Suite aux résultats présentés au chapitre 7, l’efficacité de l’électrodialyse combinée avec des membranes d’ultrafiltration à séparer un mélange d’oligomères de chitosane a été améliorée par la combinaison du pH du milieu et du temps de traitement.

Il nous a paru pertinent d’étudier l’effet du champ électrique qui est la force motrice du procédé et la vitesse de circulation (débit) des solutions sur le taux d’électromigration des oligomères de chitosane et la possibilité d’exploiter les résultats pour les séparer. Basée sur les résultats obtenus aux chapitres 6 et 7, une membrane d’ultrafiltration avec un seuil de coupure (MWCO) de 10000 Da a été utilisée. Dans cette partie de la thèse, le pH du milieu a été fixé à 4 et le temps de traitement à 4h.

La planification de ce travail et la rédaction de cet article ont été réalisées par l’étudiant Mohammed Aider qui est aussi auteur principal. Il a été supervisé avec beaucoup d’attention par le Dr. Laurent Bazinet qui a contribué au succès de ce travail par sa grande implication. Serge Brunet de ISM Biopolymer a contribué au succès de cet article par ses corrections et commentaires. Ils sont co-auteurs de ce travail. Les résultats présentés dans ce chapitre sont en attente de soumission dans Journal of Membrane Science.

8.2. Materials and methods

8.2.1. Chemicals and materials

8.2.1.1. Chitosan oligomers

A mixture of three chitosan oligomers (dimer, trimer and tetramer), lot No W-040720 was gracefully provided by ISM Biopolymer (Granby, Québec, Canada). Standards of chitosan oligomers (dimer, trimer and tetramer) were purchased from Seikagaku Corporation (Japan) (Cat. No: 800105).

8.2.1.2. Chemicals

All chemicals were of analytical grade. KCl and NaCl were purchased from Laboratoire MAT (Montreal, Qc, Canada) and HCl was from Fischer Scientific (Nepean, ON, Canada). Acetonitrile of HPLC grade was purchased from EMD Chemicals Inc, (Gibbstown, NJ, USA).

8.2.1.3. Membranes

Ultrafiltration membrane with molecular weight cut-off (MWCO) of 10000 Da from Spectrum Laboratories Inc (Rancho Dominguez, CA, USA) and anionic exchange membranes AMX-SB (Tokuyama Soda Ltd, Japan) were used. The main characteristics (thickness, electrical conductivity) of these membranes are presented in Table 8.1.

Table 8.1: Electrical conductivity (κ) and thickness (l) of the 10000 Da MWCO ultrafiltration (UFM) and anion exchange (AEM) membranes before and after electrodialysis with ultrafiltration membrane (EDUF) treatment.

8.2.2. Experimental

8.2.2.1. Electrodialysis with ultrafiltration membrane (EDUF) configuration

A MicroFlow type electrodialysis cell (ElectroCell AB, Täby, Sweden) with an effective area of 10 cm² was used. The EDUF cell configuration [Anode-AEM1-MUF-AEM2-Cathode] defines three closed compartments (Figure. 8.1). The first compartment was for NaCl (20 g/L) as electrode rinse solution. The second one was for the chitosan oligomers solution and the third one for KCl (2 g/L) solution. Each compartment was connected to a separate external reservoir to allow recirculation of the solution. The solutions were circulated using three centrifugal pumps and the flow rates in each compartment were controlled by flow meters. The anode was a dimensionally stable electrode (DSA) and the cathode was a 316 stainless steel electrode. For all experiments, the anode/cathode voltage difference was supplied by a variable 0–60V power source (HPD 30-10SX, Xantrex, Burnaby, BC, Canada). Once the EDUF cell was assembled, electrodes spacing was 2.0 ± 0.1 cm.

Figure 8.1: Schematic representation of the electrodialysis with ultrafiltration membrane cell used.

