Treatment and valorisation of winery wastewater by a new

biophysical process (ECCF)

T. Colin*, A. Bories**, Y. Sire** and R. Perrin*

* Evatex SAS, 26 rue Gay Lussac, 59147, Gondecourt, France

** I.N.R.A., Unite Experimentale de Pech Rouge, 11430, Gruissan, France

(E-mail: bories@ensam.inra.fr)

Abstract Taking account of the high specificity of the organic load of winery effluents, a new

biophysical treatment using the stripping of ethanol combined with a final concentration by

evaporation has been studied. Two options are proposed: full treatment and pre-treatment. The study

of the composition of winery wastewater has shown the large, dominant part of ethanol in the organic

load (75 to 99% of the COD). According to a linear correlation between COD and ethanol concentration,

the determination of ethanol concentration can be used to estimate the organic load of winery

wastewater. Full treatment by stripping and concentration at a pilot plant allows the separation of the

wastewater into highly purified water (COD elimination>99%), a concentrated alcoholic solution

usable as bio-fuel and a concentrated by-product. Stripping alone represents an advantageous

pre-treatment of winery wastewater. The purification rate reaches 78 to 85% and ethanol is

recovered. The process facilitates discharge into a sewage system in view of treatment with

domestic effluents and can also improve the efficiency of overloaded or old purification

plants. The economical approach of this method demonstrates its competitiveness in comparison

with biological treatments: low energy consumed, no sludge.

KeywordsWinery wastewater treatment; valorisation; ethanol

Introduction

Measurements of the organic load of winery wastewater have been increasingly reported for

many years (Mourgues and Maugenet, 1972; Maugenet, 1978) and particularly in recent

times (Rochard, 1993; Racault and Lenoir, 1994). The evaluation criteria of the organic load

such as COD and BOD are global parameters that quantify the organic load but do not

provide the composition of the pollutants. At this time, winery wastewaters are equated with

industrial or urban effluents and, as a result, they are treated with the same technologies:

spreading (Tournier, 1992; Laurens, 1996), natural or forced evaporation, aerobic degra-

dation, activated sludge (Forgeat et al., 1992; Racault, 1998) and methanisation (Bories,

1992a,b; Bories and Moulon, 1995).

Detailed studies of the composition of winery wastewater have revealed that ethanol and,

to a smaller extent and on a temporary basis, sugars (fructose, glucose) represent more than

90% of the organic load of winery effluent (Bories et al., 1998a,b; Bories, 2000). These

results prove that it is worthwhile to recover the organic load of winery wastewater rather

than dissipating it into sludge and CO2. Moreover, ethanol is a compound that is easy to

extract (stripping).

The purpose of this study concerns a new approach to the treatment of the organic load of

winery wastewater using a biophysical process known as ECCF1 (Evapo-concentration

with fractioned distillation). The first stage, when required, is the fermentation of the

sugars to ethanol. The second stage is physical. It involves the stripping of ethanol in

order to obtain a final separation of the wastewater into three phases: highly purified 99

Water

Science

andTechno

logyVol51No1pp99106

IWAPub

lishing2005

water, an alcoholic product with at least 40% ethanol and a residual concentrated

by-product.

The aim of this study was to examine the detailed composition and the flows of winery

wastewater over a one-year period. We then report some of the results obtained, on the one

hand, by the treatment of winery effluent using stripping and concentration to carry out a full

treatment and, on the other hand, by stripping alone (pre-treatment). The treatment

performances and economical and technical data are discussed.

Methods

The wastewater came from two French wineries: a large winery in the Languedoc

(200,000 hl) and a medium-size winery in Provence (40,000 hl).

Stripping and concentration were carried out with an industrial prototype (Evatex SAS,

Gondecourt, France) consisting of a steam-stripping column (14 plates, flow: 1,000 l/h)

combined with an evaporator (Falling film evaporator, flow: 500 l/h) equipped with

mechanical vapour recompression (Figure 1). Stripping and concentration were conducted at

atmospheric pressure, with additional steam supply.

Alcoholic fermentation took place on a laboratory scale using a 2-litre LSL Biolafitte

reactor, under anaerobic conditions (N2 bubbling) with mechanical agitation (100 rpm)

at room temperature (22C). Yeast inoculum was made from dehydrated active yeast(Saccharomyces cerevisiae).

