ARTICLE IN PRESS

Available at www.sciencedirect.com

WAT E R R E S E A R C H 4 1 ( 2 0 0 7 ) 2 9 8 7 – 2 9 9 5

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

�Corresponding auE-mail address:

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

Prevention of volatile fatty acids production and limitationof odours from winery wastewaters by denitrification

Andre Boriesa,�, Jean-Michel Guillotb, Yannick Sirea, Marie Couderca,Sophie-Andrea Lemairea, Virginie Kreimb, Jean-Claude Rouxb

aInstitut National de la Recherche Agronomique, Unite Experimentale de Pech Rouge, 11430 Gruissan, FrancebLaboratoire de Genie de l’Environnement Industriel, Ecole des Mines d’Ales, 30319 Ales Cedex, France

a r t i c l e i n f o

Article history:

Received 21 December 2006

Received in revised form

16 February 2007

Accepted 9 March 2007

Available online 30 April 2007

Keywords:

Volatile fatty acids

Odours

Nitrate

Denitrification

Winery wastewaters

Ponds

nt matter & 2007 Elsevie.2007.03.022

thor. Tel.: +33 4 68 49 44 00bories@supagro.inra.fr (A

a b s t r a c t

The effect of the addition of nitrate to winery wastewaters to control the formation of VFA

in order to prevent odours during storage and treatment was studied in batch bioreactors at

different NO3/chemical oxygen demand (COD) ratios and at full scale in natural

evaporation ponds (2� 7000 m2) by measuring olfactory intensity. In the absence of nitrate,

butyric acid (2304 mg L�1), acetic acid (1633 mg L�1), propionic acid (1558 mg L�1), caproic

acid (499 mg L�1) and valeric acid (298 mg L�1) were produced from reconstituted winery

wastewater. For a ratio of NO3/COD ¼ 0.4 g g�1, caproic and valeric acids were not formed.

The production of butyric and propionic acids was reduced by 93.3% and 72.5%,

respectively, at a ratio of NO3/COD ¼ 0.8, and by 97.4% and 100% at a ratio of NO3/

COD ¼ 1.2 g g�1. Nitrate delayed and decreased butyric acid formation in relation to the

oxidoreduction potential. Studies in ponds showed that the addition of concentrated

calcium nitrate (NITCALTM) to winery wastewaters (3526 m3) in a ratio of NO3/COD ¼ 0.8

inhibited VFA production, with COD elimination (94%) and total nitrate degradation, and no

final nitrite accumulation. On the contrary, in ponds not treated with nitrate, malodorous

VFA (from propionic to heptanoıc acids) represented up to 60% of the COD. Olfactory

intensity measurements in relation to the butanol scale of VFA solutions and the ponds

revealed the pervasive role of VFA in the odour of the untreated pond as well as the clear

decrease in the intensity and not unpleasant odour of the winery wastewater pond

enriched in nitrates. The results obtained at full scale underscored the feasibility and safety

of the calcium nitrate treatment as opposed to concentrated nitric acid.

& 2007 Elsevier Ltd. All rights reserved.

1. Introduction

The heavy biodegradable organic load that characterises food

industry wastewaters has multiple impacts on treatments

and on the environment: excessive sludge production, oxygen

requirements and nutrient deficiency linked to aerobic

processes, risk of acidification as a result of anaerobic

digestion and long-term storage of seasonal wastewaters in

order to prevent overload. The noxious odours during storage

and treatment of food industry wastewaters are a major

r Ltd. All rights reserved.

; fax: +33 4 68 49 44 02.. Bories).

environmental problem today (ADEME, 2005). Winery waste-

waters clearly fall within this framework (Guillot et al., 2000;

Bories, 2006; Chrobak and Ryder, 2006). Worldwide wine

production—a total of 26.5�109 Lyear�1 (OIV, 2005), of which

69% takes place in Europe—consumes approximately 0.8 L

water L�1 wine and generates large volumes of wastewater on

a seasonal basis (Rochard et al., 1996; Duarte et al., 1998; OIV,

1999; Picot and Cabanis, 1998; ITV, 2000; Rochard, 2005).

Natural evaporation in ponds, a rustic and economical

technique (Duarte et al., 1998; TV, 2000; Le Verge and Bories,

ARTICLE IN PRESS

Table 1 – Composition of the reconstituted winerywastewater

Concentration(g L�1)a

TheoreticalCOD (g O2 g�1)

CODbalance

(%)

pH 3.5

COD 22.3 100

Ethanol 4.5 2.09 42.2

Glucose 5.2 1.07 24.9

Fructose 4.9 1.07 23.4

Glycerol 0.52 1.22 2.8

Tartaric

acid

0.25 0.53 0.6

a Except pH.

WAT E R R E S E A R C H 4 1 ( 2 0 0 7 ) 2 9 8 7 – 2 9 9 52988

2004), well adapted to discharge variations and seasonal

production, is a widely used treatment method in the largest

wine production region in France, the Languedoc-Roussillon

(approximately 1.6�109 L of wine/year, 10% of the overall

European production) where 179 evaporation ponds were

identified (Lambert and Lecharpentier, 2006, personal com-

munication).

Volatile fatty acids form the major products resulting from

the fermentation of carbon compounds in winery waste-

waters (Bories et al., 2005; Bories, 2006), and are responsible

for characteristic foul odours as a result of their low olfactory

perception threshold (Le Cloirec et al., 1991). Other odorous

compounds such as esters, mercaptans and aldehydes were

also identified from winery wastewater treated in ponds

(Guillot et al., 2000).

The degradation of VFA by denitrifying microorganisms was

studied and used to eliminate nitrate in wastewaters (Min et

al., 2002; Elefsiniotis et al., 2004; Sponza and Atalay, 2004).