8.2.2.2. Protocol

Electroseparation by electrodialysis with ultrafiltration membrane (EDUF) treatments were carried-out on 200 mL of chitosan oligomer solution in a batch process at three applied voltages (5, 10 and 20 V) which corresponded to three average electric field strengths of 2.5, 5 and 10 V/cm. Solutions flow velocities of 2.77, 8.33 and 13.88 cm/s which corresponded to 100, 300 and 500 mL/min, respectively, were used. The chitosan oligomer mixture solution was obtained by dissolving chitosan oligomer mixture in HPLC grade water to obtain a final concentration of 3% (w/v). Its initial pH was adjusted to pH 4 by adding 1 N HCl. This pH value was selected following previous work on the effect of the pH value on chitosan oligomer electrophoretic mobility (Aider et al., 2006b). The pH of the 200 mL KCl solution in the adjacent compartment was also adjusted and was the same one as in the chitosan oligomer solution compartment. Samples of chitosan oligomer and KCl solutions were drawn at the beginning of the process before applying the external electric field and every 60 min. during the treatment. EDUF was stopped after 4 h of treatment. During treatment, pH and conductivity values were recorded in the KCl and chitosan oligomer solutions. Measurements of the current intensity were used for the calculation of the total system resistance. Thickness and electrical conductivity of the membranes were measured before and after each treatment to check for their integrity and verify their potential fouling. All samples were analysed by HPLC to determine the electromigration rate of the three chitosan oligomers.

8.2.2.3. Analyses

8.2.2.3.1. Chitosan oligomer profiles

Chitosan oligomers in the treated solution and in KCl (migration compartment) were analysed using a Shodex Asahipak NH2P-50 column connected to a guard column NH2P50G 4A (Shodex Separation and HPLC Group, Kanagawa, Japan). The apparatus used was a Waters 715 equipped with an RI (refractive index) differential refractometer (Model 410, Waters Corporation, Milford, MA, USA). The eluent was CH₃CN/H₂O (70/30 v/v) with a flow rate of 1.0 mL/min at 25°C, according to the instructions of Sodex Asahipak, Inc. The electro-migration rate of each oligomer at a given time was determined using the following equation:

(8.1)

where: EMR _i ( τ ), electromigration rate of the i-^thchitosan oligomer at a given time (τ) (dimer, trimer or tetramer), (%); S _i ( τ ), peak area of the chitosan oligomer (dimer, trimer or tetramer ) at a given time τ (h); and S _i ( τ _o ), peak area of the i-^th chitosan oligomer (dimer or trimer or tetramer ) at τ = 0 h in the feed chitosan oligomer solution.

8.2.2.3.2. pH and conductivity of solutions

During each treatment, pH and conductivity of chitosan oligomer solution were measured using combined pH-conductivity meter ( Model SR 601 C, SympHony, VWR Scientific Products, USA) with probe number 14002-802 and cell constant k = 1 cm^-1. In KCl solution, pH and conductivity were measured using pH-meter (Model VWR-SP20) and conductivity-meter (Model VWR-SR 60 IC, SympHony, VWR Scientific Products, USA) with cell constant k = 1 cm^-1, respectively.

8.2.2.3.3. Global system resistance

The global system resistance R ( W ) was calculated from the values of current intensity I (A) and the applied voltage U (V) using the Ohm’s law:

(8.2)

8.2.2.3.4. Membrane thickness

A Mitutoyo Corp. IDC type digimatic indicator with an absolute encoder (Model ID-C112 EB, Kanagawa-Ken, Japan), specially devised for plastic film thickness measurements, with a resolution of 1 mm and a range of 12.7–0.001 mm was used (Bazinet & Araya-Farias, 2005). The digimatic indicator was equipped with a 10 mm diameter flat contact point. The membrane thickness was measured in the active membrane area at different locations.

