sami Sayadi�

Laboratoire des Bioprocedes, Centre de Biotechnologie de Sfax; B. P. ‘‘K’’, 3038 Sfax, Tunisia

a r t i c l e i n f o

Article history:

Received 23 October 2005

Received in revised form

19 March 2006

Accepted 26 March 2006

Keywords:

Olive mill wastewater

electro-Fenton

Anaerobic digestion

Polyphenols

Abbreviations:

OMW: olive mill wastewaters

COD: chemical oxygen demand

BOD5: biological oxygen demand

TSS: total suspended solids

LMM: low molecular mass

AF: anaerobic filter

VFA: volatile fatty acids

nt matter & 2006 Elsevie.2006.03.023

uthor. Tel./fax: +00 216 74mi.sayadi@cbs.rnrt.tn (S

A B S T R A C T

In this work, we investigated an integrated technology for the treatment of the recalcitrant

contaminants of olive mill wastewaters (OMW), allowing water recovery and reuse for

agricultural purposes. The method involves an electrochemical pre-treatment step of the

wastewater using the electro-Fenton reaction followed by an anaerobic bio-treatment. The

electro-Fenton process removed 65.8% of the total polyphenolic compounds and subse-

quently decreased the OMW toxicity from 100% to 66.9%, which resulted in improving the

performance of the anaerobic digestion. A continuous lab-scale methanogenic reactor was

operated at a loading rate of 10 g chemical oxygen demand (COD) l�1 d�1 without any

apparent toxicity. Furthermore, in the combined process, a high overall reduction in COD,

suspended solids, polyphenols and lipid content was achieved by the two successive

stages. This result opens promising perspectives since its conception as a fast and cheap

pre-treatment prior to conventional anaerobic post-treatment. The use of electro-

coagulation as post-treatment technology completely detoxified the anaerobic effluent

and removed its toxic compounds.

& 2006 Elsevier Ltd. All rights reserved.

1. Introduction

Treatment and disposal of olive mill wastewater (OMW)

represents one of the main problems for olive oil producing

countries of the Mediterranean area. Tunisia is one of the

largest olive oil producers in the world with an average

annual production of 450,000 tons. This results in a by-

product of 600,000 m3 OMW. These liquid residues are

100–150 times more heavily loaded with pollutants than

ordinary domestic wastewater (Sabbah et al., 2004). The high

polluting activity of OMW is linked with their high content of

organic molecules, especially polyphenolic mixtures

(4–10 g l�1) with different molecular weights (Hamdi, 1992),

r Ltd. All rights reserved.

440 452.. Sayadi).

as well as their acidity and high concentration of potassium,

magnesium and phosphate salts (Arienzo and Capasso, 2000).

Besides aromatic compounds, OMW contain other organic

molecules including nitrogen compounds, sugars, organic

acids, and pectins (Della Greca et al., 2000), that increase their

organic load (chemical oxygen demand (COD) ¼ 80–200 g l�1;

biological oxygen demand (BOD5) ¼ 50–100 g l�1). Further-

more, the physico-chemical characteristics of OMW are rather

variable, depending on climatic conditions, olive cultivars,

degree of fruit maturation, storage time, and extraction

procedure.

Many pollution disposal methods, such as concentration,

evaporation, incineration, ultrafiltration/reverse osmosis,

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 62008

lime precipitation, aerobic treatment, lagooning, codigestion,

etc were tested on OMW, none of them led to industrial

applications. However, anaerobic digestion seems to have

some clear advantages that would make it the process of

choice (Borja et al., 1992). Indeed, this treatment process

produces energy (methane) and a digested effluent with a

significant reduction of the organic load (Marques, 2001).

However, many problems concerning the high toxicity and

inhibition of biodegradation of these effluents were encoun-

tered during anaerobic treatments, because some bacteria,

such as methanogens, were particularly sensitive to the

organic contaminants present (Andreoni et al., 1984). The

phenolic compounds severely limit the possibility of using

anaerobic digestion (Sayadi et al., 2000). Therefore, the

elimination of phenolic compounds from OMW was consid-

ered as an important objective in order to reduce its toxicity

and to permit the occurrence of microbial fermentation. For

this, research turned to a more promising alternative, namely

the physico-chemical pre-treatment to remove the toxic

compounds of OMW (Beccari et al., 1999).

In recent years, there has been increasing interest in the

use of electrochemical technologies for the treatment of

wastewaters. This technique was found to be successful in

removing pollutants in various industrial wastewaters (Lin

and Chang, 2000; Ciardelli and Ranieri, 2001; Lai and Lin,

2004). In two recent investigations, Inan et al., (2004) and

Adhoum and Moncer (2004), an electrochemical method was

used for decreasing the organic matter in OMW. Both

investigators found efficient removals of COD, colouration

and polyphenols content by electrolysis process using alumi-

nium and iron electrodes. However, a relatively new chemical

oxidation method that has not received much attention for

OMW or other industrial wastewater treatment is the electro-

Fenton method (Lin and Chen, 1997). This method represents

a combination of the electrochemical process and the Fenton

oxidation. It is based on the fact that hydrogen peroxide

(H2O2) can be used as an oxidant in advanced oxidation

processes to decompose refractory or toxic wastewaters

(Kusvuran et al., 2004). As indicated in reaction (1), when

the ferrous ion reacts with H2O2 it will generate strong

oxidant hydroxyl radicals (OH � ).

