3.1. Two-Layer Structure of Film
According to SEM pictures of films EG1, EGA1, and EGC1 shown in
Figure 1a,e,i, with the glycol content as 0.3 g, the crosslinked films maintained the same two-phase asymmetric microstructure morphology as compared to films without crosslinking in previous studies [
13,
14]. Therefore, all the complex films can be simplified to a two-layer structured film as shown in
Figure 2, with a porous CEO-rich layer and a compact polymer-rich layer. In
Figure 2,
d is the total thickness of film;
D is the equivalent diffusion coefficient of CEOs diffused through both the two layers; the thickness of the CEO-rich layer is
d1; and the equivalent diffusion coefficient of CEOs diffused through this layer is
D1. Similarly, the thickness of the polymer-rich layer is
d2, and the equivalent diffusion coefficient of CEOs diffused through this layer is
D2. It is clear that
d =
d1 +
d2,
D =
D−T,
D1 =
D−B [
13].
Thus, the relationship between
D and both
D1 and
D2 can be simplified from Equation (5) to Equation (6):
Then, the relationship between
D1 and
D2 can be transformed from Equation (6) to Equation (7):
Table 2 shows little difference between
D1/
D2 of films G, GA, and GC, which means to catch up to the significant variation of
D1,
D2 should alter at the same rate with
D1 of each film, respectively. The reason for causing a significant difference between
D1 of films G, GA, and GC was assumed to be the different electronical interaction of the emulsifier composition in previous studies. Therefore,
D2 should also be influenced by the electronical interaction of the emulsifier composition. Hence, the polymer-rich layer must be composed of not only SA but also emulsifier chains. Likewise, the CEO-rich layer includes not only the CEO and emulsifier composition but also SA chains. Similar molecular interactions bring out a similar environment for molecular thermal motion, thus allowing for a correlation between
D1 and
D2.
3.3. Influence of Glycol and Ethanol on D
The experimental and theoretical values of the CEO-release proportion out of each film are depicted in
Figure 3,
Figure 4,
Figure 5 and
Figure 6 except
Figure 4. The goodness of fits is evaluated by means of the root mean square error (RMSE) that are listed in
Table 4,
Table 5 and
Table 6. As can be inferred from the data presented in all the above figures and tables, the model satisfactorily fits the experimental data, suggesting that the adopted model can be used to obtain useful information on the mechanism of CEO release from the crosslinked SA matrices.
Since ethanol is not soluble with the polymer chains, ethanol has no significant effect on the force field formed by the atoms of the polymer chains, and the film does not swell. With solvent ethanol contacting with film, the ethanol molecule enters the film surface through the pores among polymer chains by osmosis. Moreover, because glycerol and water molecules are both soluble in ethanol, the room that is occupied by the glycerol and water well dispersed in the film works as pores or free volume for ethanol molecule. More free volume brings out easier diffusion. Likewise, because CEO is soluble in ethanol, CEOs diffused more easily in the film where ethanol is distributed. Therefore, ethanol acts as an extraction promoting the CEO diffusion.
Specifically, ethanol molecules diffuse firstly from one boundary of the film contacting the solvent through the polymer-rich layer by (1) directly entering pores, (2) dissolving in free water, and (3) dissolving in glycerol among polymer chains. Secondly, reaching the CEO-rich layer, ethanol molecules diffuse in the same way to the interior of the microspheres and dissolve CEOs. Thirdly, ethanol molecules diffuse to the other boundary of film contacting with barrier layers. Unable to penetrate the membrane, ethanol molecules start to move backward. Simultaneously, dissolved CEOs will move out of the film with ethanol, and so do the water and glycerol molecules that are dissolved in ethanol.