The chemical oxygen demand (COD) was determined by chromic oxidation (French

AFNOR standard 90103). The composition of wastewater and organic compounds (glycerol,

organic acids, ethanol) was determined by HPLC. The apparatus included: an isocratic

pump, an in-line degasser, an automatic injector, a refractive index detector (Waters,

Milford, USA), an Aminex HPX-87H organic acid analysis column (Bio-Rad, USA) and

Millennium chromatography manager software (Waters). Carbohydrates were determined

by HPLC using a Waters carbohydrate column and identical Waters apparatus. Total

suspended solids (TSS) were determined by centrifugation and drying of the bottom at

105C. Conductivity was measured with a WTW conductivity detector (Weilheim,Germany).

The ratio of the organic load of a compound on COD was calculated using the theoretical

yield of total oxidation: (in g O2/g of compound) ethanol (2.06), sugars (1.07), malic acid

(0.72), tartaric acid (0.53), lactic acid (1.07), succinic acid (0.95), acetic acid (1.07) and

glycerol (1.22).

Figure 1 ECCF process diagram

100

T.C

olin

etal.

Results and discussion

Composition of wastewater from wineries

The dissolved organic load (CODd) and the concentrations of ethanol and sugars, during one

season, are presented in Figure 2. The total organic load fluctuated between 7 to 22 g O2/l

during the year. The CODdwas not connected with active periods: effluents produced during

racking polluted as much or more than effluents from harvesting or wine production. The

fluctuations of the CODd concentrations are the result of water consumption (washing) that

effects the dilution of the effluents to varying degrees. The ethanol concentration fluctuated

between 2.5 to 8 g/l (0.3 to 1 % vol/vol). Sugars were only present during the harvest phase.

They reached a maximum of 6 g/l at the beginning of the season.

As illustrated in Figure 2, the behaviour of ethanol concentration was close to the CODd.

A linear correlation can be obtained between ethanol and CODd (Figure 3). A very good

estimation of the CODd can be obtained by measuring the ethanol concentration.

In spite of wide variations of the organic load during the year, ethanol always represented

more than 75% (up to 99%) of the total CODd during the year, except during the harvest

(>55%) (Figure 4). Sugar ratios reached 40% of COD during the harvest but rapidly

decreased in October. Addition of ethanol and sugars always represented at least 75% of

the COD.

Averages of detailed analyses of winery wastewater (Table 1) over two seasons

confirmed that ethanol was the main part (80%) of the dissolved organic load. Glucose andfructose were the main sugars encountered and represented an average of 7% of the CODd

for the year. The organic load of organic acids was lower than 10% (including tartaric acid:

5.3%) of the CODd. Acetic acid concentration was low: 0.3 g/l (2.6% of the CODd).

Glycerol represented 3% of the CODd. The organic load and compound concentrations of the

0

5

10

15

20

25

09 10 12 01 03 05 06Month

COD

d (g

O2/l

)

0

5

10

Eth

ano

l - Su

gars

(g

/l)

Figure 2 Evolution of the dissolved COD (+), of the ethanol (m) concentration and of the sugars,glucose+fructose (&), of winery wastewater over a one-year period

0

10

20

30

40

0 5 10 15Ethanol (g/l)

CODd

(g

O2/l)

CODd = 2.7144 x [ethanol] R2 = 0.9517

Figure 3 Correlation between ethanol concentration and CODd of winery wastewater 101

T.C

olin

etal.

winery wastewater varied from one year to the next. However, no significant difference

could be observed in the distribution of the CODd between the compounds.

In view of the overall results, winery wastewater can be considered to be like a hydro-

alcoholic solution (12% vol/vol) containing secondary products (glycerol, organic acids)

and sugars from time to time.

Evapo-concentration with fractioned condensation

A treatment process using fractional distillation of the components based on the separation of

ethanol and secondary products was tested and validated. If necessary, sugars contained in

the effluent can be converted to ethanol.

Fractional distillation of winery effluents using stripping and concentration (full treatment)

The stripping of the ethanol and the concentration of the stripped wastewater was directly

carried out fromwinery waste. Table 2 shows the analytical data of the raw and treatedwater.

The treatment process produced highly purified wastewater (CODd: 88224 mg O2/l),

corresponding to a global COD elimination of 99%. Many process tests have demonstrated

that the organic load of the purified effluent was between 80 and 240 mg O2/l (Figure 5).