More generally, the denitrification treatment of wastewaters

has been investigated with various carbon sources, microbial

systems and models (Bolzonella et al., 2001; Chiu and Chung,

2003; De Lucas et al., 2005; Sage et al., 2006).

However, the outlook for curative treatment of VFA

by means of denitrification in winery wastewater evaporation

ponds appears to be compromised because of the massive

quantities of VFA already accumulated and the emission

of odours previous and/or subsequent to the curative

treatment. On the other hand, the prevention of VFA

formation by orienting the degradation of organic matter in

wastewater into odourless products through anaerobic re-

spiration with nitrate as the electron acceptor (denitrification)

was studied in winery wastewaters to which nitrate had been

added in the form of concentrated nitric acid (Bories et al.,

2005).

The stoichiometric requirement in nitrate depends on the

degree of carbon reduction, according to the following Eq. (1):

CmHnOp þ 0:4ð2mþ 0:5n� pÞNO3� !

mCO2 þ 0:2ð2mþ 0:5n� pÞN2

þ ½pþ 0:8ð2mþ 0:5n� pÞ � 2m�H2O

þ 0:4ð2mþ 0:5n� pÞOH�: (1)

Since the chemical oxygen demand (COD) is a function of

the degree of carbon reduction according to the following

Eq. (2):

CmHnOp þ 0:5ð2mþ 0:5n� pÞO2 ! mCO2 þ 0:5nH2O; (2)

the nitrate requirement in relation to the COD is defined as a

molar ratio NO3/O2 ¼ 0.8 mol mol�1, resulting in a weight ratio

of NO3/COD ¼ 1.55 g g�1.

In this study, the influence of the different NO3/COD ratios

in relation to VFA formation from a winery wastewater was

studied, and the effectiveness of the preventive treatment of a

pond using calcium nitrate was assessed by measuring

olfactory intensity in addition to physico-chemical analysis.

The addition of nitrate in increasing quantities to reconsti-

tuted winery wastewater was studied on the basis of the

evolution of the redox potential and VFA formation kinetics in

order to underscore competition between acidogenic fermen-

tation pathways and denitrification, and to determine

the NO3/COD ratio required to inhibit VFA formation.

Since the general aim was to prevent odours, it was necessary

to study and to compare winery wastewater evaporation

ponds both treated and not treated with calcium nitrate

under real conditions, with discharges over a long period of

time and involving microbial systems.

2. Materials and methods

2.1. Winery wastewaters

For the study of the NO3/COD ratios, reconstituted winery

wastewater was prepared with a mixture of red grape must

and red wine, in equal proportions and then diluted 10-fold

with distilled water. The composition is presented in Table 1.

The pH of the reconstituted wastewater was then adjusted to

pH 6.5 with sodium hydroxide.

The reconstituted winery wastewater was divided between

four 1-L closed glass bioreactors, equipped with a gas exhaust

pipe (i.d.: 6 mm) and butyl septa for sampling and reagent

addition, thermostatically controlled at 25 1C by water circu-

lating through double-walled tubing, stirred at a speed of

250 rpm with a magnetic stirrer. The pH was measured with

an Ingold electrode linked to an INGOLD 2300 pH metre,

and adjusted to pH 6.5 by the addition of 10 N sodium

hydroxide.

2.2. Study of the NO3/COD ratio

Three bioreactors containing reconstituted wastewater were

supplemented with nitrate in the form of soluble concen-

trated calcium nitrate (50% w/w), according to the quantities

corresponding to the NO3/COD mass ratios (w/w): 0.4; 0.8; 1.2.

The bioreactors supplemented with nitrate and a fourth

without nitrate (control) were inoculated to a volume rate of

5% (vol/vol) with a suspension of sediments taken from

evaporation ponds whose microbial activity was tested

beforehand in the laboratory. The wastewaters were incu-

bated for 21 days and samples (4 mL) were regularly taken

with a syringe through a septum and maintained at �18 1C

before being analysed.

ARTICLE IN PRESS

Table 2 – VFA production from the reconstituted winerywastewater at various NO3/COD ratios

NO3/COD (g g�1) 0 0.4 0.8 1.2

Acetic acid (mg L�1) 1633 1369 1841 0

Propionic acid (mg L�1) 1558 1639 428 0

Butyric acid (mg L�1) 2304 809 155 59

Valeric acid (mg L�1) 298 0 0 0

Caproic acid (mg L�1) 499 0 0 0

WAT E R R E S E A R C H 41 (2007) 2987– 2995 2989

2.3. Winery wastewater treatment in ponds, with andwithout nitrate enrichment

Experiments in evaporation ponds were carried out with

wastewater from a winery that produces 180,000 hL wineyear�1

(mainly red). Wastewater was collected in a settling tank

(40 m3), sieved and then evacuated to the evaporation ponds

with a pump (20 m3 h�1).

Wastewaters were alternatively discharged into two 7000-

m2 evaporation ponds, whose impermeability was ensured by

a layer of clay and bentonite. The volume of wastewater was

measured by an integrating flowmetre.

Soluble calcium nitrate (50% w/w), (NITCALs, YARA France,

Paris, France), stored in 1-m3 tanks was injected into the

wastewater discharge pipe, using a variable flow feed pump

(100–200 L h�1), coupled with the wastewater backflow pump.

Liquid samples (250 mL) from the ponds were taken at three

different locations and then mixed together to form the

sample to be subjected to physico-chemical analyses (con-

served at �18 1C).