8.2.2.3.5. Membrane electrical conductivity

For each membrane, electrical conductivity was measured with specially designed cell (conductivity clip) from the Laboratoire des Matériaux Echangeurs d’Ions (Créteil, France) as previously described (Bazinet & Araya-Farias, 2005). An YSI conductivity meter, Model 3100, was used with an YSI immersion probe (Model 3252, cell constant k=1 cm⁻¹) (Yellow Springs Instrument Co., Yellow springs, OH, USA). The membrane conductivity (S.cm^-1) was calculated using the following equation adapted from Lteif et al. (1999) and Lebrun et al. (2003).

(8.3)

where: l: membrane thickness (cm); R _m :transversal membrane electric resistance (W) and A: electrode area (1 cm²).

8.2.2.3.6. Transport number calculation

The transport number of a given ionic specie is defined as a fraction of the total electrical current carried by this ionic specie in a solution or through a membrane (Hirata et al., 1992). Transport number of a given ion or molecule can be calculated by the following equation (Shaposhnik et al., 2000):

(8.4)

where is the transport number, Zi the molecule valence, F is the Faraday’s constant (96500 C.mole^-1), Ji the molecule flux through the membrane (mole. m^-2. s^-1) and i is the current density (A.m^-2).

8.2.2.3.7. Limiting current density

According to Cowan & Brown (1959), at the beginning of the EDUF treatment (τ = 0 h), the limiting current density determination was carried-out by quickly increasing the voltage applied to the electrodes of the electrodialysis with ultrafiltration membrane (EDUF) cell. The applied voltage was varied by 2 V from 0 up to 60 V. The current intensity was measured according to the corresponding applied voltage. The global system resistance (Eq. 8.2) was then plotted versus the reciprocal of the current intensity (1/I). From the method described by Cowan & Brown (1959), at the inflection point on this graph, the current intensity divided by the active membrane area would be considered as to be the limiting current density of the system.

8.2.2.3.8. Statistical analysis

The experimental design was a complete randomized design with three repetitions. The applied voltage and solutions velocities were of three levels. Each treatment was carried-out in triplicate for a total of 27 experiments. A multiple analysis of variance (MANOVA) was used for data analyses of the global system resistance and electro-migration rates of the chitosan oligomers as function of time, applied voltage and solution velocity. SAS software (V9.0, SAS Institute Inc., Cary, N.C.) was used. A 5% significance level was chosen.

To evaluate the membrane fouling and electric current efficiency, Statgraphics software V. 5.1 (Stat Point, Inc, Orlean, VA, USA), was used for comparison between membrane characteristics (thickness and electrical conductivity) before and after each electro-separation treatment and for comparison between different values of the current efficiency.

8.3. Results and discussion

8.3.1. Chitosan oligomers electromigration rates

Statistical analysis of the data was carried-out to check the degree of significance of the studied variables; applied average electric field (E) and solution flow velocity (d), as well as their interaction on the chitosan oligomer electromigration rate. Analysis of the data revealed that the interaction E*d was not significant (P>0.381). At the same time, data analysis showed that the effect of the solution flow velocity was not significant (P>0.785). The effect of the applied average electric filed (E) was significant (P<0.001). Because of the non significant effect of the solution flow velocity, the presented data on the electromigration rate of each chitosan oligomer as function of the processing time and applied voltage will be the averaged values obtained at the studied three flow velocities. The electromigration rate of each oligomer was expressed as a percentage of the total amount contained in the feed chitosan oligomer mixture solution.

Dimer electromigration rate as function of the processing time and applied voltage is shown in Figure 8.2. Independently of the applied voltage, the dimer electromigration rate showed linear kinetics. By applying 2.5 V/cm to the cell, the dimer electromigration rate behavior was correlated to a linear equation by a regression coefficient of R²=0.9668 and a slope of 2.36 representing the speed of the dimer electromigration through the 10000 Da cellulose ester ultrafiltration membrane used. At the end of the electroseparation process using 2.5 V/cm, the dimer electromigration rate reached 9.92 ± 2.45%. The dimer electromigration rate increased by increasing the applied electric field strength (voltage). At 5 and 10 V/cm, the dimer electromigration behavior was correlated to linear equations by regression coefficients of R²=0.9842 and 0.9974, respectively. The slopes of the corresponding equations were 6.54 at 10 V and 11.80 at 20 V. Electromigration rates at the end of the process were 26.79 ± 0.51 and 47.86 ± 3.12% when 5 and 10 V/cm, respectively.