This final condensate was highly demineralised (conductivity: 3139 mS/cm) and free of

suspended matter and microorganisms. It can therefore be recycled as industrial water or be

discharged into a natural environment. A concentrated by-product can simultaneously be

recovered from the concentration stage. The volume of the concentrated by-products

represents about 5% of the initial wastewater. It was previously shown (Bories et al., 1999;

020406080

100

09 10 12 01 03 05 06Month

COD

ra

te (%

)

Figure 4 Ethanol COD (m) and sugar COD (&) rates of the CODd over a one-year period

Table 1 Mean composition of winery wastewater in 20012002 and 20022003

20012002 20022003

Values % CODd Values % CODd

pH 4.99 6.11TSS (g/l) 3.32 0.70COD raw (g O2/l) 14.6 10.9COD dissolved (g O2/l) 12.7 100 10.1 100Ethanol (g/l) 4.9 79.5 3.9 78.5Glucose (g/l) 0.35 2.9 0.30 3.2Fructose (g/l) 0.52 4.4 0.44 4.6Tartaric acid (g/l) 1.3 5.3 0.48 2.5Malic acid (g/l) 0.07 0.4 0.06 0.4Lactic acid (g/l) 0.16 1.3 0.20 2.1Succinic acid (g/l) 0.08 0.6 0.04 0.4Glycerol (g/l) 0.32 3.1 0.27 3.3Acetic acid (g/l) 0.30 2.6 0.19 2.1

102

T.C

olin

etal.

Bories, 2000) that the concentrated residues can be used for composting or spreading, and

animal feed is being considered. The recovery of an alcoholic phase (ethanol 3050%) was

obtained at the same time.

Treatment performances are optimal as soon as treatment begins and are not sensitive to

the variations of the organic load or the shutdown or start-up of plant operations.

Stripping (pre-treatment)

Stripping alone of the ethanol from winery wastewater resulted in the elimination of about

80% of the CODd (Table 3) represented by alcohol.

Many wineries are equipped with biological treatment plants. Most of these plants have a

small treatment capacity and it is for this reason that treatment is often not effective. In this

case stripping could improve the treatment. In fact, ethanol stripping could be considered as a

pre-treatment. The stripped effluent that consists of partially purified wastewater can be

more easily treated in a biological purification plant than crude effluent. It can also be

discharged and treated by an urban purification plant. Moreover, this system makes it

possible to highly reduce sludge production. The alcohol (3050% vol/vol) is recovered

under the same conditions as those used in the ECCF1 process.

Table 2 Analytical data of treatment of winery wastewater using the ECCF process

Parameter Before treatment After treatment Removal rate (%)

pH 3.653.82 4.033.88 CODd (mg O2/l) 23,37621,204 22488 99.2Ethanol (g/l) 7.917.28 0.060.02 99.5Conductivity (mS/cm) 2,9801,650 3931 98.4

0

5 000

10 000

15 000

20 000

25 000

30 000

35 000

40 000

1 2 3 4 5 6 7 8 9 10Test

COD

befo

re tr

eatm

ent

(mgO

2/l)

(mg

O 2/l)

0

125

250

375

500

625

750

875

1000

COD

afte

r tr

eatm

ent

Figure 5 COD before (^) and after (&) treatment during several tests of the ECCF process onwinery wastewater

Table 3 Analytical data of pre-treatment of winery wastewater by stripping

Parameter Before treatment After treatment Removal rate (%)

pH 3.82 3.80 Conductivity (mS/cm) 1,5071,650 1,8131,400 CODd (mg O2/l) 28,72821,204 5,1454,318 80Ethanol (g/l) 12.07.28 00.77 93

103

T.C

olin

etal.

Fermentation of wastewater

When sugars are present, the effluent can be fermented for optimal alcohol recovery. For

example (Figure 6), the kinetics of fermentation of wastewater containing 10.3 g/l of sugars

and 11.6 g/l of ethanol showed that sugars were totally fermented to ethanol (yield: 0.4 g

ethanol/g sugars). Glycerol (0.8 g/l) was produced whereas there was no production of

acetic acid.

Alcoholic fermentation of winery wastewater is different from the fermentation of the

must that involves low sugar and ethanol concentrations and variable pH. Effluent

fermentation conditions are favourable. There are no inhibitory effects like those produced

by ethanol.

Technical and economical considerations

The energy required for the stripping and concentration by mechanical vapour recompres-

sion is provided by a power input of 20 kWh/m3 of wastewater treated and a steam supply of

about 80 kg/m3. Total energy consumption (fermentation, stripping and concentration) is

estimated at 25 kWh/m3 (about 1.7 kWh/kg CODd for average winery wastewater) and is

independent of the organic load of the wastewater. The operating costs can be valued at

1.5 euro/m3 (0.06 euro/kWh). In comparison, biological treatment of winery effluents

requires a power supply, proportional to the organic load, of at least 45 kWh/m3 (3 kWh/kg

CODd) (ITV, 2000) and generates sludge (0.3 kg/kg BOD).