2.4. Chemical analyses

Nitrate and nitrite were measured using an ion HPLC: Waters

IC Pak A column (50�4.6 mm) with eluent: borate/gluconate

buffer, pH 8.5 and 10% acetonitrile (1.2 mL min�1), Waters 717

autosampler, Waters 432 conductivity detector and EMPOWER

software (Waters, Millipore, USA). Volatile fatty acids, sugars,

ethanol, glycerol and organic acids were analysed by HPLC:

mobile phase water/H2SO4 0.002 M (0.4 mL min�1) with on-line

degassing, Waters 717 autosampler, Aminex HPX-87H column

(Bio-Rad, Hercules, CA, USA), Waters RI 2410 refractive index

detector and EMPOWER software (Waters, Millipore, USA).

The COD was measured with an MN029 COD-1500 test kit

(Macherey-Nagel, Duren, Germany). The redox potential (ORP)

was measured using a SenTix ORP (Ingold) electrode and a

WTW millivoltmetre.

2.5. Determination of olfactory intensity

Odour intensities were determined by comparison with an n-

butanol reference. Five solutions of 0.01, 0.05, 0.1, 0.5 and

1 g L�1 were placed in glass flasks corresponding to five

intensity levels (from 1 to 5), respectively, characterised as

follows: level 1 (very low), level 2 (low), level 3 (medium), level

4 (slightly strong) and level 5 (strong).

The odour intensity of each liquid effluent (pond with or

without nitrate during two different periods) placed in a glass

flask, was compared with n-butanol solutions by a jury of five

selected panellists, and the average level was then calculated.

Panellists were submitted to tests of sensitivity to n-butanol

solutions and VFA solutions beforehand. The quantitative

data obtained per intensity test were accompanied by a

qualitative description of the smell.

An experiment with 150 L of effluent without nitrate was

carried out in a wind tunnel expressly designed for emission

studies from areal sources (Leyris et al., 2000, 2005). The

purpose of this experiment was to simulate the emission

from the pond under controlled wind speed conditions

(1 m s�1). Air samples taken at 3 cm from the water surface

in Tedlar bags, were studied in both situations without

nitrate: on site at the pond and in the tank of the wind

tunnel. The odour intensity of these gas samples was

quantified according to the levels as described above.

A final experiment was carried out to determine the

relationship between odour and VFA from the water. To do

this, the most odorous VFA were selected and synthetic

solutions of VFA were made into a buffer composed of

Na2HPO4 (6 mM), NaH2PO4 (2 mM), Na2EDTA (1 mM) and NaCl

(185 mM), as described by Comanici et al. (2006). The

conditions of effluent without nitrate were reproduced,

including VFA concentration and a pH value of 6.8. The

concentrations (November period) were 513, 301 and

307 mg L�1 for propionic, butyric and valeric acids, respec-

tively. The concentrations (December period) for these acids

were the same as those values given in Table 5 (without

nitrate). The synthetic solutions were also evaluated by the

jury in order to compare odour intensity on an n-butanol

scale.

3. Results and discussion

3.1. Influence of the NO3/COD ratio on the formation ofVFA

3.1.1. VFA production from winery wastewaterReconstituted winery wastewater has a COD of 22.3 g O2 L�1,

91% of which consists of ethanol and sugars: glucose and

fructose (Table 1). Its organic load and its composition

correspond to that of winery wastewaters produced during

the winery’s maximum activity period: grape harvest, vinifi-

cation (Bories et al., 2005, 1998; Colin et al., 2005). The use of a

reconstituted winery effluent (mixture of wine and must,

diluted 10-fold) made it possible to carry out a laboratory

study of the NO3/DCO ratio under pre-established and

reproducible conditions, using a medium representative of

winery effluents and easily available, with a pre-determined

and stable composition. In this way, we were able to avoid the

high degree of variability of the composition of winery

effluents (Rochard et al., 1996; Duarte et al., 1998; Picot and

Cabanis, 1998; ITV, 2000; Rochard, 2005).

Composition in VFA produced during incubation of the

reconstituted winery wastewater with evaporation pond

microflora and at different NO3/COD ratios is given in Table 2.

In the absence of nitrate, microflora produces a mixture of

VFA: butyric acid (2304 mg L�1), acetic acid (1633 mg L�1),

ARTICLE IN PRESS

0

400

800

1200

1600

2000

2400

0

Time (h)

Buty

ric a

cid

(m

g L

-1)

NO3/COD = 0

NO3/COD = 0.4

NO3/COD = 0.8

NO3/COD = 1.2

100 200 300 400 500

Fig. 1 – Effect of various NO3/COD ratios (w/w): 0, 0.4, 0.8 and

1.2 on butyric acid production from reconstituted winery

wastewater.

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

0

Time (h)

OR

P (

mV

)

NO3/COD = 0

NO3/COD = 0.4

NO3/COD = 0.8

NO3/COD = 1.2

100 200 300 400 500

Fig. 2 – ORP behaviour from reconstituted winery wastes at

various NO3/COD ratios (w/w): 0; 0.4; 0.8 and 1.2.

WAT E R R E S E A R C H 4 1 ( 2 0 0 7 ) 2 9 8 7 – 2 9 9 52990

propionic acid (1558 mg L�1), caproic acid (499 mg L�1) and

valeric acid (298 mg L�1), through fermentation of the carbon

compounds in the wastewater: sugars, ethanol, glycerol and

organic acids (Table 1). VFA represent 45% of the initial COD of

the reconstituted wastewater.

Butyric, propionic, caproic, valeric and heptanoic acids

form the major fraction of products resulting from the

anaerobic degradation of the constituents of winery waste-

waters by evaporation pond microflora. Their low perception

threshold, their unpleasant odour, their high concentrations

and large accumulated quantities all contribute to the typical

characteristic noxious odour. Butyric and propionic acids are

generally produced by clostridia and propionibacteria, re-

spectively, from simple substrates such as sugars, lactic acid,

glycerol, etc., according to well-known fermentation path-

ways (Barbirato et al., 1997; Colin et al., 2001). However, the

formation of butyric, valeric, caproic and heptanoic acids

from ethanol, a characteristic constituent of winery waste-

waters (Colin et al., 2005), appears to be related to particular

fermentation processes, given its reduction level (Cred ¼ 6).

