Electrodialysis as a useful technique for lactic acid separation from a model solution and a fermentation broth* 1

Page 1

Desalination 163 (2004) 361–372

0011-9164/04/$– See front matter © 2004 Elsevier B.V. All rights reserved

Presented at PERMEA 2003, Membrane Science and Technology Conference of Visegrad Countries (Czech Republic, Hungary,Poland and Slovakia), September 7–11, 2003, Tatranské Matliare, Slovakia.

*Corresponding author.

Electrodialysis as a useful technique for lactic acid separationfrom a model solution and a fermentation broth

Vera Hábová, Karel Melzoch*, Mojmír Rychtera, Barbora SekavováDepartment of Fermentation Chemistry and Bioengineering, Institute of Chemical Technology Prague,

Technická 5, CZ-16628 Prague 6, Czech RepublicTel. +420224354035; email: [email protected]

Received 25 July 2003; accepted 1 December 2003

Abstract

A two-stage electrodialysis (ED) method was used for lactic acid recovery. In the first step sodium lactate wasconcentrated with desalting electrodialysis using ion exchange membranes Ralex (Mega, Czech Republic). Thesecond step was the electroconversion of sodium lactate to lactic acid by electrodialysis with bipolar membranes(EDBM) Neosepta (Tokuyama Corp., Japan). The lactic acid was recovered from model solutions and from realfermentation broth as well. The trials with model solutions were focused on determination of the suitable conditionsfor electrodialysis experiments and on investigation of the time course under different conditions. The fermentationbroth from lactic acid fermentation had to be pretreated before electrodialysis experiments. The pretreatment consistedof ultrafiltration, decolourisation and removing of multivalent metal ions. In the first ED step the final lactateconcentration of 175 g/l was obtained and afterwards the final lactic acid concentration of 151 g/l was reached inthe second ED step.

Keywords: Lactic acid recovery; Electrodialysis; Desalting electrodialysis; Bipolar membrane; Pretreatment offermentation broth

1. Introduction

Lactic acid is one of the organic acids havinga wide use in a number of fields, e.g. food industry,beverage production, pharmaceutical industry,

chemical industry, medicine [1]. Exploitation oflactic acid for the production of biodegradablepolymers is one of the recent applications [2].Today, two thirds of the world production of lacticacid is done by the fermentation method [3]. Theconventional fermentation process produces

ě

Page 2

362 V. Hábová et al. / Desalination 162 (2004) 361–372

calcium lactate precipitate, which must be concen-trated by evaporation and reacidified by a strongacid [4]. The disadvantages of the conventionalfermentation process are a low reaction rate, anelaborate product recovery, a large amount of by-products and thereby negative impact on theenvironment. There are other possibilities forlactic acid recovery, but solvent extraction, directdistillation, adsorption and other relatively simplemethods have certain limitations, which obstructtheir wider use [5,6]. Electrodialysis is one of verypromising and perspective methods provided bythe rapid development of the membrane processes,especially the membranes in the 80’s and 90’s [7].

Electrodialysis is applied to remove salts fromsolutions or to concentrate ionic substances. Alarge number of existing applications of electro-dialysis has been described in the literature [8]. Aspecial type of electrodialysis process is electro-dialysis with bipolar membranes (EDBM) [9]. Abipolar membrane (BPM) consists of an anion andcation selective layer joined together. When anelectric field is applied, water splitting occurs atthe junction of the bipolar membrane, thus gener-ating protons and hydroxyl ions. The configura-tion of the EDBM process depends on the applic-ation (Fig. 1). Using a three-compartment config-

Fig. 1. The process scheme of three- and two-compartment EDBM (BPM — bipolar membrane, AEM — anion exchangemembrane, CEM — cation exchange membrane).

AEM

Water

NaOH

H+

+

CEM

Water NaX

HX

OH-

Na+ _

BPM BPM

X-

NaOH

H+

+

CEM

Water NaX

OH-

Na+_

BPM BPM

HX

uration, consisting of an anion, a bipolar and a cationexchange membrane as a repeating unit, the con-version of salt into a corresponding acid and baseis achieved. A two-compartment configuration,consisting of a BPM and a cationic (anionic) mem-brane allows the recovery of the base (acid).