Fe2þ þH2O2 ! Fe3þ þOH� þOH�: (1)

This Fe2+/H2O2 system, often referred to as Fenton’s reagent

(Fenton’s), has dual functions of OH radical peroxidation as

well as ferrous/ferric coagulation.

During this process, the non-biodegradable organics and

toxic pollutants present in the wastewaters such as poly-

phenols are usually destroyed by direct or indirect anodic

oxidation via the production of oxidants such as hydroxyl

radicals and complex coagulants that promote the floccula-

tion of the matter (Israilides et al., 1997; Panizza et al., 2000;

Chen et al., 2002).

This paper will attempt to apply the electro-Fenton process

to reduce the organic load and the toxicity of OMW in order to

improve the anaerobic digestion in terms of biomethane

yield. Electro-coagulation was assayed as a post-treatment for

complete detoxification and colour removal allowing water

recovery and reuse for agricultural purposes.

2. Materials and methods

2.1. OMW characterisation

Fresh OMW was obtained from an olive oil continuous

processing plant located in Sfax (southern Tunisia). The

OMW was characterised by high total suspended solids

(TSS) content, COD concentration up to 100 g l�1 and poly-

phenols up to 12 g l�1. Raw OMW was pre-decanted in a 120 l

decanter before being treated by electro-Fenton in order to

remove suspended solids (Fig. 1).

To confirm the role of electro-Fenton in polymerising and

removing the highly polymerised phenolic fraction, experi-

ments were carried out with a low-molecular-mass (LMM)

polyphenolic fraction (o2 kDa) obtained by the ultrafiltration

of crude OMW using a polysulphone 2 kDa cut-off membrane.

The purpose of this ultrafiltration was to study the effect of

electro-Fenton reaction on the toxic fraction of OMW which is

composed of LMM phenolics such as simple phenolics

(hydroxytyrosol, tyrosol, p-OH benzoic acid, p-OH phenyl

acetic acid, vannilic acid, caffeic acid, coumaric acid, vanni-

line, ferilic acid, catechol, methylcatechol), tannins, antocya-

nins, catechin (Sayadi et al., 2000). The C18-HPLC

chromatogram of this OMW phenolic fraction is presented

in Fig. 2.

2.2. Electro-Fenton and electro-coagulation treatment

Preliminary experiments were carried out in a 0.25 l�1 glass

reactor for the electro-Fenton of OMW fraction. The aqueous

solution of reactants was homogenised by magnetic agitation

to avoid concentration gradients. The electro-Fenton reactor

was formed by one pair of anodic and cathodic electrodes

(cast iron plates) which were positioned approximately 1.5 cm

apart from each other and were dipped in the effluent. The

total effective surface area of electrodes was 0.2 dm2. The

current input was supplied by a convergy power supply. In

each run, approximately 0.2 l of OMW fraction was placed in

the electrolytic cell. The pH of the solution was adjusted to 4.

H2O2 was added to the electrolytic cell before the electrical

current was turned on. A batch study was conducted to

optimise parameters like H2O2 concentration and current

density governing the electro-Fenton process. These para-

meters were examined in the range of 0–1.5 g l�1 and

1.25–10 A dm�2, respectively. The optimum H2O2 concentra-

tion and current density were found to be 1 g l�1 and

7.5 A dm�2, respectively. At these conditions, maximum

removal of monomer concentration, COD and colour were

attained. For this reason, these conditions were chosen as the

optimised parameters and were subsequently used for

preparing the pre-treated OMW for the biomethanisation.

Experiments of electro-Fenton of crude OMW were conducted

in a 5 l glass reactor using iron electrodes having an effective

surface area of 150 dm2 (Fig. 1). In each run, 3 l of crude OMW

were treated and operated in batch mode.

Electro-coagulation of anaerobic effluent was carried out in

the same reactor as for the electro-Fenton of crude OMW

without stirring. This electrolysis process lasted 2 h at

1.8 A dm�2 and without adjustment of pH.

ARTICLE IN PRESS

Fig. 2 – C18-HPLC chromatogram of phenolic compounds present in OMW fraction. 1: Hydroxytyrosol; 2: 3, 4 dihydroxyphenyl

acetic acid, 3: tyrosol; 4: p-OH benzoıc acid; 5: p-OH phenylacetic acid; 6: vanillic acid; 7: caffeic acid; 8: coumaric acid; 9:

vanillin; 10: ferulic acid.

Fig. 1 – Schematic representation of OMW treatment process.