VFA accumulation subsequent to the overload of an anaerobic

digester supplied with ethanol was demonstrated by Smith

and McCarty (1988) and attributed to a reverse b-oxidation

process. Butyric and caproic acid production by co-fermenta-

tion of ethanol and acetic acid was observed in Clostridium

kluyveri (Gottschalk, 1986), and comparable metabolic proper-

ties were demonstrated in Eubacterium pyruvativorans (Wallace

et al., 2004). Ethanol alone is not fermented and propionic

acid can be used instead of acetic acid. On the basis of the

energy production of this pathway 1 ATP/6 ethanol

(Gottschalk, 1986), it is possible to project low microbial

growth and a low VFA production rate from the anaerobic

catabolism of the ethanol, which could explain the formation

of VFA with long carbon chains during extended storage of

winery wastewaters in evaporation ponds.

3.1.2. Effect of COD/NO3 ratiosFor a ratio of NO3/COD ¼ 0.4 g g�1, caproic and valeric acids,

whose carbon reduction degrees (Cred) are high (5.33 and 5.2,

respectively), were not produced. Butyric acid (Cred ¼ 5)

concentration was reduced by 64.9% and that of propionic

and acetic acids (Cred ¼ 4.67 and 4, respectively) was not

considerably modified. At a ratio of NO3/COD ¼ 0.8, butyric

acid production was reduced by 93.3% in relation to the

control, and that of propionic acid by 72.5%. The formation of

acetic, propionic, caproic and valeric acids was null for the

ratio of NO3/COD ¼ 1.2, and butyric acid production was very

low (59 mg L�1), a decrease of 97.4% in relation to the control

without nitrate.

Butyric acid production at different NO3/COD ratios is

presented in Fig. 1. In the absence of nitrate, butyric acid

formation was observed at the very beginning of wastewater

incubation with microflora, linked to the consumption of

sugars, glycerol and organic acids, and it occurred regularly

throughout the 21 days of incubation, at an average rate of

4.6 mg L�1 h�1. The addition of nitrate delayed and decreased

butyric acid production. At a ratio of NO3/COD ¼ 0.4, butyric

acid production only began after 185 h of incubation and, as of

the 330th hour, the production rate of butyric acid

(4.1 mg L�1 h�1) was as high as the one observed in the

absence of nitrate. With a ratio of NO3/COD ¼ 0.8, the

production rate of butyric acid remained low (0.7 mg L�1 h�1).

For the ratio of NO3/COD ¼ 1.2, butyric acid production was

only observed after the 330th hour and at a very low rate:

0.35 mg L�1 h�1.

The behaviour of the redox potential of the reconstituted

wastewater during incubation with microflora and with

different NO3/COD ratios is illustrated in Fig. 2. This figure

shows that in the absence of nitrate, the ORP is very low

(o�400 mV) during the most active butyric production phase

and then remains within the value range of �77 and �185 mV.

At a ratio of NO3/COD ¼ 0.4, ORP values ranged from +12 to

�32 mV during the phase when butyric acid was not

produced, and rapidly dropped to �186 mV when the produc-

tion of butyric acid began. The ORP always remained negative

when butyric acid production took place. For the ratio of NO3/

COD ¼ 0.8, the ORP remained positive until the 402nd hour

and only slightly decreased (�30 mV) at the end of the test

when butyric acid production began. The ORP remained

ARTICLE IN PRESS

Table 4 – Composition of the winery wastewater dis-charged into ponds

Concentrationa (g L�1)b COD balance (%)

pH 4.19

COD 16.62 100

Ethanol 3.91 49.2

Glucose 1.96 12.6

Fructose 2.86 18.4

Glycerol 0.40 2.9

Tartaric acid 0.33 1.1

a Except pH.b Mean on eight samples.

0

100

200

300

400

500

600

700

800

900

1000

29/0

9/05

06/1

0/05

13/1

0/05

20/1

0/05

27/1

0/05

03/11/

05

10/11/

05

17/11/

05

24/11/

05

01/1

2/05

08/1

2/05

15/1

2/05

Bu

tyric,

va

leric a

cid

(m

g L

-1) Butyric acid with NO

3

Valeric acid with NO3

Butyric acid without NO3

Valeric acid without NO3

Fig. 3 – Behaviour of butyric and valeric acids in winery

wastewater ponds supplemented or not with nitrate.

WAT E R R E S E A R C H 41 (2007) 2987– 2995 2991

positive (4+40 mV) throughout the test made with the ratio of

NO3/COD ¼ 1.2.

The effect of increasing nitrate concentrations was revealed

by the suppression of VFA production according to decreasing

order of the carbon reduction degree: caproic and valeric

acids first, butyric acid followed by propionic acid and, finally,

acetic acid. Competition between catabolic pathways fa-

voured the oxidative pathway through anaerobic respiration,

as opposed to fermentation processes. The prevention of

malodorous VFA formation (propionic to caproic acids)

obtained for NO3/COD ratios ranging from 0.8 to 1.2 g g�1, less

than the stoichiometric ratio (NO3/COD ¼ 1.55 g g�1), suggests

that acidogenic substrates of winery wastewaters (sugars,

ethanol) were preferentially used by anaerobic respiration

with nitrate whose redox potential conditions inhibited

anaerobic catabolic pathways. When nitrate is no longer

available, carbon is the final electron acceptor and acidogenic

fermentation pathways appear. Sobieszuck and Szewczyk

(2006) demonstrated that for carbon substrates with a

reduction degree lower that 4.67, the critical denitrification

ratio, COD/N, is equal to 7.6 g O2 g�1 N, corresponding to a

NO3/COD ratio of 0.58 g g�1.