The calculated
D values listed in
Table 4 show that for CEOs diffused from films to absolute ethanol, the
D values increased with a glycerin content increase in all the films. This may be due to two reasons. (1) With solubility among CEOs, ethanol, and glycerin, glycerin itself acts as free volume. More glycerin means more free volume, which increases the space between chains and weakens the interactions among atoms of chains, thus bringing out larger
D values. (2) Addition of glycerol content as plasticizer increases the space between molecular chains. Meanwhile, with strong hygroscopicity, more glycerol means higher moisture content, and the water molecules also play a similar plasticizing effect as glycerol [
21]. Thus, this improves the mobility of the molecular chain segments, reduces the energy required for free volume redistribution, increases the generation frequency of CEOs transition channels, and promotes the diffusion of CEO molecules.
Figure 4 shows that with the increase of glycerol content (0.3, 0.6, 0.9, 1.2 mL), the
D values of CEOs diffused in films EG, EGA, and EGC showed a similar growth pattern. All of the
D values increased by small degrees, by a large margin, and in small amounts in the range of 0.3–0.6 mL, 0.6–0.9 mL, and 0.9–1.2 mL, respectively. This may be caused by different distributions of glycerol and the different interactions between chains.
During the increase of glycerol content from 0.3 to 0.6 mL, glycerol might distribute mostly in the original interchain pores, that is, most of the glycerol volume replaces the pore volume without increasing extra spacing between chains. Therefore, the total equivalent free volume in the film does not significantly increase. Glycerol only slightly improves the mobility of molecular segments. Thus, the decrease of free volume redistribution energy is small, and the increase of D is small.
In the range of 0.6–0.9 mL, glycerol might distribute more in the extra space produced by increasing spacing between chains. Thus, the total equivalent free volume in the film increases. Moreover, in this range of chain spacing, the intermolecular force decreases rapidly with the increase of spacing. Therefore, the energy required for free volume redistribution reduces greatly, that is, the formation frequency of diffusion channel is greatly increased. Thus, the increase of D is large.
In the range of 0.9–1.2 mL, glycerol might as well distribute in the extra space produced by increasing more spacing between chains, which means the total equivalent free volume in the film is further increased and the wall diameter of the diffusion channel is larger. However, in this range of spacing, the intermolecular force decreases slowly with the increase of spacing. Therefore, the energy required for free volume redistribution decreases slightly. Thus, the increase of D is small.
In conclusion, the size and distribution of diffusion channels and the intermolecular forces jointly determine the diffusion ability of molecules.
With the increase of glycerol content, the
D values of CEOs diffused in the film always follow the relationship of
DEG >
DEGA >
DEGC. This may be due to the fact that the electrostatic interaction between chains conform to the relationship of GC > GA > G, as discussed in previous work [
13]. Furthermore, the
D values of external crosslinked films are significantly smaller than the
D values of films without the crosslinking process [
13]. It might be the bridging effect that emerged by electrostatic interaction between the –COO
− of sodium alginate chains and the Ca
2+ that increases the interaction between molecular chains [
22]. In addition, since the electrostatic interactions enhanced by calcium ion crosslinking are between sodium alginate molecular chains, the ordering of electrostatic interactions between the emulsifier combinations are not affected.
With either the emulsifier combination or the crosslinking process by calcium ions, the mechanism for them to affect CEO diffusion is both through altering the electrostatic interaction between the molecular chains. Higher electrostatic interaction increases the binding effect of chains on CEO molecules and reduces free volume. Therefore, chain segments move more difficultly, and the energy needed for redistribution of free volume increases. Thus, it is more difficult to form a diffusion channel which is not conducive to the diffusion of CEO molecules, and which brings out smaller D.
Moreover, it was found that the D values of CEOs diffused to absolute ethanol through film EG1 and GA were very close. This might be due to the fact that the electrostatic interaction between –COO− of sodium alginate chains and Ca2+ is likely to be the same as that between –COO− of acacia gum and –NH3+ of the gelatin chain. The D values of CEOs diffused to absolute ethanol through films EGA1 and GC were very close. This might be due to the fact that the electrostatic interaction between –COO− of sodium alginate chains and Ca2+ and that between –COO− of acacia gum and –NH3+ of the gelatin chain is likely to be the same as that between –COO− of sodium carboxymethyl cellulose and –NH3+ of the gelatin chain.