Hongo proposed to use electrodialysis for insitu lactate recovery to reduce the productinhibition [10]. The obtained productivity wasthree times higher than that in non-pH controlledfermentation. However, fouling of anion exchangemembranes by cells was observed in the electro-dialysis fermentation. To solve this problem,Nomura used immobilized growing cells entrappedin calcium alginate [11]. The amount of lactic acidproduced by semicontinuous electrodialysis fer-mentation using immobilized cells was 8-timeshigher than that produced by non-pH controlledfermentation, but some free cells were found inthe solution. Czytko found that the electrodialysisunit could only be operated with cell free solutionsin order to prevent deposition of bacteria on themembranes [12]. To increase the productivity ofthe lactic acid fermentation and to reduce theamounts of effluents, Boyaval chose for the lacticacid fermentation a bioreactor with total cell re-cycling, which was coupled with an ultrafiltration

Page 3

V. Hábová et al. / Desalination 162 (2004) 361–372 363

module and an electrodialysis unit. The outletconcentration of lactate was stabilized at 85 g/l[13]. A similar system was studied by Yao. H2SO4

was used as a donor of protons and lactic acidwas the final product, reaching a concentration of90 g/l [14].

The two-stage electrodialysis method wasdescribed by Glassner [15]. Lee studied the two-stage electrodialysis method for lactic acid recovery,too[6]. In the desalting electrodialysis, 115 g lactateper litre was obtained in the concentrate, only 1 glactate per litre remained in the diluate, the currentefficiency was about 90% and the energy con-sumption for the lactate transfer from the diluateto the concentrate was 0.25 kWh/kg. In the secondstep, 88–93% of the total amount of lactate wasconverted to lactic acid, and the current efficiencywas about 80%. The total energy consumption forlactic acid recovery was in the range of 0.78–0.97 kWh/kg.

Heriban dealt with electrodialysis with doubleexchange [16]. A 4-times higher lactic acid con-centration in the continuous mode compared tothe lactic acid concentration was reached in theprocessed solution. A concentration of 236.8 g lacticacid per litre was obtained in the course of theexperiment with model solutions, and the energyconsumption was in the range of 1.3–2.3 kWh/kg.

Kim studied two- and three-compartmentEDBM for lactic acid recovery [17]. High volum-etric productivity (71.7 g/l.h) was reached.

Choi compared the conventional electro-dialysis consisting of cation and anion exchangemembranes and the ion substitution electrodialysisconsisting of only cation exchange membranes[18]. Both electrodialysis operations removed over95% of sodium ions from the feed solution.

Finally Bailly describes performances of aplant with a production capacity 5,000 tons oflactic acid per year which makes use of a two-stageelectrodialysis process. An electricity consump-tion of 1.8 kWh/kg of produced acid is reported[19].

2. Materials and methods

2.1. Chemicals

Sodium lactate (p.a. purity) was purchasedfrom Sigma. The other chemicals were fromLachema (Brno, Czech Republic). Demineralisedwater (resistivity of 18.2 MΩcm) was preparedfrom distilled water in a device Millipore–Qgradient (Molsheim, France). Real fermentationbroth was obtained from the continuous lactic acidfermentation using Lactobacillus plantarum L10as a producent strain (the strain was obtained fromthe Collection of the Department of Dairy and FatTechnologies, ICT Prague, Czech Republic). Thelactate concentration ranged from 17 to 88 g/l.Granulated active charcoal Purolite AC 20G(Purolite International Ltd., UK) was used fordecolourisation of fermentation broth. Chelatingresin Purolite S 940 (Purolite International Ltd.,UK) was used for the recovery of multivalentmetal ions.

2.2. Electrodialysis equipment

The electrodialysis laboratory unit BEL-500(Berghof, Germany) consisted of a control unit(adjustable outputs of voltage from 0 to 50 V andcurrent from 0 to 3.9 A), a measuring device (con-ductivity and voltage) and 3 independent circuitswith pumps and storage containers (for the diluate,the concentrate and the electrode solution). Themembrane stack ED 0 (Mega, Czech Republic)with 20 pairs of ion exchange membranes RalexCMH and Ralex AMH was used for the desaltingelectrodialysis experiments. The effective mem-brane area was 180 cm2, the distance between themembranes was 1 mm. Stack Type 500 with 4bipolar membranes Neosepta BP-1 and 5 cationexchange membranes Neosepta CMB (TokuyamaCorp., Japan) was used for EDBM, the distancebetween the membranes was 0.5 mm. The effect-ive membrane area was 57.6 cm2.