Figure 8.2: Dimer electromigration rate as a function of time and the applied electric field strength at the averaged solution flow velocity.

Trimer electromigration rate evolution as a function of the processing time and applied voltage to the electrodialysis with ultrafiltration membrane (EDUF) cell is shown on Figure 8.3. At 2.5 V/cm, the trimer electromigration rate kinetic was correlated to a curvilinear function with a regression coefficient R²=0.9909. The average electromigration rate at the end of the electroseparation in this electric field condition reached a mean value of 8.58 ± 1.03%. When a voltage of 5 or 10 V/cm was used, the electromigration rate kinetic of the trimer followed a linear behavior. At 5 V/cm, the electromigration kinetic of the trimer was correlated to a linear function with a coefficient R²=0.9919 and a slope of 5.13. Electromigration rate was 20.28 ± 1.56% at τ = 4h. In the case when a voltage of 10 V/cm was used, the trimer electromigration rate was correlated to a linear function with R²=0.9953. The slope of the corresponded equation was 9.14. Under this applied voltage, trimer electromigration rate was 36.19 ± 2.24% at τ = 4h of the electroseparation procedure.

Figure 8.3: Trimer electromigration rate as function of time and the applied electric field strength at the averaged solution flow velocity.

The electromigration behavior of the tetramer is shown on Figure 8.4. At an applied voltage of 2.5 V/cm, the tetramer was found in the adjacent KCl compartment only after 3h of treatment with an average electromigration rate of 5.06 ± 0.61%, which increased up to 6.46 ± 0.67% at the end of the electroseparation process. By increasing the applied average electric field up to 5 and 10 V/cm, respectively, the tetramer was found in the adjacent KCl compartment since the first hour of treatment. In both cases, it migrated with a linear kinetic. At 5 and 10 V/cm, the kinetics were fitted to linear curve regression coefficients of R²=0.9669 and 0.9950, respectively. The equations slopes indicating the electromigration speed of the tetramer through the 10000 Da MWCO ultrafiltration membrane were 3.75 and 6.86, respectively. At the end of the process, electromigration rates of the tetramer under these conditions were 14.46 ± 1.52 and 26.39 ± 2.33%, respectively.

Figure 8.4: Tetramer electromigration rate as function of time and the applied electric field strength at the averaged solution flow velocity.

8.3.1.1. Effect of the flow velocity

Effect of solution flow velocity (flow rate) is directly related to the residence time of a given molecule in the side of the separation chamber. The higher the solution flow velocity the lower the residence time. Solution flow velocity (flow rate) must ensure a minimum time to the molecule to reach the membrane surface to be able to penetrate it. Such explanation was previousely reported in the literature (Galier & Roux-de Balmann, 2004; Roux-de Balmann et al., 1998).