WAT E R R E S E A R C H 40 (2006) 2007– 2016 2009

After electro-Fenton reaction and electro-coagulation, trea-

ted effluents were placed in decanter tanks (Fig. 1) to

eliminate sludge formed during electrolysis. During the

experiments, samples were withdrawn and immediately

analysed for water quality measurements.

2.3. Anaerobic biotreatment and biogas analysis

Two anaerobic filters (AFs) were used in this study. These

reactors were made of a glass column having a working

volume of 3 l. The inner tubes were enclosed in a jacket

through which hot water was circulated to maintain the

temperature of the filter at 37 1C. These AFs were packed with

polyurethane foam cubes 2 cm�2 cm�1 cm (Filtren T45,

from Recticel, Wetteren, Belgium) as support and inoculated

with an 8-year-old digester operated with pre-treated OMW.

The influent was fed in six times into the reactor using a

pump connected to a programmer. For monitoring the volatile

fatty acids (VFA) inside the reactor, three sampling points

were made in the AF. Level (A) was at the bottom of the

reactor. Level (B) corresponded to the middle and level (C) was

at the top of the reactor.

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 62010

The gas flow rates of the AFs were measured by liquid

displacement. Gas samples were taken with a syringe from

the tank of biogas. CH4, CO2 and N2 were measured using a

gas chromatograph GC11 (Delsi instruments) equipped with a

Haye SepQ 60/80 (SUPELCO) column (maintained at 60 1C), a

thermal conductivity detector (current intensity of 160 mA)

and a servotrace integrator (SEFRAM). Helium was used as a

carrier gas at a pressure of 1.3 bar.

2.4. Analytical methods

As the presence of residual H2O2 introduces a positive error in

COD determination (Kang et al., 1999), in electrolysis

experiments with H2O2, the pH of the samples was raised to

above 10 with NaOH 6 N prior to analysis. The value of

COD was estimated using the method described by Knechtel

(1978) and fading colour was monitored by measuring the

absorbance at 395 nm, the length of the maximum absor-

bance, using a spectrophotometer (ANTHLIE ADVANCED

5 SECOMAM). Samples were centrifuged for 20 min at

4000 t/min and diluted appropriately before each COD deter-

mination.

BOD5 was determined by the manometric method with a

respirometer (BSB-Controller Model 620T (WTW)).

Concentration of ortho-diphenols was determined by the

colorimetric reaction with Folin–Ciocalteau reagent. An

aliquot of the OMW aqueous methanol extract was mixed

with 2 ml of Folin–Ciocalteau reagent (Fluka, Switzerland). A

sodium hydroxide solution (6% v/v) was added, and the

mixture was shaken. The blue colour formed was measured

at 727 nm. The ortho-diphenol concentration of OMW sam-

ples, as determined by the Folin–Ciocalteau method, (Folin

and Ciocalteau, 1927) were reported as caffeic acid equiva-

lents by reference to a standard curve. As about total

polyphenols, they were quantified using the method de-

scribed by Sayadi et al. (2000).

Concentration of aromatic compounds was determined by

high-performance liquid chromatography (HPLC) using a

Shimadzu 10AVP chromatograph equipped with a Shimadzu

10AVP UV detector. Separation was made by a column (Shim-

pack CLC-ODS (M) 250 mm�4.6 mm) washed with acetoni-

trile/water (70/30) before and after analysis. A mixture of 50%

acetonitrile in 50% water was chosen as optimal mobile

phase. Data were analysed by class VP Shimadzu software.

A Progel TSK-G 2000-SW Supelco column (300 mm�

7.8 mm) was used with the same Shimadzu apparatus to

analyse molecular-mass distribution of the OMW polyphe-

nols. The elution was carried out using a phosphate buffer of

pH 6.8 and 0.6 ml min�1 flow rate. The wavelength of the

detector was adjusted to 280 nm.

The standard method of Soxhlet solid/liquid (organic solids

of OMW/hexane) was utilised for the dosage of lipids.

VFAs (acetate, propionate, butyrate, isobutyrate and vale-

rate) were measured by HPLC using the method described by

Mechichi and Sayadi (2005).

The microtoxicity test consists of the inhibition of the

bioluminescence of Vibrio fischeri LCK480 using the LUMIStox

system (Dr Lange GmbH, Dusseldorf, Germany) and was

carried out according to ISO 11348-2 (1998). Percentage

inhibition of the bioluminescence was achieved by mixing

0.5 ml of OMW and 0.5 ml luminescent bacterial suspension.

After a 15 min exposure at 15 1C, the decrease in light

emission was measured. The toxicity of the OMW is

expressed as the percent of the inhibition of bioluminescence

(%IB) relative to a non-contaminated reference. A positive

control (7.5% NaCl) was included for each test.

Phytotoxicity test was estimated by the determination of

the germination index (GI) according to Wong et al. (2001)

using Lycopersicon esculentum (tomato) seeds.