Page 4

364 V. Hábová et al. / Desalination 162 (2004) 361–372

2.3. Operating conditions

Electrodialysis experiments were carried outin a batch mode.

Desalting electrodialysis: The electrode solution(Na

2SO

4 — 25 g/l), the concentrate (sodium

lactate — initial concentration 3–127 g/l) and thediluate (sodium lactate — initial concentrationfrom 15 to 88 g/l) were circulated through the cor-responding compartments of the desalting stackwith the flow of 2.4 l/min. For the constant currentperiod, the voltage of 1.5 V per pair and the currentdensity of 7.8 mA/cm2 were used. For the constantvoltage period, the voltage of 18 V was used. Theexperiments were terminated when the lactateconcentration in the diluate dropped to 1–2 g/l.

For EDBM the following solutions were used:NaOH — 20 g/l (electrode solution), NaOH —1g/l (concentrate), sodium lactate 43 and 178 g/lor concentrate from desalting electrodialysis(diluate). The current density of 67.7 mA/cm2 andthe voltage of 12 V were applied. The circulationflow was 2 l/min. When the conductivity in thediluate reached 5 mS/cm, the experiments wereterminated.

2.4. Ultrafiltration

UF cartridge TFF 300 kD (Millipore, USA) wasused for ultrafiltration of the fermented broth.

2.5. Decolourisation, removing of multivalentmetal ions

These operations were carried out in a glasscolumn filled with the above-mentioned materials.The flow rate of the fermentation broth throughthe column was 1 bed volume/h (decolourisation)or 2 bed volumes/h (removing of multivalentmetal ions).

2.6. Analytical methods

The lactate was analysed by HPLC (Labora-torní prístroje Praha, Czech Republic): column—Ostion LG KS in H-cycle; refractometer detector

RIDK 101; the mobile phase — H2SO4 (c =0.005 mol/l), the flow rate of the mobile phase0.5 ml/min; the column temperature 85°C. Theconcentration of lactic acid and the concentrationof NaOH were analysed by titration by the stan-dard solution of NaOH (c = 0.025 mol/l), and HCl(c = 1 mol/l), respectively, using phenolphthaleinas the indicator. The colour intensity of the fer-mentation broth was measured by a spectro-photometer at a wavelength of 400 nm relative towater. Multivalent metal ions were determined byAAS method and the amount of biomass wasestimated by weight after drying at 105°C.

2.7. Calculations

The calculation equations were taken from Lee[6].

3. Results and discussion

The fermentation broth had to be pretreatedbefore the electrodialysis experiments due to highdemands of the electrodialysis membranes, espe-cially bipolar ones, for the quality of the feedsolutions used. The pretreatment consisted ofultrafiltration, decolourisation and removing ofmultivalent metal ions. The recovery and purifica-tion of lactic acid by the two-stage electrodialysismethod followed. The first step was desaltingelectrodialysis, and EDBM was the second step.The trials with model solutions were focused onthe determination of the parameters for the electro-dialysis experiments and on the investigation ofthe time course under different conditions.

3.1. Ultrafiltration

The fermentation broth from the lactic acidfermentation was ultrafiltrated to remove the cellsin order to prevent the deposition of bacteria onthe membrane surface and the creation of bacteriaclusters in the space between the membranes.

The fermentation broth contained 3.35 g/l ofbiomass, the cells were removed from it by a

ř

Page 5

V. Hábová et al. / Desalination 162 (2004) 361–372 365

spiral-wound module. The cell free permeate andconcentrate with 51.1 g/l of biomass were obtained.

3.2. Decolourisation

The fermentation broth was decolourised inorder to prevent a decrease in the electrodialysisefficiency due to the dye fixing on the membrane.

The flow of the fermentation broth through thecolumn significantly influenced the decolour-isation degree. When the flow rate was relativelyslow, only 1 bed volume/h, the amount of the fer-mentation broth corresponding to the 10-fold bedvolume was decolourised from 90%. If the de-colourisation was carried out in a batch mode, thedecolourisation degree was equal to the onementioned above, but more decolourising agentwas consumed.

