Influence of experimental parameters on the formation of chitosan microspheres by the emulsion cross-linking method

In the classical cross-linking method, chitosan solution is subjected to straightforward cross-linking by mixing with a cross-linking agent, which leads to gelling, followed by crushing the resulted gel into particles. Glutaraldehyde is commonly used as cross-linking agent, accomplishing in the same activation of the carrier. In this approach, enzymes are entrapped in the formed gel if they have been mixed with chitosan before performing the cross-linking. (Krajewska, 2004). In our work, the emulsion cross-linking approach was used (Denkbas and Odabasi, 2000). Chitosan microemulsion was obtained in an oil phase and the resulted microspheres were hardened by addition of glutaraldehyde cross-linker. In the present study, 3% glutaraldehyde (v/v, related to the chitosan solution volume) has been used, while influence of different glutaraldehyde concentrations was studied in the enzyme immobilization experiments (section 3.3). The effect of particle preparation parameters (oil phase composition, emulsifier nature and emulsifier concentration) on the mean size of particles was investigated.

Effect of oil phase composition

The principal characteristic required in the preparation of microparticles was the uniformity of their size expressed by the span value (span < 1 means narrow particle size distribution). Although particles prepared with ionotropic gelation and precipitation with spraying were smaller, they exhibited a wide size distribution. The narrowest size distribution was presented by microparticles prepared with the emulsion cross-linking method. That is why this method was chosen for further investigation. The main goal of this study was the reduction of size of particles, but maintaining the narrow size distributions.

To achieve this aim, single component organic phases have been tested (paraffin oil, sunflower oil, n-octane and n-hexadecane), as well as two concentrations, 1 and 2% (v/v)

of surfactant Tween 80. The concentration of chitosan solution in all these experiments was 2% (w/v), in aqueous acetic acid (2%, v/v). According to our observations, from these organic compounds only n-hexadecane was suitable for microparticle preparation, as the others caused strong aggregation of the chitosan droplets and the mean size of particles could not be measured (Table 3.1).

Table 3.1. Mean sizes of the chitosan microspheres obtained by various compositions of oil phases with 1% Tween 80.

The explanation resides in that a too high (paraffin oil, sunflower oil) or too low (n-octane) viscosity of the dispersion phase is non-adequate for particle formation. Therefore, mixtures of oil and an organic solvent were subsequently tested. As organic solvent, petroleum ether, used by other authors (Denkbas and Odabasi, 2000), has been found unsuitable for our experiments due to its high volatility, and it was replaced with n-octane.

The effect of the dispersion medium composition on the particle size was investigated using 1% Tween 80 as emulsifier. When the oil phase consisted from different ratios of paraffin oil and n-octane, both the viscosity and density of the mixtures increased with increasing paraffin oil amount (Table 3.1). The physical characteristics of the oil phase, expressed by density and viscosity, had a slight influence on the size of formed chitosan microparticles. While at lower paraffin oil/n-octane ratios, up to 40% relative paraffin oil content, the size of particles was practically not modified (taking into account the standard deviation), at increasing paraffin oil content this size slightly decreased. At the highest paraffin oil/n-octane ratio (75:25%, v/v), the size of particles could not be measured, as formation of irregularly shaped chitosan particles and strong aggregation were observed.

Using two-component oil phases with paraffin oil, but replacing octane with n-hexadecane, a less volatile hydrocarbon with higher viscosity, the same decreasing effect on the size of particles was noticed. Ultimately, the combination of paraffin oil with n-hexadecane resulted in the lowest particle size, as shown in Table 3.1. Since paraffin oil is more viscous than sunflower oil, it results that increase of viscosity of the dispersion medium leads to decrease of the mean size of particles.

Effect of various surfactants on particle size

The surfactant has emulsifying role in the process, therefore, it was necessary to investigate the effect of surfactant nature on chitosan microparticle formation, as well. In the experiments performed with different surfactants, a mixture of 40% sunflower oil and 60% n-hexadecane was used as dispersion medium at the same surfactant concentration (1%). As results from Table 3.2, the sizes of obtained particles were strongly influenced by the chemical structure of surfactant.

Table 3.2. Effect of various surfactants on mean particle size of chitosan microspheres obtained in 40% sunflower oil and 60% n-hexadecane oil phase, containing 1% surfactant Surfactant name Chemical structure HLB value* Particle mean size

[µm]

Lutensol XL80 C10-guerbet alcohol ethoxylate 14.0 535.3

Triton X 100 octyl phenol ethoxylate 13.4 499.4

Tween 80 polyoxyethylene-sorbitan

monooleate 15.0 453.3

Emulsogen M fatty alcohol polyethyleneglycol

ether 9.0 400.8

VEM-75/03 polyethylene glycol oleate 7.9 281.3

Span 80 sorbitan monooleate 4.3 21.3

*from Flick, 1991.