3. Results and discussion

3.1. Electro-Fenton treatment

3.1.1. Electro-Fenton treatment of OMW fractionPreliminary tests were conducted to study the effect of

electro-Fenton reaction on the pollutant characteristics of

the LMM phenolic fraction. Experiments were realised with

1 g l�1 H2O2 added and a current density of 7.5 A dm�2 (see

materials and methods section). The initial pH of the OMW

fraction (4.8) was decreased to pH 4 in order to allow the

Fenton’s peroxidation. Fig. 3 shows the evolution of pH, COD

removal, colouration, monomers removal, hydroxytyrosol

concentration and toxicity using V. fischeri based on LUMISTox

system during the electro-Fenton treatment. The pH in-

creased from 4 to 9 (Fig. 3a). The final pH differed in function

of the quantity of the H2O2 added (data not shown) and the

duration of the treatment. For 1 g l�1 of H2O2 and a 6 h

treatment, the final pH obtained (pH 10.5) was not favourable

for anaerobic post-treatment. The COD removal was 26% (Fig.

3b) at the steady state (after 4 h) but could be higher for a

prolonged treatment period. The colour intensity of the

effluent fraction monitored by measuring absorbance at

395 nm (Fig. 3c) doubled after 30 min of reaction, probably

due to the polymerisation of monoaromatic compounds.

Then it decreased to 75% of the initial colour after 4 h. This

can be explained by the coagulation of the highly polymerised

polyphenolic compounds.

Chromatography techniques confirmed the removal of the

most phenolics of LMM (Fig. 3d) which resulted in decreasing

the toxicity from 100% to 67% after 30 min (TF1) and to 28%

after 4 h of incubation (TF2) (Fig. 3f). Indeed, approximately

90% of the mono-aromatic compounds were removed after

4 h of incubation. As an example, the concentration of

hydroxytyrosol, the major ortho-diphenol, decreased rapidly

in the OMW fraction (Fig. 3e); a removal of 98% was achieved

at the end of treatment.

To test whether the dark colour was caused by

polymerisation of the OMW fraction, the molecular-mass

distribution of the reaction mixtures was measured by fast

SE-HPLC and compared with the control. A polyphenol

fraction with high hydrodynamic volumes was formed after

OMW oxidation by electro-Fenton oxidation, suggesting that

polymerisation had taken place (Fig. 4). After this polymer-

isation step which was related to the colour increase, a

coagulation of the highly condensed polymers followed by

rapid sedimentation occurred. This resulted in the decolour-

isation of the OMW fraction.

ARTICLE IN PRESS

4

5

6

7

8

9

10

0 2 4

pH

Time (h)

0

20

40

60

80

100

1 3 8 10

Max

imum

rem

oval

(%

)

Monomers

0

10

20

30

0 1

CO

D r

emov

al (

%)

Time (h)

0

200

400

600

800

1000

0 2 3 4

Hyd

roxy

tyro

sol

conc

entr

atio

n (m

g l-1

)

Time (h)

0

50

100

150

200

250

0 2 3 4

Res

idua

l col

orat

ion

(%)

Time (h) TF2TF1

F0

20

40

60

80

100

Rel

ativ

e in

hibi

tion

(%)

1 3 2 4 5 6 7 9

2 3 4 1

1

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 3 – Variations of pH (a), COD removal (b), residual colour (c), monomer removal (d) (1: hydroxytyrosol; 2: 3, 4 di-OH

phenylacetic acid, 3: tyrosol; 4: p-OH benzoıc acid; 5: p-OH phenylacetic acid; 6: vanillic acid; 7: caffeic acid; 8: coumaric acid;

9: vanillin; 10: ferulic acid), hydroxytyrosol concentration (e) and the relative toxicity of OMW fraction (F) and the reactants

collected during the Fenton process at 30 min (TF1) and at 4 h (TF2) of treatment (f).

WAT E R R E S E A R C H 40 (2006) 2007– 2016 2011

3.1.2. Electro-Fenton treatment of crude OMWTable 1 shows the characteristics of OMW before and after

electro-Fenton treatment. After decantation, the TSS of raw

OMW decreased from 59 to 12 g l�1. The residual TSS in crude

OMW was unsettlable suspended matter, which presents a

major difficulty in the treatment and handling of OMW.

During electro-Fenton treatment, pH increased from 4 to 7.6,

which may be attributed to the smaller production of H+ than

OH� as was explained by Israilides et al. (1997) and the

reduction in phenol concentration. Indeed, phenols are acids

in liquids, and their removal from a solution reduces its

acidity. The pH value of electro-Fenton-treated OMW can be

considered favourable for anaerobic bio-treatment.

Biodegradability is determined by measuring the ratio

between COD and BOD5, whose value must be in the range

of 2–2.5. After electro-Fenton, the COD of crude OMW drops to

approximately 68% of the initial value. This result points out

the ability of the electrolysis process to eliminate soluble

compounds present in OMW. BOD5 values decreased from

19.25 to 15.5 g l�1 before and after treatment, respectively.