3.3. Removal of multivalent metal ions

Multivalent metal ions (Ca, Mg, Fe, Zn, etc.)have to be removed from the fermentation brothto prevent the irreversible damage to the electro-dialysis membranes, especially bipolar ones. Theyrequire less than 1 mg/l multivalent metal ions(this value corresponds approximately to theconcentration of 0.03 mmol/l).

Chelating resin Purolite S 940 was used forremoving multivalent metal ions from the ferment-ation broth for the electrodialysis experiments(Table 1). The composition of the initial ferment-ation broth and of that treated broth by resin isgiven in Table 2. Depleted broth contained lessthan 1 mg/l of multivalent metals ions and thebipolar membrane could be used without hazardof the membrane irreversible damage.

3.4. Desalting electrodialysis

The initial measurements were focused on thedetermination of the limiting current. If the operat-ing current is beyond the limiting value, mem-branes can be irreversibly damaged. It has beenfound that the limiting current depends on thelactate concentration — the limiting current in-

Table 1Removal of multivalent metal ions (flow rate was 2 bedvolumes/h) in a column filled by Purolite S 940

Initial multivalent ions concentration, mmol/l 1.85 Final multivalent ions concentration, mmol/l 0.025 Treated bed volumes 42 Retained ion efficiency,% 98.6

Table 2Composition of initial and treated fermentation broth byion exchanger Purolite S 940

Initial fermentation broth

Treated fermentation broth

Ca, mg/l 28.3 0.39 Mg, mg/l 38.8 0.47 Fe, mg/l 0.64 0.04 Mn, mg/l 4.36 <0.02 Zn, mg/l 0.44 0.02 Na, mg/l 16500 16800 K, mg/l 935 868

creases with the increase of the lactate concentra-tion. The maximum limiting current density of8.8 mA/cm2 was found with the lactate concen-tration of 6.5 g/l. The limiting current density wasnot found at lactate concentrations higher than theone mentioned above. The working current den-sity was a bit lower than the determined maximallimiting current density to prevent the operationbeyond the limiting value. For a constant currentperiod, the current density of 7.8 mA/cm2 wasused. When the lactate concentration decreasedto the limiting value, the operating mode wasswitched from the constant current mode to theconstant voltage mode and the voltage wasadjusted to 18 V. The characteristic course of theelectrodialysis experiment is illustrated in Fig. 2.The results of the experiments performed areshown in Table 3.

It was found that the current density influencedsignificantly the transport rate of the lactate ionsthrough the membrane. A decrease of the current

Page 6

366 V. Hábová et al. / Desalination 162 (2004) 361–372

Tabl

e 3

Des

altin

g el

ectr

odia

lysi

s —

exp

erim

ents

with

the

mod

el s

odiu

m la

ctat

e so

lutio

n

DS

3 —

add

ition

of g

luco

se in

the

dilu

ate

(9 g

/l); D

S 4

— a

dditi

on o

f glu

cose

(9 g

/l) a

nd s

alts

(K2H

PO4,

MgS

O4,

MnS

O4,

sodi

um a

ceta

te, a

mm

oniu

m c

itrat

e)in

the

dilu

ate;

VD

0/VK

0 rat

io o

f the

initi

al d

iluat

e vo

lum

e to

the

initi

al c

once

ntra

te v

olum

e

Exp

erim

ent

Ope

ratin

g tim

e, m

in

Switc

hing

tim

e,

min

In

itial

dilu

ate

conc

entr

atio

n, g

/l Fi

nal d

iluat

e co

ncen

trat

ion,

g/l

Initi

al c

once

ntra

te

conc

entr

atio

n, g

/l Fi

nal c

once

ntra

te

conc

entr

atio

n, g

/l V

D0/

VK

0 W

ater

tran

spor

t in

dex,

mg/

l

DS

1 80

50

43

.7

0.8

3.0

34.1

1

7.08

D

S 2

160

100

39.6

0.

8 3.

9 55

.6

1.87

5.

27

DS

3 15

0 10

0 42

.4

1.2

4.1

59.5

1.

87

4.71

D

S 4

165

100

40.4

1.

2 4.

2 55

.9

1.87

5.

20

DS

5 75

15

15

.3

0.7

4.5

26.9

1.

87

4.78

D

S 6

240

120

36.6

0.

9 3.

9 12

1 7.

50

4.66

D

S 7

235

190

85.0

1.