Also, taking into account the effectiveness of the promoter of turbulence to offer a particularly effective agitation of the solutions at the both sides of the ultrafiltration membrane used, it would be expected that the contribution of the solution flow velocity; which was the same in both chitosan oligomer mixture solution and KCl compartments to the hydrodynamic conditions at the interfaces membrane/solution was not important. The promoter of turbulence is mainly used to diminish the concentration polarisation phenomenon. The range of the used solution flow velocities was sufficient to offer this condition. Consequently, the effect of solution flow velocity on chitosan oligomer electromigration rate at a fixed value of the applied voltage was not significant. Nevertheless, it was reported that it is necessary that the solution flow velocity must be sufficient to make the role of the promoter of turbulence effective in order to minimise the concentration polarisation phenomenon. In our study, it seems that the used solution flow velocities offer this condition. This explanation is in good agreement with the information reported in the literature (Lteif et al., 1999). They showed that the solution flow velocity did not have any significant effect on the electrolyte transport through a membrane if the solution was well mixed. In their study (Lteif et al., 1999), magnetic bars were used to insure an adequate agitation of the solutions in the cell compartments. Varying simultaneously the solutions flow velocities, a series of experiments were carried out at 16.66 and 25 mL/min. They reported that taking into account the presence of a particularly effective magnetic agitation, the contribution of the solution flow velocity to the solute transfer through the membrane was not significant. In our study, an adequate agitation of the solution was obtained by the promoter of turbulence and the flow rate. At the other hand, the effect of solution flow velocity on the volumetric permeate flux of dissolved organic molecules through 15 kDa ceramic ultrafiltration membrane was evaluated (De la Rubia et al., 2006). In this study, it was reported that the solution flow velocity had a significant effect on solute permeation through the membrane used. The effect of solution flow velocity was expressed by Reynolds number. They used Reynolds number values varied between 2500 and 20000. They showed that decreasing the Reynolds number, the volumetric permeate flux decreased. At the difference with our study where the driving force of the process is an applied external electric filed, in the study described by (De la Rubia et al., 2006), the driving force of the system was a pressure gradient. The impact of modifying the flow velocity of the feed solution on the yield and selectivity of an electrodialysis with ultrafiltration (EDUF) process for the fractionation of peptides was studied (Poulin, 2007; Poulin et al., 2006a). Solution flow rates of 100, 150, 200 and 250 mL/min were used. In this study, it was shown that the solution flow velocity had no affect on the total peptide migration but on the migration of particular peptides when a solution flow rate of 250 mL/min and up was used. These authors explained that this behavior could be attributed to the limited electrophoretic mobility of these peptides.

8.3.1.2. Effect of the electric field strength

The dimer showed a linear electromigration kinetic whatever the applied average electric field. This behavior indicates that the dimer migrated quite freely through the 10000 Da MWCO ultrafiltration membrane used, under the effect of the applied external electric field as a driving force. This is in a good agreement with the result reported in our previous studies (Aider et al., 2007a,b) in which it was shown that the dimer electromigration kinetic followed a linear behavior. In these studies, the cell configuration used was the same as in the present work with 10000 Da MWCO ultrafiltration membrane and a voltage of 5 V corresponded to average electric field strength of 2.5 V/cm. The trimer electromigration behavior followed a curvilinear type equation and showed an inflexion point after 2h of treatment, when an electric field strength of 2.5 V/cm was applied. But in the cases when electric field strengths of 5 and 10 V/cm were used, its electromigration behavior followed a linear kinetic. This behavior could be explained by an enhancement effect of the electric field on this molecule migration through the 10000 Da MWCO ultrafiltration membrane used. Indeed, by increasing the applied voltage, this molecule was subjected to a higher driving force that makes it possible to significantly minimize the friction exerted by the membrane. As result, the trimer migrated quite freely at the applied electric fields of 5 and 10 V/cm. The given explanation about the trimer is also applicable to the tetramer. Results explaining the effect of the applied voltage on solute electromigration through cellulose ester ultrafiltration membrane was previously reported (Roux de-Balmann & Sanchez, 1992). In a membrane contactor system, they studied a separation of model proteins. The studied mixture was composed of Poly-L-glutamic acid, α- lactalbumin and bovine haemoglobin. They showed that the solute migration trough the membrane increased by increasing the electric field. The electroseparation system used by Roux de-Balmann & Sanchez (1992) is similar to our system considering that both systems are based on the same principle (Figure 8.1). The difference is that in their study, the membrane active area was 32 cm² and the separation was carried out at a fixed current density varied from 3 to 22 A/m² corresponded to an electric field varying from 1 to 8 V/cm. Solution flow velocities of 1.66 and 83.33 mL/min were used. In our case, the active membrane area was 10 cm², fixed electric field ranging from 2.5 to 10 V/cm was studied. Our study was carried out at solution flow velocities of 100, 300 and 500 mL/min. On the other hand, the electrophoretic transfer of proteins across 10⁶ Da MWCO polyacrylamide membrane in an electrophoresis-based separation technique (Gradiflow technique) was studied (Rylatt et al., 1999). They reported that bovine serum albumin (BSA) transfer across a polyacrylamide membrane was directly proportional to the applied electric field strength. It was shown that as the electric field increased; there was a proportional increase of the amount of the BSA that migrated through the used 10⁶ Da MWCO polyacrylamide membrane. In the study described by Rylatt et al. (1999), the electric voltages used were 25, 50, 100 and 200 V, and the assembled cartridge was 15 cm long, 3 cm wide and 1 cm deep. These voltages corresponded to average electric field strengths of 25, 50, 100 and 200 V/cm. In this study, the heat generated at higher voltages imposed a practical upper limit to the voltage which can be applied to the Gradiflow electro-separation system. A comparison of our study to the work of Rylatt et al. (1999), revealed some particularities. Our upper electric field was 10 V/cm while their lower one was 25 V/cm. The BSA is a protein with a molecular weight of 67000 Da (Oussedik & Mameri, 2001), isoelectric point (PI) of 4.9, negatively charged at pH values above the PI, and electrophoretic mobility of -1.7x10^-6 m²/V.s (Takeda et al., 1992). The ratio of the polyacrylamide membrane MWCO used in this study to the solute (BSA) molecular weight is approximately 15, while in our study the ratio of the 10000 Da MWCO cellulose ester ultrafiltration membrane used to the dimer, trimer and tetramer molecular weight was 12.5, 17 and 25, respectively. In the contrary of the work of Rylatt et al. (1999), in our case, there was no significant heat generation because the applied voltage was lower. The temperature increased by 8-9 °C along the electroseparation procedure.