Thus, COD/BOD5 ratio decreased from 5.84 before to 2.26 after.

It appears that a significant proportion of the non-biodegrad-

able matter present in OMW was removed by electro-Fenton.

Degradation and mineralisation of phenolic compounds can

occur during Fenton reaction. Kavitha and Palanivelu (2004)

reported that in Fenton process, biodegradable aliphatic

ARTICLE IN PRESS

Fig. 4 – Molecular-mass distribution of phenolics from the untreated ultra-filtrated OMW phenolic fractions (___) and treated

fraction ( ). Arrows indicate standard elution times: from left to right: 1: blue dextran (MW ¼ 2000 kDa), 2: lysozym

(MW ¼ 15 kDa), 3: syringic acid (MW ¼ 198 Da).

Table 1 – Compositions of olive mill wastewaters before and after the treatment with the electro-Fenton process (S1, seeFig. 1) then with anaerobic bio-treatment (S2, see Fig. 1) and finally by electro-coagulation (S3, see Fig. 1)

Parameter Crude OMW S1 S2 S3

pH 5.44 7.60 7.80 9.20

Color (absorbance 395 nm) 73.00 16.10 13.16 1.19

UV absorbance 280 nm 167.30 54.20 49.60 4.60

BOD5 (g l�1) 19.25 15.50 — —

COD (g l�1) 112.50 36.00 8.30 2.50

COD/BOD5 5.84 2.26 — —

Total solids (%) 12.20 3.90 1.95 0.97

Total volatiles (%) 10.90 2.10 1.50 0.65

Total suspended solids (g l�1) 59.00 2.70 1.80 0.53

Volatiles suspended solids (g l�1) 55.17 2.30 1.65 0.32

Ortho-diphenols (mg l�1) 6025.50 1536.70 861.80 28.57

Total polyphenols (g l�1) 11.75 4.20 1.20 Not detected

Residual oils (g l�1) 12.00 1.30 Not detected Not detected

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 62012

compounds such as acetic acid and oxalic acid were identified

as the major products during the degradation of synthetic

phenol. However, transformation of phenolic polymers to

simple phenolic compounds was not demonstrated.

Chromatography analysis (data not shown) confirmed the

removal of most LMM phenolics. Besides, the concentration

of ortho-diphenols, monitored by Folin–Ciocalteau method,

was significantly reduced during the electro-Fenton process.

Removal efficiency was about 65.8% for total polyphenols and

74.5% for ortho-diphenols.

Crude OMW was highly coloured due to its high content of

polyaromatic compounds. In the beginning of the electrolysis

treatment, the colour intensity of the effluent increased

(data not shown) as a result of phenolic compounds

polymerisation. However, colour intensity decreased to 78%

of the initial colour at the end of treatment.

During the electrolysis treatment, a part of the solute and

particle matter present in OMW turned out to be a suspended

solid that could reach 40 g l�1 at the end of the electrolysis

reaction. These TSS were rapidly eliminated by simple

sedimentation. After decantation, the obtained effluent has a

weak quantity of TSS (2 g l�1) in comparison with the decanted

crude OMW (12 g l�1). The formation of suspended particles

was caused presumably by electro-coagulation process. The

polymers were precipitated with iron which was continuously

dissolved into the wastewater from the cast iron anodes, as

governed by the Faraday’s law (Pletcher and Walsh, 1990).

This result confirms the hypothesis that the electro-Fenton

reaction would have a strong ability to eliminate polyphenols

from OMW. Furthermore, as can be seen in the Table 1, the

concentration of lipids was decreased by 89.2%. The pH, COD,

colouration, polyphenols and lipids removal were consis-

tently very good. Indeed, the effluent quality of the

pre-treated OMW by electro-Fenton process (S1, see Fig. 1)

was rather excellent (Table 1). It could be directly fed as

influent to anaerobic reactor.

3.2. Anaerobic bio-treatments

3.2.1. Anaerobic digestion of non pre-treated OMWThe anaerobic treatment of non pre-treated OMW was

performed in a 3-l AF reactor. The yield of methanisation of

ARTICLE IN PRESS

WAT E R R E S E A R C H 40 (2006) 2007– 2016 2013

this untreated diluted OMW was higher than 0.3 l CH4 g�1 COD

introduced at low loading rates. However, since the 26th day,

when the loading rate reached a mean of 4 g l�1 d�1 of COD, a

decrease in the biogas production and yield was observed

(Fig. 5A). This toxicity was accompanied by a pH decrease in

the three levels of the reactor and an accumulation of the VFA

(Fig. 6). This test of the anaerobic digestion of untreated OMW

by an 8-year OMW-acclimated consortium will serve as a

control for comparing the efficiency of the electro-Fenton pre-

treatment in the detoxification of this effluent.