Chitosan has a polycationic character, containing cationic NH₃⁺ and polar –OH groups distributed along the polysaccharide backbone, without large hydrophobic groups. This implies that chitosan will strongly interact with cationic surfactants, whereas the mechanism of interaction with nonionic surfactants is much more complicated, especially because chitosan itself exhibits emulsifier properties (Schulz et al., 1998). We tested several nonionic emulsifiers with different hydrophilic-lipophilic balance (HLB) values in order to control the particle size of microspheres by the nature of the added surfactant. As

the data show, there is a strong relationship between the HLB value of the applied surfactant and the obtained particle size when the HLB value is below 10 (Table 3.2). In the case of oil phase soluble surfactants, decreasing HLB values induced a decrease of the particle size. Very fine particles were formed when Span 80 surfactant was used, while any other investigated surfactants yielded particles with one order of magnitude larger size.

Unfortunately, particles formed with Span 80 and VEM-75/03, even having smaller sizes, were not appropriate for our immobilization purposes, because the elimination of surfactant in the washing stage of particles was very difficult and the particles showed strong tendency of aggregation. The presence of surfactant, even in very low amounts, had negative effect on binding of enzyme to the particle, while aggregated particles were not suitable for biocatalytic function. The polyethylene glycol oleate VEM-75/03, manufactured at the University of Pannonia, is not yet a commercial product, being currently under patenting procedure.

Based on these results Tween 80, a polyoxyethylene (20) sorbitan monooleate, was selected as surfactant. It is a representative, commercially available and low priced surfactant, accepted as emulsifier in the food industry and able to stabilize colloidal systems that contain chitosan (Klinkesorn and Namatsila, 2009).

Effect of Tween 80 concentration

Optimization of Tween 80 concentration has been carried out in dispersion media which had the oil phase composed from 40% oil and 60% complementary organic solvent. Two oil phase compositions, paraffin oil/n-octane and sunflower oil/n-hexadecane, both at 40/60 (v/v) ratio, were selected to study the influence of surfactant concentration. The Tween 80 concentration was altered in the range of 1-5%. (Fig. 3.6 and 3.7).

Generally, the sunflower oil/n-hexadecane oil phase resulted in smaller particles than paraffin oil/n-octane. This effect can be related to the increased oil phase viscosity, as it was discussed earlier. Comparison of experimental data obtained by using 40% oil and 60% organic solvent demonstrated that the increase of the Tween 80 concentration gave different results for the two oil phases. When the oil phase consisted of paraffin oil and n-octane, the smallest particles were obtained at 3% Tween 80 concentration (Fig. 3.6), but the character of the relationship between particle mean size and emulsifier concentration was not unambiguous because of the extent of standard deviation of the particle size data.

Tween 80 concentration [%]

1 2 3 4 5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 200.0 400.0

600.0 Span

Mean size [µµµµm]

Fig. 3.6. Effect of Tween 80 concentration on the particle size in 40% paraffin oil and 60% n-octane

Tween 80 concentration [%]

1 2 3 4 5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 200.0 400.0

600.0 Span

Mean size [µµµµm]

Fig. 3.7. Effect of Tween 80 concentration on the particle size in 40% sunflower oil and 60% n-hexadecane

When sunflower oil/n-hexadecane oil phase has been used, the mean particle size decreased with increasing Tween 80 concentration in the 1-4% range, but no more decrease occurs by applying 5% emulsifier (Fig. 3.7). Denkbas and Odabasi (2000) obtained similar results with mineral oil + petroleum ether dispersion medium and Span-85 emulsifier concentration in the range of 0.5-2%. In the mixture of 40% paraffin oil and

60% n-octane the effect of Tween 80 on the particle size is not unambiguous. In this case, the influence of emulsifier concentration was similar to the other organic medium in the 1-3% range, but the mean particle size increased at higher Tween 80 concentrations. It seems that an optimum emulsifier concentration has been achieved, and further increase of this concentration was not useful to reduce the size of the obtained particles. However, it must be pointed out that differences between the mean particle sizes at 2-5% emulsifier content were not significant (the difference between the highest and lowest value was less than 10%). Moreover, if the relative standard deviation values (having values between ±2.4%

and ±6.3%) are taken in consideration, the most plausible conclusion is that in the emulsifier range mentioned above, the Tween 80 concentration has no significant influence on the particle size in paraffin oil/n-octane medium.

As concerns the sunflower oil/n-hexadecane medium the mean particle size decrease was more than 30% at increasing Tween 80 concentration from 1% to 4%, therefore the tendency in this case is much more evident, even if the influence of relative standard deviations is taken into account. The explanation of this phenomenon needs further investigations considering the complicated interactions which exist between chitosan, oil, solvent and Tween 80. Although paraffin oil was suitable for preparation of chitosan particles, it was difficult to remove it from the surface of particles. This fact could be disadvantageous in view of enzyme immobilization. For this reason, further investigations were carried out with sunflower oil as component of the dispersion medium.

Microspheres prepared in sunflower oil /n-hexadecane oil phase exhibited good sphericity and rather smooth surface showing small peaks, as well as some pores (Fig. 3.8).

Fig. 3.8. SEM micrograph of a microsphere obtained by emulsion cross-linking method in 40% sunflower oil and 60% n-hexadecane with 1% Tween 80 (scale bar: 50 µm)