The prevention of H2S and mercaptan formation with

nitrate was studied and applied to urban wastewater (Bentzen

et al., 1995; Hobson and Yang, 2000). Garcıa de Lomas et al.

(2005) showed that the mechanism for preventing H2S is due

to the development of Thiomicrospira denitrificans, bacteria that

reduce nitrate by oxidising sulphide. The strategy and the

mechanisms involved in preventing VFA formation by nitrate

from wastewaters rich in carbon studied here, based on the

oxidation of carbon substrates by denitrification, differ from

the H2S prevention treatment with nitrate that does not

prevent H2S production by sulphate reduction bacteria but

that biologically oxidises H2S by reducing the nitrate.

3.2. Full-scale study of winery wastewater pondssupplemented or not with nitrate

A comparative study of the two winery wastewater evapora-

tion ponds, treated or not with nitrate, was carried out in

order to assess the benefits of nitrate for the prevention of

VFA formation in actual winery wastewater, at full scale, over

a period of time corresponding to industrial activity, and with

the natural pond microflora. Table 3 presents the character-

istics of the evaporation ponds (7000 m2 each) and of

preventive wastewater treatment through the addition of

calcium nitrate. The evaporation ponds were supplied with

Table 3 – Characteristics of the winery wastewater pondssupplemented or not with nitrate

With nitrate Without nitrate

Area (m2) 7000 7000

Wastewater intake (m3) 3521 2236

Nitrate supply Calcium nitrate None

Nitrate (g NO3 L�1) 13.3 0

NO3/COD (g g�1) 0.8 0

wastewaters produced during the winery’s maximum activity

period: grape harvest/vinification (29 September–14 Decem-

ber). The average composition of wastewaters discharged into

the ponds (Table 4) is characterised by a COD (16.6 g O2 L�1)

and a proportion of ethanol (49% of the COD) and sugars (31%

of the COD) close to those of the reconstituted winery

wastewaters (Table 2), confirming the representativeness of

the study carried out with reconstituted wastewater. One

pond was supplied with winery wastewaters enriched with

nitrate (3521 m3), with an average concentration of

13.3 g NO3 L�1, or a ratio of NO3/COD ¼ 0.8. The other pond

was supplied with winery wastewaters (2236 m3), without the

addition of nitrate (control).

As can be seen in Fig. 3, no accumulation of butyric and

valeric acids, nor any other VFA (results not shown) were

observed in the pond supplied with wastewaters enriched in

nitrate during the wastewater discharge period (212 months).

On the other hand, the accumulation of butyric acid

(143–997 mg L�1) and valeric acid (99–343 mg L�1) was ob-

served throughout the study period in the pond supplied

with wastewater without nitrate. Table 5 provides the VFA

composition in the two ponds at the end of the maximum

discharge period (14 December). In addition to the VFA

observed in reconstituted winery wastewaters (acetic, pro-

pionic, butyric, valeric and caproic acids), the formation of

ARTICLE IN PRESS

0

10

20

30

40

50

60

70

29/9

/05

6/10/

05

13/1

0/05

20/1

0/05

27/1

0/05

3/11/

05

10/1

1/05

17/1

1/05

24/11/0

5

1/12

/05

8/12

/05

15/1

2/05

VF

AC

OD/C

OD

(%

)

With NO3

Without NO3

Fig. 4 – Part of the VFA (propionic to heptanoic) in the COD of

winery wastewater ponds supplemented or not by nitrate

(each VFA is expressed as equivalent COD, and the sum of

VFA-COD is expressed as a percent of the COD measured).

0

1000

2000

3000

4000

5000

29/0

9/05

06/1

0/05

13/1

0/05

20/1

0/05

27/1

0/05

03/11/

05

10/11/

05

17/11/

05

24/11/

05

01/1

2/05

08/1

2/05

15/1

2/05

Nitra

te,

nitrite

(m

g L

-1)

0

1

2

3

4

5

6

7

8

CO

D (

gO

2 L

-1)

Nitrite

Nitrate

raw COD

Fig. 5 – Behaviour of COD, nitrate and nitrite in winery

wastewater pond supplemented by nitrate.

Table 5 – VFA composition of the winery wastewaterponds supplemented or not with nitrate

With nitrate Without nitrate

Acetic acid (mg L�1) n.d.a 822

Propionic acid (mg L�1) n.d. 211

Butyric acid (mg L�1) n.d. 354

Valeric acid (mg L�1) n.d. 178

Caproic acid (mg L�1) n.d. 83

Heptanoic acid (mg L�1) n.d. 69

a n.d.—not detected.

WAT E R R E S E A R C H 4 1 ( 2 0 0 7 ) 2 9 8 7 – 2 9 9 52992

heptanoic acid was observed in untreated pond of winery

wastewaters. None of these VFAs were detected in the

wastewater pond enriched with nitrate (Table 5).

VFA concentrations in the untreated pond were lower than

those obtained in batch tests with reconstituted wastewaters.

These differences can be explained by dilution due to autumn

rains (Mediterranean climate) and the eventual degradation

by the pond’s microflora. Fig. 4 illustrates the evolution of the

portion of VFA (propionic to heptanoic acids) in the COD of

the ponds. It shows that 30–60% of the COD of the untreated

pond consisted of VFA higher than acetic acid, revealing the

preponderance of malodorous compounds and acidogenic

fermentation. On the contrary, the portion of VFA in the

residual COD of wastewaters treated with nitrate was

negligible (Fig. 4).