3.2.2. Anaerobic digestion of electro-Fenton pre-treated OMWThe AF was loaded with undiluted pre-treated OMW at a

starting loading rate of 2 g COD l�1 d�1. The reactor was

operated at influent OMW concentration of 35.5 g COD l�1

0

2

4

6

8

10

0 20 40 60

Load

ing

rate

(g

CO

D l-1

d-1

)

Time (days)

Aa

0

1

2

3

4

5

0 20 40 60

Bio

gas

(l l-1

)

Ab

Time (days)

0

0.1

0.2

0.3

0.4

0 20 40 60

Met

hane

yie

ld (

Y)

Ac

Time (days)

Fig. 5 – Evolution of the loading rate (g COD l�1 reactor d�1) (a), b

introduced) (c) during anaerobic digestion of crude OMW (A) an

(mean value). The hydraulic retention time (HRT) varied

between 17.7 and 3.5 days. In general, the percentage of

COD removal decreased with increased loading rate during

the fermentation of OMW in the AF. The percentage of COD

removal decreased from 88.8% to 68% when the organic

loading rate increased from 2 to 10 g COD l�1 d�1 (Fig. 7). The

evolutions of the loading rate, biogas productivity and

methane yield are presented in Fig. 5B.

At the higher loading rates (9–10 g COD l�1 d�1), the yields

obtained were approximately 0.3 l CH4 g�1 COD introduced.

The volume of biogas reached 12 l d�1 (4-fold of the volume of

the digester). The higher values of yields (0.32–0.34 l CH4 g�1

COD introduced) were obtained for loading rates lower than

8 g COD l�1 d�1. In addition, Fig. 5Ba and Bc show that the

methane yield increased with the increase of the loading rate

0

2

4

6

8

10

0 20 40 60 80 100

Ba

Load

ing

rate

(g

CO

D l-1

d-1

)

Time (days)

0

1

2

3

4

5

0 20 40 60 80 100

Bb

Bio

gas

(l l-1

)

Time (days)

0

0.1

0.2

0.3

0.4

0 20 40 60 80 100

Bc

Time (days)

Met

hane

yie

ld (

Y)

iogas production (l l�1) (b) and methane yield (l CH4 g�1 COD

d electro-Fenton pre-treated OMW (B).

ARTICLE IN PRESS

4

5

6

7

8

9

10

0 20 40 60

A B C

Time (days)

pH

0

3

6

9

12

15

0 20 40 60

A B C

Time (days)

VF

A g

l-1

(a)

(b)

Fig. 6 – Evolution of pH (a) and concentration of VFA (b) in the

anaerobic filter during the methanisation of crude OMW (A:

bottom of the reactor, B: middle of the reactor, C: top of the

reactor).

0

10

20

30

40

50

0 20 40 60 80 100

COD inf COD eff

Time (days)

CO

D g

l-1

Fig. 7 – Evolution of the COD of influent and effluent of the

anaerobic filter alimented by electro-Fenton-treated OMW.

4

5

6

7

8

9

10

0 20 40 60 80 100

A B C

Time (days)

pH

(a)

0

2

4

6

8

10

0 20 40 60 80 100

A B C

Time (days)

VF

A g

l-1

(b)

Fig. 8 – Evolution of pH (a) and concentration of VFA (b) in the

anaerobic filter during the methanisation of electro-Fenton-

pre-treated OMW (A: bottom of the reactor, B: middle of the

reactor, C: top of the reactor).

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 62014

up to the theoretical steady yield value, reported to be around

0.35 l CH4 g�1 COD introduced. This observation can be

considered a solid proof for the ability of the anaerobic

biomass to degrade most organic matter present in the

electro-Fenton-pre-treated OMW. Besides, it may confirm

the gradual increase of the methanogenic activity.

The biomethanisation process was found to be stable

during 3 months of operation. No toxicity phenomenon was

observed. VFAs have long been recognised as the most

important intermediates in the anaerobic process and were

proposed as a control parameter (Ahring et al., 1995; Mechichi

and Sayadi, 2005). Therefore, VFA and pH were analysed in

the three levels of the AF (Fig. 8). Level A is at the bottom of

the reactor, level B corresponds to the middle and level C is in

the top of the reactor.

The pH at these three levels of the reactor was higher than

7.0 for all the loading rates applied. The VFA concentrations

were low even at the higher loading rates. Knowing that

untreated OMW causes inhibition of methanisation at a

loading rate of 2–4 g COD l�1 d�1 (Kang and Chang, 1997;

Hamdi, 1991; Rozzi et al., 1989, this work), it can be concluded

that electro-Fenton of OMW resulted in decreasing the toxic

effect of this wastewater on anaerobic digestion. Moreover,

this experiment was stopped at a loading rate of

10 g COD l�1 d�1 while the biological process did not show

any apparent toxicity. These results also suggest that

anaerobic digestion can be a practical alternative for the

treatment of OMW.