0 74

.8

162

4.38

4.

74

DS

8 24

5 19

0 84

.4

1.5

103

173

4.38

4.

68

DS

9 11

0 85

31

.2

1.9

2.4

44.7

1.

80

4.56

D

S 10

12

0 90

33

.7

0.8

45.0

81

.9

1.80

4.

47

DS

11

120

100

33.4

1.

0 77

.0

104

1.80

5.

11

DS

12

120

97

34.5

1.

0 10

1 12

3 1.

80

5.26

D

S 13

13

0 10

5 34

.4

1.4

118

133

1.80

5.

58

DS

14

125

105

41.4

1.

6 12

7 14

9 1.

80

4.65

Tot

al p

roce

ss p

erio

d C

onst

ant c

urre

nt p

erio

d E

xper

imen

t

Lac

tate

tr

ansp

ort,

g

Rat

e of

lact

ate

tran

spor

t, g/

h C

urre

nt

effi

cien

cy,%

E

nerg

y co

nsum

ptio

n,

kWh/

kg

Rat

e of

lact

ate

tran

spor

t, g/

h C

urre

nt e

ffic

ienc

y,

%

Ene

rgy

cons

umpt

ion,

kW

h/kg

DS

1 83

.3

62.5

68

0.

34

79.6

67

0.

35

DS

2 16

1 60

.4

66

0.31

77

.7

65

0.32

D

S 3

180

72.1

74

0.

27

91.1

77

0.

26

DS

4 16

7 60

.8

60

0.32

67

.0

56

0.34

D

S 5

60.7

48

.6

76

0.29

80

.8

68

0.40

D

S 6

213

53.2

84

0.

32

91.7

86

0.

33

DS

7 29

2 74

.5

73

0.24

88

.8

75

0.22

D

S 8

288

70.4

71

0.

26

83.1

70

0.

26

DS

9 13

1 71

.7

71

0.31

83

.4

70

0.31

D

S 10

14

6 72

.8

73

0.27

85

.9

72

0.27

D

S 11

14

4 72

.2

67

0.30

81

.7

69

0.29

D

S 12

15

0 74

.9

70

0.28

84

.3

71

0.27

D

S 13

14

8 68

.5

66

0.31

77

.5

66

0.31

D

S 14

17

7 85

.2

80

0.25

95

.2

81

0.24

Page 7

V. Hábová et al. / Desalination 162 (2004) 361–372 367

Fig. 2. The time course of the desalting electrodialysis — experiment DS 9.

0

10

20

30

0 20 40 60 80 100

Time (min)

Vo

ltag

e (V

)

0

2

4

6

8

10

Cu

rren

t d

ensi

ty (

mA

/cm

2 )

Voltage

Current density

0

10

20

30

40

50

0 20 40 60 80 100

Time (min)

Lac

tate

co

nce

ntr

atio

n (

g/l)

Diluate

Concentrate

density by 30% resulted in the prolongation ofthe constant current period by 60% and in asignificantly decreased rate of the lactate transport.The value of the lactate concentration in the con-centrate is a very important factor for the success-ive recovery step — the electroconversion. Thedegree of the concentration (the ratio of the finalconcentration in the concentrate to the initialdiluate (feed) concentration) can be influenced bythe increase of ratio of the initial diluate volumeto the initial concentrate volume. During thecourse of the electrodialysis experiments theconcentrate and diluate volumes changed due towater passage through the membranes simul-taneously with the lactate ions by electroosmosis.Water transport index (the volume of water passedto the amount of the transported lactate) was in

the range of 4.6–7.1 ml/g in our experiments. Theinfluence of other components (glucose and salts,which are commonly present in the fermentationbroth) on the electrodialysis run was studied.While the addition of glucose in the diluate didnot show any effect, the addition of salts resultedin the decreasing current efficiency. The currentefficiency was in the range of 60–80% during theelectrodialysis experiments with the modelsolutions and the energy consumption was about0.3 kWh/kg.