Fig. 5 shows the evolution of the raw COD, of the nitrate and

nitrite in the pond supplied with wastewaters enriched in

nitrate. The COD of the pond never exceeded 5 g L�1 from the

beginning of the discharge period, decreased to 1 g L�1 during

the first month and remained close to this value throughout

the discharge period. In relation to the average COD of

discharged winery wastewaters (16.6 g L�1), the apparent

COD degradation rate in the pond treated with nitrate almost

reached 94%, revealing that the organic load of the waste-

water enriched in nitrate was oxidised and eliminated as

carbon dioxide through anaerobic respiration. Moreover, the

complementary analysis of the effluent in the treated pond

revealed a suspended matter content of 0.5 g L�1, and a

dissolved COD and BOD of less than 300 and 50 mg L�1,

respectively (data not shown). These levels of dissolved COD

and BOD are consistent with effluent standards in the

receiving environment and correspond to the characteristics

of winery effluents treated biologically. Therefore, discharge

in a receiving environment of an effluent treated with nitrate

in the pond could be considered after clarification. Related

research and experiments are being carried out at this time by

the laboratory.

The nitrate concentration in the treated pond developed

parallel to the COD. The nitrate totally disappeared by the end

of the first month and did not accumulate thereafter. The

transitory formation of nitrite in the pond was observed at the

beginning of the discharge period of wastewater enriched in

nitrate (first month) and occasionally thereafter when nitrate

input was greater. Under these conditions of strong denitrify-

ing activity, the degradation rate of nitrite into nitrogen was

less than the degradation rate of nitrate into nitrite. However,

the nitrite totally disappeared during the subsequent periods

and, at the end of the discharge of wastewaters enriched with

nitrate, no nitrate or nitrite accumulation could be observed

in the evaporation pond. The evolution of the ammoniacal

nitrogen concentration in the pond, ranging from 0 to

14 mg NH4+ L�1 (results not shown), did not reveal an accumu-

lation during nitrate degradation, confirming that the nitrate

was totally transformed into N2. The phenomenon of the

reductive dissimilation of nitrate into ammonia (Payne, 1973;

Samuelsson, 1985; Vigneron et al., 2005) did not take place in

the pond.

The absence of VFA accumulation for a ratio of NO3/

COD ¼ 0.8 and at full scale (3526 m3 of wastewater) confirmed

that anoxic respiration process (denitrification) were in

competition with the acidogenic pathways, although the

NO3/COD ratio was clearly lower than the stoichiometry.

Compared to previous studies of preventive treatments

based on concentrated nitric acid (a less expensive source of

nitrate), the implementation of the calcium nitrate treatment

(a non-corrosive product) proved itself to be technically

feasible both for determining the quantity of nitrate to be

ARTICLE IN PRESS

1

2

3

4

5

nov-05 dec-05

Levels

of odour

inte

nsity

Real effluent VFA solutions

Fig. 7 – Comparison of odour intensity of wastewater

without nitrate and a solution with the major odorous

VFA at the same concentration.

WAT E R R E S E A R C H 41 (2007) 2987– 2995 2993

used (concentrated calcium nitrate solution) and in relation

to human safety (neutral product) and facility maintenance

(no corrosion, ease of storage), matters of particular impor-

tance because wineries do not always have the adequate

structures or the means to enable them to use large

quantities of concentrated acids.

3.3. Contribution of VFA to the odour of winerywastewater and effect of nitrate enrichment

Among the VFA produced from winery wastewaters, acetic

acid is the one that has the least pronounced foul odour. All

the others, from propionic to heptanoic acid, have unpleasant

odours and low perception thresholds of only several dozen

mg m�3 (Le Cloirec et al., 1991). The fraction of malodorous

VFA (higher than acetic acid) represents 74% and 52% (w/w) of

the VFA produced from reconstituted wastewaters and in

winery wastewater ponds, respectively. In order to complete

chemical measurements of VFA and to assess the odours of

the VFA and the winery wastewater ponds, olfactory intensity

measurements were made on model VFA solutions and on

liquid samples taken from the ponds at two different times

during the discharge period (middle and end).

3.4. Relationship between odour intensity and VFA

The odour intensity of liquid samples, given in Fig. 6, clearly

shows that nitrate addition decreases the intensity level. If

treated wastewater is not odourless, the decrease from strong

to medium odour impression is significant even if the RSD of

these experiments was calculated as 0.5. The decrease is also

linked to a major transformation of odour quality. The odour

is very unpleasant without nitrate, whereas it is very

acceptable with nitrate. The unpleasant odour from liquid

samples, subsequently confirmed by gas samples (3/11/2005)

from the site and from the experiment in the wind tunnel,

reveal average levels of 3 and 2.75, respectively. This small

decrease may be due to a slight evolution of the liquid (1 day

between sampling at the site and the experiment in the wind

tunnel) and mainly to the volatilisation of some odorous

compounds during operations to transfer the effluent into the

tank of the wind tunnel. Because the procedure of smelling

air from a Tedlar bag is different than the procedure involving

direct sniffing of a flask, levels estimated on gas samples are

1

2

3

4

5

nov-05 dec-05

Levels

of odour

Inte

nsity

With nitrate Without nitrate

Fig. 6 – Comparison of odour intensity of wastewater with or

without nitrate during two different periods.

lower than those obtained directly on liquid samples that

contain many volatile odorous compounds.

All these results indicate that the decrease of odour is due

to VFA decrease obtained by the addition of nitrate to

wastewater. To confirm this close relationship between odour

intensity and VFA, synthetic solutions were made with the

same selected VFA concentrations as those in the pond

without nitrate for both periods (November and December).

The VFA selected were propionic, butyric and valeric acids.