The results also show that the D values of CEOs diffused from films EG1 and EGA2 to absolute ethanol are very close. This might be because the electrostatic interaction between –COO− of acacia gum and –NH3+ of gelatin chain decreased D by the same amount with the increment of D promoted by 0.3 mL of glycerol in the range of 0.3–0.6 mL. Similarly, the D values of CEOs diffused from films EG2 and EGC3 to absolute ethanol are very close. This might be because the electrostatic interaction between –COO− of sodium carboxymethyl cellulose and –NH3+ of the gelatin chain decreased D by the same amount with the increment of D promoted by 0.3 mL of glycerol in the range of 0.6–0.9 mL. Moreover, D values of CEOs diffused from films EGA1 and EGC2 to absolute ethanol are very close. This might be because the electrostatic interaction between –COO− of sodium carboxymethyl cellulose and –NH3+ of the gelatin chain minus that between –COO− of acacia gum and –NH3+ of the gelatin chain decreased D by the same amount with the increment of D promoted by 0.3 mL of glycerol in the range of 0.3–0.6 mL.
In addition, in the range of 0.3–0.6 mL, the increment of DEGA and DEGC per unit glycerol is significantly smaller than that of DEG. This may be due to the stronger electrostatic interaction between the chains in films EGA and EGC, which squeezes more glycerol into the holes originally existing between chains. Moreover, there are more hydroxyl groups in acacia gum and sodium carboxymethyl cellulose molecular chains than in gelatin chains. Since the hydroxyl groups have stronger adsorption on glycerol, glycerol might distribute more finely and stably in the holes between chains, thus reducing the influence on diffusion.
3.5. Influence of Aqueous Ethanol on D
For the solvent containing both ethanol and water, i.e., aqueous ethanol, the influence on D consists of the blocking effect of polymer chains (reduced by water), the blocking effect of water, and the extraction-promoting effect by ethanol. Here, the one that plays a leading role determines the diffusion ability of CEO.
Table 6 shows that for those
D values of CEOs diffused to ethanol that are much greater or less (two orders of magnitude, for instance) than those which diffuse to water, the diffusion of CEOs in the film is very sensitive to water. As shown in
Figure 7a,
D can be greatly changed by a low moisture content and small swelling rate. For instance, for film EG3,
D decreases dramatically in the range of a low percentage of moisture in the solvent (
w = 0 to 0.3) and film (
s = −0.06 to −0.03). This might be due to the long spacing between polymer chains, which has already made the film reach the range of weak and slow change of a molecular force field (as discussed in
Section 3.3). Thus, the promoting effect by swelling and ethanol extraction is not significant. On the contrary, with the increase of percentage of moisture in the solvent, the partition coefficient decreased greatly, and the blocking effect of water plays a leading role, hence greatly reducing the
D of CEOs.
In another case, for film EGC1, as shown in
Figure 7b,
D increased dramatically in the range of low percentage of moisture in the solvent (
w = 0–0.3) and in film (
s = 0.07–0.34). This might be because the short spacing between polymer chains (which appeared by strong molecular force field as discussed in
Section 3.3) is significantly enlarged by the water swelling effect. Thus, the reduced blocking effect of polymer chains by water plays a leading role, hence greatly increasing the
D of CEO.
Figure 7 also shows that in the range of
w = 0–0.6, the swelling ratio of two films increased to 0.47 (EG3) and 0.77 (EGC1), respectively. Both the
D values varied considerably first and then slightly. In the range of
w = 0.6–1, the swelling ratio of two films increased to 2.6 (EG3) and 2.9 (EGC1), respectively, while both the
D values changed only slightly. The results implicate that (1) for film EG3, the partition coefficient might decrease slightly when
w > 0.6; and (2) for film EGC1, when
s > 0.77, the spacing between chains might have already made the film reach the range of weak and slow change of the molecular force field as described before. Therefore, the
D values of CEOs varied slightly with vigorously increased swelling rates. In addition, the swelling rate of the film EGC1 is higher than that of film EG3, which may be due to the retention of more water molecules by more hydroxyl groups on sodium carboxymethyl cellulose in film EGC1.