The electrodialysis experiments with thepretreated fermentation broth were carried out andthe results are shown in Table 4. The comparisonof the lactate recovery from the model solutionsand from the fermentation broth has shown thatin the case fermentation broth with lower initial

Page 8

368 V. Hábová et al. / Desalination 162 (2004) 361–372Ta

ble

4D

esal

ting

elec

trod

ialy

sis

— e

xper

imen

ts w

ith th

e pr

etre

ated

ferm

enta

tion

brot

h

Exp

erim

ent

Ope

ratin

g tim

e, m

in

Switc

hing

tim

e,

min

In

itial

dilu

ate

conc

entr

atio

n, g

/l Fi

nal d

iluat

e co

ncen

trat

ion,

g/l

Initi

al c

once

ntra

te

conc

entr

atio

n, g

/l Fi

nal c

once

ntra

te

conc

entr

atio

n, g

/l V

D0/

VK

0 W

ater

tran

spor

t in

dex,

mg/

l

DM

1

120

40

18.1

1.

3 3.

9 29

.2

1.87

7.

10

DM

2

165

50

17.3

1.

3 28

.7

45.4

1.

87

9.04

D

M 3

19

5 70

22

.2

1.2

43.0

60

.6

1.87

9.

07

DM

4

195

150

36.3

2.

0 2.

8 10

8 7.

50

5.58

D

M 5

18

0 15

0 36

.9

2.8

104

152

7.50

5.

36

DM

6

215

170

36.6

0.

9 4.

8 11

3 7.

50

5.45

D

M 7

19

5 15

0 36

.6

1.2

107

147

7.50

5.

76

DM

8

215

170

88.3

3.

7 96

.4

185

4.38

4.

18

DM

9

220

170

85.8

1.

2 93

.6

175

4.38

4.

50

Tot

al p

roce

ss p

erio

d C

onst

ant c

urre

nt p

erio

d E

xper

imen

t

Lac

tate

tr

ansp

ort,

g

Rat

e of

lact

ate

tran

spor

t, g/

h C

urre

nt

effi

cien

cy,%

E

nerg

y co

nsum

ptio

n,

kWh/

kg

Rat

e of

lact

ate

tran

spor

t, g/

h C

urre

nt e

ffic

ienc

y,

%

Ene

rgy

cons

umpt

ion,

kW

h/kg

DM

1

72.7

36

.4

53

0.41

62

.3

52

0.46

D

M 2

66

.9

24.3

55

0.

54

52.4

64

0.

57

DM

3

81.3

25

.0

39

0.53

48

.2

41

0.53

D

M 4

20

6 63

.5

66

0.32

76

.7

67

0.32

D

M 5

20

5 68

.4

69

0.32

76

.0

68

0.33

D

M 6

21

7 60

.5

64

0.34

71

.5

64

0.35

D

M 7

19

8 61

.0

64

0.34

73

.0

65

0.34

D

M 8

30

0 83

.6

82

0.21

97

.6

83

0.20

D

M 9

29

5 80

.4

81

0.22

97

.4

82

0.21

Page 9

V. Hábová et al. / Desalination 162 (2004) 361–372 369

lactate concentrations the transport rate decreased,the energy consumption increased and the currentefficiency decreased. Nevertheless at the higherlactate concentration in the fermentation brothvery good agreement with the model solutions wasfound.

Fig. 3 shows the time course of the two-levelelectrodialysis with the fermentation broth. Theinitial lactate concentration was 36.6 g/l, and thefinal concentration of 146 g/l was obtained in theconcentrate stream. About 1 g/l lactate remainedin the diluate at the end of each level. The concen-tration of lactate obtained during the two-levelelectrodialysis was about 4-times higher than thatin the fermentation broth. The current efficiencywas 64% and the energy consumption was0.34 kWh/kg.

Fig. 3. The time course of the two-level desalting electrodialysis with the pretreated fermentation broth.

0

10

20

30

40

0 50 100 150 200 250 300 350 400

Time (min)

Vo

ltag

e (V

)

0

2

4

6

8

10

Cu

rren

t d

ensi

ty (

mA

/cm

2 )

Current density

Voltage

0

50

100

150

0 50 100 150 200 250 300 350 400

Time (min)

Lac

tate

co

nce

ntr

atio

n (

g/l)

Concentrate

Diluate

addition of the fresh fermentation broth into diluate circuit

3.5. Electrodialysis with bipolar membranes

Two-compartment EDBM was the second step,where sodium lactate was converted to lactic acidby removing Na+ through the cationic membrane.Fig. 4 shows an example of EDBM.