The ethanol recovered from the wastewater can be sold or used as fuel in steam generators

or boilers. In all cases, the recovery of ethanol (0.3 euro/L) can cover the costs of the ECCF1

process. Moreover, considerable savings can be made when treated water is recovered for

industrial uses (washing, boiler water, etc.) rather than discharged. For these reasons, the

ECCF1 process can be considered within the framework of sustainable development.

Concerning the investment costs for ECCF1, estimates for industrial plants have shown

levels similar to those of conventional biological treatments. The non-production of sludge,

which is difficult and expensive to separate and eliminate from the effluent, is a great

advantage of the biophysical process compared to biological treatments.

Conclusions

Results of the treatment of winery wastewater by the new biophysical process that involves

fermentation and stripping/concentration, obtained on an industrial scale at the winery site,

validated the technical and economical feasibility of the process. These results demonstrated

the high degree of water purification obtained and the valorisation potential of the

co-products (ethanol, concentrated fraction).

0

4

8

12

16

20

0 20 40 60 80 100Time (h)

Suga

rs - Et

hano

l (g/

l)0

200

400

600

800

1000

Gly

cero

l - Ac

etic

ac

id

(mg/

l)

Figure 6 Kinetics of the fermentation of winery wastewater. m Ethanol,& sugars

(glucose+fructose),* acetic acid, glycerol

104

T.C

olin

etal.

The study confirmed the high specificity of the composition of winery wastewater.

Ethanol is always the main component (2.5 to 8 g/l) and represents, on the average, about

80% of the organic load. Secondary compounds include residual sugars, glycerol and organic

acids. These data showed the uniqueness of winery wastewater compared to industrial and

urban effluents and led to new considerations concerning the treatments. First, they under-

scored the limits of characterising the organic load with only global parameters such COD,

BOD or TOC. Secondly, the prevalence of ethanol in the organic load made it possible

to propose new technologies and improved treatments and technical and economical

performances.

The physical separation of ethanol (stripping) and water (concentration) offers two

possibilities for the treatment of winery wastewater: full treatment and pre-treatment.

Full treatment of effluents by ethanol stripping and concentration results in a total

de-alcoholisation and produces highly purified water: COD

Bories, A., Conesa, F., Boutolleau, A., Peureux, J.L. and Tharrault, P. (1998). Nouveau procede de traitement

des effluents vinicole par fractionnement des constituants et thermo-concentration. In: Congre`s

international sur le traitement des effluents vinicoles, Bordeaux, France.

Bories, A., Conesa, F., Boutolleau, A., Peureux, J.L. and Tharrault, P. (1999). Nouvelle approche et nouveau

procede de traitement des effluents vinicole par fractionnement des constituants et thermo-concentration.

Revue francaise dnologie, 171, 2629.

Bories, A., Conesa, F. and Boutolleau, A. (1998b). Procedes et dispositifs biophysiques de traitement deaux

residuaires sucrees ou alcoolisees. French Patent 9804784, European Patent 9943004.4.

Forgeat, J.C., Jacquart, J.C. and Trousseau, M. (1992). Analyse des techniques depuration applicables a` la

profession vinicole. Revue des nologues, 61, 4548.

ITV. (2000). Les filie`res depuration des effluents vinicoles. ITV France/Onivins/Groupe technique des

effluents vinicoles, Paris, France.

Laurens, N. (1996). Traitement des effluents vinicoles par epandage ou evaporation. In: Journee technique:

Les caves vinicoles, Agence de leau RMC, Avignon, France.

Maugenet, J. (1978). Les eaux residuaires dans les industries viti-vinicoles, leur origine et les possibilites

de traitement. Revue francaise dnologie, 71, 2329.

Mourgues, J. and Maugenet, J. (1972). Les eaux residuaires de caves de vinification. Ind. Alim. Agric.,

3, 261273.

Racault, Y. (1998). Les procedes de traitement biologiques aerobies applicables aux effluents vinicoles.

In: Journee technique des effluents vinicoles, Agence de leau Adour Garonne, Senouillac, France.

Racault, Y. and Lenoir, A. (1994). Evolution des charges polluantes de deux caves vinicoles du Sud-Ouest

de la France sur un cycle annuel. In: Congres international du traitement des effluents vinicoles,

Narbonne Epernay, France, pp. 3743.

Rochard, J. (1993). Prise en compte des aspects lies a` lenvironnement au cours de lelaboration des vins.

In: La gestion des effluents vinicoles, Association francaise de genie rural, Montpellier, France.

Tournier, R. (1992). Cave de Raze`s: le succe`s de lepandage. Viti, Mars, pp. 4950.

106

T.C

olin

etal.