3.2.3. Characterisation of the anaerobic effluentThe anaerobic effluent (S2) was characterised with common

parameters (pH, COD, colouration, TSS, ortho-diphenols).

Main results were plotted in Table 1. Results showed that

the colouration and the residual COD (hardly biodegradable

compounds) of S2 remained relatively high. The phytotoxicity

test of OMW samples were carried out using the germination

index (GI) of L. esculentum (tomato). Results showed that

electro-Fenton treatment increased the GI percentage of

L. esculentum from 4.4% (for crude OMW) to 30% while

the anaerobic effluent led to an increase of the GI to

ARTICLE IN PRESS

Table 2 – Inhibition of Vibrio fisheri luminescence (IB) afterexposure with different OMW samples during 15 minand the GI percentage of Lycopersicon esculentum

OMW sample IB (%) GI (%)

Untreated OMW 100 4.4

Electro-Fenton-treated OMW (S1) 66.9 30

Electro-Fenton anaerobic-treated OMW (S2) 45.2 121

Electro-coagulated S2 (S3) — 140

WAT E R R E S E A R C H 40 (2006) 2007– 2016 2015

121% compared to 100% for the control (Table 2). Indeed,

approximately 43.9% of the ortho-diphenols were removed

after anaerobic bio-treatment.

As shown in Table 2, untreated OMW exercised 100%

inhibition on V. fischeri. It was reduced to 66.9% after pre-

treatment by electro-Fenton and to 45.2% in the anaerobic

effluent. Microtoxicity of S2 remained high due to the residual

VFA. Yet, the characteristics of anaerobic effluent do not

comply with legal requirements. To overcome this problem, a

tertiary treatment step was necessary if we want to reach the

Tunisian standard requirements. For this purpose, experi-

ments of electro-coagulation of S2 were carried out using the

same electro-Fenton reactor, in order to remove the residual

polyphenols, COD and colour.

3.3. Improvement of the quality of the effluent usingelectro-coagulation

The purpose of this part of study was directed to treat the

anaerobic effluent (S2) by electro-coagulation process. During

this process, when direct current passed though the Fe

anodes, Fe2+ and Fe3+ correspondingly dissolved and com-

bined with hydroxyl ions in the water. They formed metal

hydroxyls ions, which are partly soluble in water under

definite pH values and play the role of coagulant.

The electro-coagulation step was performed at a current

density of 1.8 A dm�2 and without adjustment of pH. The

determination of the physico-chemical parameters of the

electro-coagulated anaerobic effluent (S3) showed that the

electrolysis process was able to remove 70.55% of TSS, 91% of

the colour and 70% of the residual COD (Table 1). Moreover,

the analysis of ortho-diphenols showed a removal efficiency

of 97% while polyphenolic compounds were not detected.

Hence, the final effluent (S3) was free of toxic compounds as

can be seen in Table 1. Furthermore, the phytotoxicity of S3

was significantly reduced by the application of electro-

coagulation, reaching 140% germination index (Table 2). As

conclusion, the proposed process of OMW treatment reduces

significantly its biotoxicity. For this, OMW can be used as

fertiliser.

4. Conclusion

The electro-Fenton method applied on raw OMW as pre-

treatment resulted in removing a large amount of recalcitrant

polyphenolic compounds as well as in decreasing toxicity.

The anaerobic process applied as post-treatment reached a

loading rate of 10 g COD l�1 d�1 without any apparent toxicity.

Finally, electro-coagulation of the anaerobic digestion effluent

could be used as polishing step for improving the quality of

the treated water for potential reuse.

Acknowledgements

This research was supported by C.I.U.F. (Belgium), EEC

Contract no. ICA3-CT-2002-00034 and Contract Programmes

(MRSTDC, Tunisia).

R E F E R E N C E S

Adhoum, N., Moncer, L., 2004. Decolourization and removal ofphenolic compounds from olive mill wastewater by electro-coagulation. Chem. Eng. Process 43 (10), 1281–1287.

Ahring, B.K., Sandberg, M., Angelidaki, I., 1995. Volatile fatty acidsas indicators of process imbalance in anaerobic digestors.Appl. Microbiol. Biotechnol. 43, 559–565.

Andreoni, V., Ferrari, A., Ranali, G., Sortini C., 1984. The influenceof phenolic acids present in oil mill waters on microbialgroups for the methanogenesis. In: Proceedings of the Inter-national Symposium on Olive By-products Valorization, FAO,Espana.

Arienzo, M., Capasso, R., 2000. Analysis of metal cations andinorganic anions in olive oil mill wastewaters by atomicabsorption spectroscopy and ion chromatography. Detectionof metals bound to the organic polymeric fraction. J. Agric.Food Chem. 48, 1405–1410.

Beccari, M., Majone, M., Riccardi, C., Savarese, F., Torrisi, L., 1999.Integrated treatment of olive oil mill effluents: Effect ofchemical and physical pretreatment on anaerobic treatability.Water Sci. Technol. 40, 347–355.