Concerning the separation possibility of the chitosan oligomers, it was not possible to obtain any separation at 5 and 10 V/cm because the applied electric field strength was sufficient for the tetramer to migrate through the membrane together with the dimer and trimer since the first hour of the treatment.

8.3.3. Mathematical validation of the flow velocity and electrif field effects

At the same time, each chitosan oligomer is submitted to the effect of the applied electric field and solution flow velocity. As consequence, the resultant molecule velocity is the sum of two vectors resulted from the electric filed strength and solution flow velocity effects. The following equation can be written (Doneddu, 1984):

(8.3.2.1)

being the molecule linear velocity caused by the applied external electric field, in a direction perpendicular to the membrane, (m/s).

being the molecule linear velocity caused by the applied solution flow velocity, in a direction parallel to the membrane, (m/s).

These above mentioned velocities can be expressed through chitosan oligomer electrophoretic mobility (μ), electric field strength (E), solution flow velocity (d) and area (S) calculated from compartment thickness and membrane length (Sun, 2004; Young et al., 2004):

(8.3.2.2)

(8.3.2.3)

By replacing Vx and Vy in eq.3.2.1, we obtained the following resultant velocity:

(8.3.2.4)

If we consider X as being the maximal available distance for each molecule to be in contact with the active membrane area, and τ the corresponding time, then, the following equation could be written:

(8.3.2.5)

By combining equations (3.2.4) and (3.2.5), the result will be given by the following equation:

(8.3.2.6)

From equation (3.2.6), the maximal time needed for a given molecule to reach the membrane surface will be as follows:

(8.3.2.7)

According to the equation (8.3.2.7), the above calculated time for a given chitosan oligomer in the feed solution compartment depends on the compartment geometry (X), solute electrophoretic mobility (μ) which is an intrinsic molecule characteristic at given concentration, pH value and ionic strength of the medium (Vishu-Kumar et al., 2005), the applied external electric field strength (E), surface (S) and solution flow velocity (d).