The intensity levels given in Fig. 7 show the high degree of

correspondence between odours from real wastewater with-

out nitrate and, therefore, with VFA solutions. These three

acids are very odorous and present low perception thresholds

(around 80, range 4–50 and around 5mg m�3 for C3–C5 acids,

respectively, according to data from INERIS, a research

institute under the supervision of the French Ministry of the

Environment and Sustainable Development). These three VFA

mainly contribute to overall odour. When acetic acid, the

most abundant but potentially less odorous with a perception

threshold at 900mg m�3 (INERIS data), was added to these

synthetic solutions, it is more difficult to evaluate because of

the irritant aspect of this acid. This acid (C2) gives a sour

qualitative feeling without increasing the odour intensity.

Acetic acid is less responsible for odour intensity than others,

so the mixture of the three acids (C3–C5) is an easy way to

simulate the global intensity level of winery wastewaters.

4. Conclusion

This study revealed, both in the laboratory and at full scale,

the effect of the addition of an electron acceptor, nitrate in

this case, on the degradation of the components of winery

wastewaters through anaerobic respiration, to the detriment

of acidogenic and odorogenic fermentations, as well as its

effect on odour prevention. The following conclusions can be

drawn from this study:

�

VFA and particularly those acids ranging from propionic to

heptanoic that are very odorous and have low perception

levels are produced by the degradation of the organic

components of winery wastewaters by microflora. Butyric

acid is preferentially produced by fermentation of sugars,

ARTICLE IN PRESS

WAT E R R E S E A R C H 4 1 ( 2 0 0 7 ) 2 9 8 7 – 2 9 9 52994

glycerol and organic acids, whereas ethanol, the major

component of winery wastewaters, leads to the formation

of propionic, butyric, valeric, caproic and heptanoic acids.

�

Increasing NO3/COD ratios (0.4–1.2) successively affect VFA

production, going from the most highly reduced first

(caproic, valeric) to the least reduced (butyric followed by

propionic). The combination of the degradation of organic

compounds and nitrate, and the increase of the redox

potential shows that microflora activity is focused on an

anaerobic respiration process (denitrification), in competi-

tion with acidogenic fermentation pathways.

�

Full-scale experimentation (evaporation pond) with the

addition of calcium nitrate at a ratio of NO3/COD ¼ 0.8 in

winery wastewaters (3526 m3) revealed the absence of the

formation of VFA and the complete degradation of the

nitrate into nitrogen gas, with the considerable decrease of

the COD.

�

Olfactometric measurements underscored the relation-

ship between the VFA concentration in the wastewaters

and the odour, and confirmed the role of nitrate in relation

to odour prevention.

Acknowledgements

This study received financial support from the French Agency

for the Environment and Energy Management (ADEME) within

the framework of a request for proposals, ‘‘Odours and

Industries’’ (Contract No. 0374C0036). We thank Cave Anne

de Joyeuse in Limoux, France, for granting us access to its

wastewater treatment facilities and for its contribution to test

procedures. We are also grateful to the Societe YARA France

for its assistance with calcium nitrate treatment.

R E F E R E N C E S

ADEME, 2005. Pollutions Olfactives. Agence de l’Environnement et

de la Maıtrise de l’Energie, Dunod, Paris, 388pp.

Barbirato, F., Chedaille, D., Bories, A., 1997. Propionic acid

fermentation from glycerol: comparison to conventional

substrate. Appl. Microbiol. Biotechnol. 47, 441–446.

Bentzen, G., Smith, A.T., Bennett, D., Webster, N.J., Reinholt, F.,

Sletholt, E., Hobson, J., 1995. Controlled dosing of nitrate for

prevention of H2S in a sewer network and the effects on the

subsequent treatment process. Water Sci. Technol. 31 (7),

293–302.

Bolzonella, D., Innocenti, L., Pavan, P., Cecchi, F., 2001. Denitrifica-

tion potential enhancement by addition of anaerobic fer-

mentation products from the organic fraction of municipal

solid waste. Water Sci. Technol. 44 (1), 187–194.

Bories, A., 2006. Prevention et traitement des odeurs des effluents

vinicoles. In: Techniques de l’Ingenieur, Paris, G 1 960, pp.

1–12.

Bories, A., Sire, Y., Colin, T., 2005. Odorous compounds treatment

of winery and distillery effluents during natural evaporation in

ponds. Water Sci. Technol. 51 (1), 129–136.

Bories, A., Conesa, F., Boutolleau, A., Peureux, J-L., Tharrault, P.,

1998. Nouvelle approche et nouveau procede de traitement

des effluents vinicoles par fractionnement des constituants et

thermo-concentration. Revue Franc-aise d’Oenologie 171,

26–29.

Chiu, Y.C., Chung, M.S., 2003. Determination of optimal COD/nitrate ratio for biological denitrification. Int. Biodeterior.Biodegrad. 51, 43–49.

Chrobak, R.S., Ryder, R.A., 2006. Odors and control methods in winerywastewater treatment. In: IV International Specialized Confer-ence on Sustainable Viticulture: Winery Wastes and EcologicalImpact Management, 5–8 November, Vina del Mare, Chile.

Colin, T., Bories, A., Lavigne, C., Moulin, G., 2001. Effects of acetateand butyrate during glycerol fermentation by Clostridiumbutyricum. Curr. Microbiol. 43, 238–243.

Colin, T., Bories, A., Sire, Y., Perrin, R., 2005. Treatment andvalorisation of winery wastewater by a new bio-physicalprocess. Water Sci. Technol. 51 (1), 99–106.

Comanici, R., Gabel, B., Gustavsson, T., Markovitsi, D., Cornaggia,C., Pommeret, S., Rusu, C., Kryschi, C., 2006. Femtosecondspectroscopic study of carminic acid-DNA interactions. Chem.Phys. 325 (2), 509–518.