Borja, R., Martin, A., Maestro, R., Alba, J., Fiestas, J.A., 1992.Enhancement of the anaerobic digestion of OMW by theremoval of phenolic inhibitors. Process Biochem. 27, 231–237.

Chen, X., Chen, G., Yue, P.L., 2002. Novel electrode system forelectroflotation of wastewater. Environ. Sci. Technol. 36 (4),778–783.

Ciardelli, G., Ranieri, N., 2001. The treatment and reuse ofwastewater in the textile industry by means of ozonationand electrofloculation. Water Res. 35 (2), 567–572.

Della Greca, M., Fiorentino, A., Monaco, P., Previtera, L., Temussi,F., 2000. Phenolic components of olive mill wastewaters. Nat.Prod. Lett. 14, 429–434.

Folin, O., Ciocalteau, U., 1927. On tyrosine and tryptofandeterminations protein. J. Biol. Chem. 73, 627–650.

Hamdi, M., 1991. Effects of agitation and pretreatment on thebatch anaerobic digestion of olive mill wastewater. Biores.Technol. 36, 173–178.

Hamdi, M., 1992. Toxicity and biodegradability of olive millwastewaters in batch anaerobic digestion. Appl. Biochem.Biotechnol. 37, 155–163.

Inan, H., Dimoglo, A., Himsek, H., Karpuzcu, M., 2004. Olive oilmill wastewater treatment by means of electrocoagulation.Sep. Purif. Technol. 36, 23–31.

ISO 11348-2, 1998. Water quality—determination of the inhibitoryeffect of water samples on the light emission of Vibrio fischeri(Luminescent bacteria test)—Part 2: Method using liquid-driedbacteria.

Israilides, C.J., Vlyssides, A.G., Mourafeti, V.N., Karvouni, G., 1997.Olive oil wastewater treatment with the use of an electrolysissystem. Biores. Technol. 61 (2), 163–170.

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 0 ( 2 0 0 6 ) 2 0 0 7 – 2 0 1 62016

Kang, S.F., Chang, H.M., 1997. Coagulation of textile secondaryeffluents with Fenton’s reagent. Water. Sci. Technol. 36, 215–222.

Kang, Y.W., Cho, M.J., Hwang, K.H., 1999. Correlation of hydrogenperoxide interference on standard chemical oxygen demandtest. Water Res. 33, 1247–1251.

Kavitha, V., Palanivelu, K., 2004. The role of ferrous ion in Fentonand photo-Fenton process for the degradation of phenol.Chemosphere 55, 1235–1243.

Knechtel, R.J., 1978. A more economical method for the determi-nation of chemical oxygen demand. J. Water Pollut. ControlFed. (May/June), 25–29.

Kusvuran, E., Gulnaz, O., Irmak, S., Atanur, O.M., Yavuz, H.I.,Erbatur, O., 2004. Comparison of several advanced oxidationprocesses for the decolorization of reactive red 120 azo dye inaqueous solution. J. Hazard. Mater. B 109, 85–93.

Lai, C., Lin, S., 2004. Treatment of chemical mechanical polishingwastewater by electrocoagulation: system performances andsludge settling characteristics. Chemosphere 54, 235–542.

Lin, S., Chang, C., 2000. Treatment of landfill leachate bycombined electro-Fenton oxidation and sequencing batchreactor method. Water Res. 34 (17), 4243.

Lin, S.H., Chen, M.L., 1997. Textile wastewater treatment byenhanced electrochemical method and ion exchange. Environ.Technol. 18, 739–747.

Marques, I.P., 2001. Anaerobic digestion treatment of olive millwastewater for effluent reuse in irrigation. Desalination 137,233–239.

Mechichi, T., Sayadi, S., 2005. Evaluating process imbalance ofanaerobic digestion of olive mill wastewaters. Process Bio-chem. 40 (1), 139–145.

Panizza, M., Bocca, C., Cerisola, G., 2000. Electrchemical treatmentof wastewater containing polyaromatic organic pollutants.Water Res. 34 (9), 2601–2605.

Pletcher, D., Walsh, F.C., 1990. Industrial Electrochemistry, seconded. Chapman and Hall, London, UK.

Rozzi, A., Passino, R., limoni, M., 1989. Anaerobic treatment ofolive mill effluents in polyurethane foam bed reactor. ProcessBiochem. 4, 68–74.

Sabbah, I., Marsook, T., Basheer, S., 2004. The effect of pre-treatment on anaerobic activity of olive mill wastewater usingbatch and continuous systems. Process Biochem. 39,1947–1951.

Sayadi, S., Allouche, N., Jaoua, M., Aloui, F., 2000. Detrimentaleffects of high molecular mass polyphenols on olive millwastewater biotreatment. Process Biochem. 35, 725–735.

Wong, J.W.C., Li, K., Fang, M., Su, D.C., 2001. Toxicity evaluationof sewage sludges in Hong Kong. Environ. Int. 27,373–380.