Considering that the electrodialysis with ultrafiltration membrane (EDUF) cell is a multi-compartment separation chamber and that the electrical conductivity of the solutions in the different compartments are not the same, the electric field strength is consequently not the same in these compartments. Using the nominal averaged electric field strength in equation 8.3.2.7 is less precise. According to (Galier & Roux-de Balmann, 2004; O’Connor et al., 1996), the local electric field (E _l ) can be calculated through the current intensity (I), active membrane surface (S) and solution electrical conductivity (κ) as follows:

(8.3.2.8)

By replacing the electric field in equation (3.2.7) by equation (3.2.8), we obtain the following equation:

(8.3.2.9)

For each chitosan oligomer, the equation (8.3.2.9) was plotted in 3D surface as function of the two studied variables (electric field strength and solution flow velocity). Only data of the dimer were shown because the tendencies were the same for all oligomers even if the calculated times were different. Figure 8.5 shows the time needed for the dimer to reach the membrane surface as a function of the electric field and solution flow velocity. We can see that the parameter influencing this time is the electric filed, and that the higher the electric field the lower the time needed independently on the solution flow velocity (rate). Using the first derivative of the equation (8.3.2.9) as a function of the solution flow velocity (d), it was possible to calculate the minimal solution flow velocity from which this last (flow velocity) is not significant. It was found that at a solution flow velocity of 0.83 cm/s corresponding to a flow rate of 30 mL/min and above, this parameter does not exert any effect on the time needed for the solute to reach the membrane surface. The same values were calculated for the trimer and tetramer and were found to be equal to 1.16 and 1.49 cm/s (42 and 54 mL/min), respectively.

8.3.4. Chitosan oligomer transport numbers

Ions or charged organic molecules may carry drastically different portions of the total current. This depends on their respective electrophoretic mobilities, valences and concentrations (Usobiaga et al., 2000). Transport numbers of the studied chitosan oligomer mixture are shown in Table 8.2. The solution flow velocity and electric field strength did not affect the transport numbers of each given chitosan oligomer (P>0.171). At the same time, the chitosan degree of polymerisation had a significant effect on the transport number. Multiple comparisons of the chitosan oligomers transport numbers did not reveal any difference between the dimer and trimer transport numbers (P>0.299), while the tetramer transport number was significantly different from those of the dimer and trimer (P<0.006).

Results presented in Table 8.2 showed that approximately 7-8% of the total electric current passing through the electrodialysis with ultrafiltration membrane (EDUF) cell was transported by the chitosan oligomers. Considering that the total transport number can not exceed 1, we can deduce that approximately 92-93% of the total electric current was carried by the electrolytes present in the solutions (Na⁺, K⁺, Cl^-, H⁺, OH^-). The tetramer showed the lowest transport number because of its lower concentration in the feed solution compared to that of the dimer and trimer, its lower electrophoretic mobility as shown by (Aider et al., 2006a) and membrane restriction to its migration as shown in our previous study on the effect of membrane molecular weight cut-off on the chitosan oligomer electromigration kinetics (Aider et al., 2007a).

Table 8.2: Average transport number of chitosan oligomers as a function of the electric field strength applied to the electrodialysis with ultrafiltration membrane cell.

8.3.5. System performance

The EDUF system performance was evaluated according to the measurement of the limiting current density, global electrical resistance of the EDUF cell, ultrafiltration and anion-exchange membranes integrity by measuring their main characteristics (electrical conductivity and thickness) before and after each electroseparation process.

8.3.5.1. Limiting current density

In the present study, the procedure described by Cowan & Brown (1959) was used to determine the limiting current density. According to the electrodialysis with ultrafiltration membrane cell configuration (Figure 8.1), data obtained by plotting the global system resistance versus 1/I (Cowan & Brown, 1959)did not reveal any inflexion point indicating that the limiting current density was reached (data not shown). This result is due to the fact that the chosen cell configuration permits to keep a good electrical conductivity in all compartments, and thus water splitting did not occur to compensate electrolyte deficiency for electric current carrying.