The experiments were carried out at a constantcurrent density (67.7 mA/cm2), which was higherthan that during the desalting electrodialysis. Theresults of the experiments with the model lactatesolution are summarized in Table 5. The amountof converted sodium lactate is proportional to theamount of supplied charge. A final lactic acid con-centration of 29.7–156.8 g/l (c = 0.33–1.74 mol/l),corresponding to 85–98% of conversion wasobtained; the energy consumption was about1.1 kWh/kg. The final base concentration was

Page 10

370 V. Hábová et al. / Desalination 162 (2004) 361–372

Fig. 4. The time course of EDBM using for electroconversion of sodium lactate to lactic acid.

0

5

10

15

20

0 50 100 150 200 250 300 350 400

Time (min)

Vo

ltag

e (V

)

0

10

20

30

40

50

60

70

Cu

rren

t d

ensi

ty (

mA

/cm

2 )

Current density

Voltage

0

20

40

60

80

100

120

0 50 100 150 200 250 300 350 400Time (min)

Co

nce

ntr

atio

n (

g/l)

Lactic acid

NaOH

c = 0.35–1.45 mol/l and the current efficiency was70–80%.

The results confirmed the possibility of sodiumlactate conversion to lactic acid with the use ofEDBM.

Table 6 shows the results of the two experi-ments with sodium lactate which was recoveredfrom the fermentation broth and concentrated bythe desalting electrodialysis. The values obtainedare very similar to those with the model lactatesolutions. A final lactic acid concentration of 121and 151 g/l (c = 1.34 – 1.67 mol/l), correspondingto 92% and 95% of conversion was obtained; theenergy consumption was about 1 kWh/kg. Thefinal base concentration was c = 1.07 – 1.32 mol/land the current efficiency was 70–80%.

4. Conclusions

Lactic acid separation from the fermentationbroth was studied comprehensively and the resultsobtained showed very good agreement with theliterature sources [6,19].

The results confirm that the two-stage electro-dialysis is a suitable and efficient technique forthe recovery of lactate ions from the pretreatedfermentation broth and subsequent conversion intolactic acid with respect to environmental aspects.

In the first ED step the final lactate concen-tration up to 175 g/l was obtained and the finallactic acid concentration of 151 g/l was reachedin the second step. Total required energy in bothelectrodialysis processes consisting of energy

Page 11

V. Hábová et al. / Desalination 162 (2004) 361–372 371

Table 5EDBM — experiments with model sodium lactate solutions

Experiment Operating time, min

Initial lactate concentration, g/l

Final lactic acid concentration, g/l

Final lactate concentration, g/l

Final NaOH concentration, g/l

Na+ transport, g

WS 1 82.5 43 30 0.5 14.0 16 WS 2 406 178 157 1.3 58.1 71 WS 3 200 61 50 1.5 21 23 WS 4 210 66 53 2.0 22 27 WS 5 180 69 54 2.3 21 26 WS 6 300 85 63 1.8 26 32 WS 7 260 100 86 1.0 33 42 WS 8 345 139 127 2.1 40 53 WS 9 420 175 148 5.1 57 70

Experiment Water transport index, ml/g

Rate of Na+

transport, g/h

Rate of lactic acid conversion, g/h

Current efficiency, %

Energy consumption, kWh/kg

% of conversion

WS 1 3.8 11.6 45.3 92 0.91 85 WS 2 3.0 10.5 41.0 78 0.84 98 WS 3 3.3 6.9 26.9 73 1.20 95 WS 4 3.4 7.6 29.8 77 1.14 93 WS 5 6.7 8.7 34.2 73 1.20 87 WS 6 2.8 6.5 25.4 61 1.38 92 WS 7 3.9 9.6 37.6 77 1.10 96 WS 8 4.2 9.3 36.4 72 1.10 89 WS 9 2.3 10.0 39.2 76 0.98 94

Table 6EDBM — experiments with lactate recovered from the fermentation broth

Experiment Operating time, min

Initial lactate concentration, g/l

Final lactic acid concentration, g/l

Final lactate concentration, g/l

Final NaOH concentration, g/l

Na+ transport, g

WM 1 390 145 121 4.0 43 53 WM 2 430 171 151 2.6 53 71

Experiment Water transport index, mg/l

Rate of Na+

transport, g/h

Rate of lactic acid conversion, g/h

Current efficiency, %

Energy consumption, kWh/kg

% of conversion

WM 1 3.9 8.2 32.2 66 1.16 92 WM 2 3.3 9.9 38.9 79 0.95 95

Page 12

372 V. Hábová et al. / Desalination 162 (2004) 361–372

consumption for lactate transfer and for itselectroconversion to lactic acid was about 1.5 kWhper 1 kg of lactic acid obtained.