De Lucas, A., Rodrigues, L., Villasenor, J., Fernandez, F.J., 2005.Denitrification potential of industrial wastewaters. Water Res.39 (15), 3715–3726.

Duarte, E., Martins, M.B., Carbalho, E.C., Spranger, I., Costa, S.,1998. An integrated approach for assessing the environmentalimpacts of wineries in Portugal. In: Second InternationalSpecialized Conference on Winery Wastewaters, Bordeaux,5–7 May, CEMAGREF ed., pp. 61–69.

Elefsiniotis, P., Wareham, D.G., Smith, M.O., 2004. Use of volatilefatty acids from an acid-phase digester for denitrification.J. Biotechnol. 114 (3), 289–297.

Garcıa de Lomas, J., Corzo, A., Gonzalez, J.M., Andrades, J.A.,Iglesias, E., Montero, M.J., 2005. Nitrate promotes biologicaloxidation of sulfide in wastewaters: experiments at plant-scale. Biotehnol. Bioeng. 93 (4), 801–811.

Gottschalk, G., 1986. Bacterial Metabolism, second ed. Springer,pp. 234–236.

Guillot, J.M., Desauziers, V., Avezac, M., Roux, J.C., 2000. Char-acterization and treatment of olfactory pollution emitted bywastewater in winery Mediterranean region. Fresen. Environ.9, 243–250.

Hobson, J., Yang, G., 2000. The ability of selected chemicals forsuppressing odour development in rising mains. Water Sci.Technol. 41 (6), 165–173.

ITV, 2000. Les filieres d’epuration des effluents vinicoles. Nouvelleedition. ITV, France.

Lambert, J., Lecharpentier, D., 2006. Federations des CavesCooperatives de l’Aude et du Gard, Narbonne France. Perso-nal Communication.

Le Cloirec, P., Fanlo, J.L., Degorce-Dumas, J.R., 1991. Traitement desodeurs et desodorisation. Innovation 128, Paris.

Le Verge, S., Bories, A., 2004. Les bassins d’evaporation naturelle desmargines. Le Nouvel Olivier, 41, Septembre–Octobre, pp. 5–10.

Leyris, C., Maupetit, F., Guillot, J.M., Pourtier, L., Fano, J.L., 2000.Laboratory study of odour emissions from areal sources:evaluation of a sampling system. Analusis 28, 199–206.

Leyris, C., Guillot, J.M., Pourtier, L., Fanlo, J.L., 2005. Comparisonand development of dynamic flux chambers for the determi-nation of odorous compound emission rate from area source.Chemosphere 59, 415–421.

Min, K.S., Park, K.S., Khan, J.Y.J., Kim, Y.J., 2002. Acidogenicfermentation: utilization of wasted sludge as a carbon sourcein the denitrification process. Environ. Technol. 23 (3), 293–302.

OIV, 1999. Gestion des effluents de cave et de distillerie. Cahierscientifique et technique, 78pp.

OIV, 2005. World statistics, Third General Assembly of the OIV,June, Paris.

Payne, J.W., 1973. Reduction of nitrogenous oxides by micro-organisms. Bacteriol. Rev. 37, 409–452.

Picot, B., Cabanis, J.C., 1998. Caracterisation des effluents vini-coles: evolution des charges polluantes de deux caves

ARTICLE IN PRESS

WAT E R R E S E A R C H 41 (2007) 2987– 2995 2995

vinicoles du sud de la France sur deux cycles annuels. In:Second International Specialized Conference on WineryWastewaters, Bordeaux, 5–7 May, CEMAGREF ed., pp. 312–317.

Rochard, J., 2005. Traite de Viticulture Et D’œnologie Durables.Oenoplurimedia, Chaintre, Fr., 310pp.

Rochard, J., Kerner, S., Finazzer, E., 1996. Reglementationsrelatives aux effluents vinicoles dans les principaux paysproducteurs de vin. 76eme Assemblee Generale de l’OIV,Novembre.

Sage, M., Dauphin, G., Gesan-Guiziou, G., 2006. Denitrificationpotential and rates of complex carbon source from dairyeffluents in activated sludge system. Water Res. 40 (14),2747–2755.

Samuelsson, M-O., 1985. Dissimilatory nitrate reduction tonitrate, nitrous oxide, and ammonium by Pseudomonas putre-faciens. Appl. Environ. Microbiol. 50 (4), 812–815.

Smith, D.P., McCarty, P.L., 1988. Hydrogen partial pressure: effecton methanogenesis of ethanol and propionate in a perturbed

CSTR. In: Tilche, A., Rozzi, A. (Eds.), Fifth InternationalSymposium on Anaerobic Digestion. Monduzzi Editore, Bo-logna, pp. 75–80.

Sobieszuck, P., Szewczyk, K.W., 2006. Estimation of (C/N) ratio formicrobial dentrification. Environ. Technol. 25 (1), 103–106.

Sponza, D.T., Atalay, H., 2004. Influence of nitrate and COD onphosphorus, nitrogen and dinitrotoluene (DNT) removalsunder batch anaerobic and anoxic conditions. Anaerobe 10 (5),287–293.

Vigneron, V., Bouchez, T., Bureau, C., Mailly, N., Mazeas, L.,Duquennoi, C., Audic, J.M., Hebe, I., Bernet, N., 2005. Leachatepre-treatment strategies before recirculation in landfill bior-eactors. Water Sci. Technol. 52 (12), 289–297.

Wallace, R.J., Chaudhary, L.C., Miyagawa, E., McKain, N., Walker,N.D., 2004. Metabolic properties of Eubacterium pyruvativorans,a ruminal ‘hyper-ammonia-producing’ anaerobe with meta-bolic properties analogous to those of Clostridium kluyveri.Microbiology 150, 2921–2930.