Solution flow velocity at the both sides of the membranes, presence of a promoter of turbulence, cell configuration type, and concentration of the feed solution have a direct impact on the electrical current utilization as well as on the general cost of the process (Tanaka, 2006). It was demonstrated that an electroseparation process shows higher electrical resistance or lower current efficiency when it is operated at current density above its limiting value (Vera et al., 2007). Therefore, the limiting current density should be considered as one of the critical system parameters since it determines the efficiency of the electroseparation process unit (Tanaka, 2003). In our study, the, the electric current was efficiently used since the limiting current density was not reached.

8.3.5.2. Global system electrical resistance

One of the parameters giving an indication of the system performance is the global system electrical resistance evolution (Tanaka, 2003). Figure 8.6 shows the kinetic of this parameter as function of the processing time according to the electrodialysis with ultrafiltration membrane cell configuration used. Data analysis showed that the processing time had a significant effect on the global system electrical resistance evolution (P<0.007), the electric field strength had also a significant effect (P<0.001) on this parameter, whereas the solution flow velocity had no significant effect on the measured global system electrical resistance (P>0.902). The higher global system electrical resistance was recorded when electric field strength of 2.5 V/cm was used. It decreased by increasing the electric field strength. The lowest global electrical resistance values were recorded with 10 V/cm.

Figure 8.6: Global EDUF system electrical resistance evolution as function of time, electric field strength and solution flow velocity.

8.3.5.3. Membrane integrity

Tables 8.1 shows the main characteristics of the 10000 Da MWCO UF-membrane used as well as those of the anion exchange membranes at the anodic and cathodic compartments, before and after electroseparation process. Data analysis showed that the applied voltage (P>0.251) and solution flow velocity (P>0.761) did show any significant effect on the characteristics of these membranes. The comparisons of the electrical conductivity and thickness of each membrane before and after each electroseparation treatment did not show any significant difference. This means that all the membranes kept their integrity and that no significant fouling was occurred. This is in good agreement with the previous results obtained in the same conditions (Aider et al., 2007a,b).

According to the results of the present study, we can extrapolate that it will be possible to work continuously with a voltage of 5 V that corresponds to an average electric field strength of 2.5 V/cm during 16h without any significant loss of the membranes integrity, since the effect of a voltage of 20 V (E = 10 V/cm) during 4h is similar to that of a voltage of 5 V (E = 2.5 V/cm) during 16h.

Conclusion

In the present study, it was found that the solution flow velocity did not exert any significant effect on the electromigration rate of chitosan oligomer, as well as on system performances. At the same time, it was found that the applied electric field strength had a significant effect on all chitosan oligomer electromigration rates. The dimer average electromigration rate after 4h of treatment varied from 9.92 ± 2.45 to 47.86 ± 3.12% for an electric field strength range of 2.5 to 10 V/cm (voltage from 5 to 20 V). The trimer migrated together with the dimer with electromigration rate of 8.58 ± 1.03 up to 36.19 ± 2.24% for the above mentioned values of the electric field. The tetramer showed the lowest electromigration rates. When electric field strength of 2.5 V/cm was applied, a solution composed only by the dimer and trimer was obtained until 2h of electroseparation treatment. At the pH used in the present study, no separation was possible at any time when electric field strengths of 5 and 10 V/cm were applied to the electroseparation system with the 10000 Da MWCO cellulose ester ultrafiltration membrane stacked in the EDUF cell.

The results obtained can be used to optimise the performance of the electroseparation process of the studied chitosan oligomers by increasing the effective ultrafiltration membrane area under applied electric field strength of 2.5 V/cm. This technique can be successfully used in the basic research laboratory as well as for commercially important scales for chitosan oligomers separation.

Other studies are needed to verify if the separated chitosan oligomers by EDUF technology keep their biological functions considering that soft conditions were used for their separation.

Acknowledgments

The authors would like to thank Ms. Monica Araya-Farias for her technical assistance, Mr. Alain Gaudreau for his technical support for chitosan oligomer HPLC analysis. The financial support of the Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT) is also acknowledged.