Acknowledgement

The work was supported by EUREKA Σ! 1820BIOLACTATE and grant No. CEZ:J19/98:223300005 of the Ministry of Education, Youthand Sports of the Czech Republic.

References[1] T.B. Vick Roy, Lactic acid. In: M.M. Young, Ed., Com-

prehensive Biotechnology, Pergamon Press, NewYork, 1985, pp. 761–775.

[2] R. Datta, S.P. Tsai, P. Bonsignore, S.H. Moon and J.R.Frank, Technological and economic potential ofpoly(lactic) and lactic derivates. FEMS MicrobiologyReviews, 16 (1995) 221–231.

[3] J.S. Ka cák, J. Komínek and M. Roehr, Lactic acid.In: M. Roehr, Ed., Biotechnology. Vol. 6, Products ofPrimary Metabolism, VCH, Weinheim, 1996, pp. 294–306.

[4] K. Buchta, Lactic acid. In: H. Dellweg, Ed., Biotech-nology. Vol. 3, Verlag Chemie, Weinheim, 1983, pp.410–417.

[5] V. Heriban, J. kára, E. turdík and J. Ilavsk , Isolationof free lactic acid using electrodialysis. Biotech. Tech.,7 (1993) 63–68.

[6] E.G. Lee, S.-H. Moon, Y.K. Chang, I.-K. Yoo and H.N.Chang, Lactic acid recovery using two-stage electro-dialysis and its modelling. J. Membr. Sci., 145 (1998)53–66.

[7] J.A. Asenjo, Separation Processes in Biotechnology.Marcel Dekker, New York, 1990.

[8] M. Mateus, J.A.L. Santos and J.M.S. Cabral, Mem-brane separation processes. In: J.F. Kennedy and

J.M.S. Cabral, Eds., Recovery Processes for BiologicalMaterials. John Wiley & Sons, Chichester, 1993, pp.177–222.

[9] A.J.B. Kemperman, Handbook on Bipolar MembraneTechnology. Twente University Press, Enschede,2000.

[10] M. Hongo, Y. Nomura and M. Iwahara, Novel methodof lactic acid production by electrodialysis fermenta-tion. Appl. Environ. Microbiol., 52 (1986) 314–319.

[11] Y. Nomura, M. Iwahara and M. Hohgo, Lactic acidproduction by electrodialysis fermentation usingimmobilized growing cells. Biotechnol. Bioeng., 30(1987) 788–793.

[12] M. Czytko, K. Ishii and K. Kawai, Continuous glucosefermentation for lactic acid production: recovery ofacid by electrodialysis. Chem. Ing. Tech., 59 (1987)952–954.

[13] P. Boyaval, C. Corre and S. Terre, Continuous lacticacid fermentation with concentrated product recoveryby ultrafiltration and electrodialysis. Biotechnol. Lett.,9 (1987) 207–212.

[14] P.-X. Yao and K. Toda, Lactic acid production inelectrodialysis culture. J. Gen. Appl. Microbiol., 36(1990) 111–125.

[15] D.A. Glassner and R. Datta, Process for Productionand Purification of Lactic Acid. EP 0 393 818 A1,1990.

[16] V. Heriban, J. kára, E. turdík and and J. Ilavsk ,Isolation of free lactic acid using electrodialysis.Biotech. Tech., 7 (1993) 63–68.

[17] Y.H. Kim and S.-H. Moon, Lactic acid recovery fromfermentation broth using one-stage electrodialysis. J.Chem. Technol. Biotechnol., 76 (2001) 169–178.

[18] J.-H. Choi, S.-H. Kim and S.-H. Moon, Recovery oflactic acid from sodium lactate by ion substitutionusing ion-exchange membrane. Sep. Purif. Technol.,28 (2002) 69–79.

[19] M. Bailly, Production of organic acids by bipolarelectrodialysis: realizations and perspectives. Desa-lination, 144 (2002) 157–162.

č