Oil Core−Polymer Shell Microcapsules Prepared by Internal Phase

Nov 19, 2004 - Cross-linking the polymer shell after capsule formation was also found to decrease both the release rate and the yield of the active in...
2 downloads 15 Views 253KB Size
11374

Langmuir 2004, 20, 11374-11379

Oil Core-Polymer Shell Microcapsules Prepared by Internal Phase Separation from Emulsion Droplets. I. Characterization and Release Rates for Microcapsules with Polystyrene Shells Peter J. Dowding, Rob Atkin, Brian Vincent,* and Philippe Bouillot School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K. Received June 10, 2004. In Final Form: September 27, 2004 Microcapsules with an oil core surrounded by a polymeric shell have been prepared by the controlled phase separation of polymer dissolved within the oil droplets of an oil-in-water emulsion. The dispersed oil phase consists of the shell polymer (polystyrene), a good solvent for the polymer (dichloromethane), and a poor solvent for the polymer (typically hexadecane). Removal of the good solvent results in phase separation of the polymer within the oil droplets. If the three interfacial tensions between the core oil, the shellforming polymer, and the continuous phase are of the required relative magnitudes, a polymer shell forms surrounding the poor solvent. A UV-responsive organic molecule was added to the oil phase, prior to emulsification, to investigate the release of a model active ingredient from the microcapsules. This molecule should be soluble in the organic core but also have some water solubility to provide a driving force for release into the continuous aqueous phase. As the release rate of the active ingredient is a function of the thickness of the polymeric shell, for controlled release applications, it is necessary to control this parameter. For the preparative method described here, the thickness of the shell formed is directly related to the mass of polymer dissolved in the oil phase. The rate of volatile solvent removal influences the porosity of the polymer shell. Rapid evaporation leads to cracks in the shell and a relatively fast release rate of the active ingredient. If a more gentle evaporation method is employed, the porosity of the polymer shell is decreased, resulting in a reduction in release rate. Cross-linking the polymer shell after capsule formation was also found to decrease both the release rate and the yield of the active ingredient. The nature of the oil core also affected the release yield.

Introduction Encapsulation involves the surrounding of a core material (solid or liquid) with a polymeric or some other solid shell.1 Often such core-shell microcapsules are used to effect protection, or the controlled time-release profile, of some active material contained in the core. Many methods for preparing microcapsules have been described, such as interfacial polymerization reactions utilizing emulsions2-5 and microemulsions,6-8 spray drying,9,10 phase separation from the continuous phase,11,12 solvent extraction,13,14 layer-by-layer addition,15,16 and sol* To whom correspondence should be addressed. Phone: +44 (0)117 9288160. Fax: +44 (0)117 9250612. E-mail: brian.vincent@ bris.ac.uk. (1) Microspheres, Microcapsules & Liposomes; Arshady, R., Ed.; Plenum: New York, 1998. (2) Janssen, L.; Tenijenhuis. K. J. Membr. Sci. 1992, 65, 59. (3) Lambert, G.; Fattal, E.; Pinto-Alphandary, H.; Gulik, A.; Couvreur, P. Pharm. Res. 2000, 17, 707. (4) Fallouh, N. A.; Roblot-Treupel, L.; Fessi, H.; Devissaguet, J. Ph.; Puisieux, F. Int. J. Pharm. 1986, 28, 125. (5) Fresta, M.; Cavallaro, G.; Giammona, G.; Wehrli, E.; Puglisi, G. Biomaterials 1996, 17, 751. (6) Watnasirichaikul, S.; Davies, N. M.; Rades, T.; Tucker, I. G. Pharm. Res. 2000, 17, 684. (7) Munshi, N.; Chakarvorty, K.; De, T. K.; Maitra, A. N. Colloid Polym. Sci. 1995, 273, 464. (8) Daubresse, C.; Grandfils, C.; Jerome, R.; Teyssie, P. J. Colloid Interface Sci. 1994, 168, 222. (9) Mathiowitz, E.; Bernstein, H.; Giannos, S.; Dor, P.; Turek, T.; Langer, R. J. Appl. Polym. Sci. 1992, 45, 125. (10) Bodmeier, R.; Chen. H. G. J. Pharm. Pharmacol. 1988, 40, 754. (11) Ruiz, J. M.; Busnel, J. P.; Benoit. J. P. Pharm. Res. 1990, 7, 928. (12) Vachon, M. G.; Nairn, J. G. J. Microencapsulation 1995, 12, 287. (13) Leelarasamee, N.; Howard, S. A.; Malanga, C. J.; Ma, J. K. H. J. Microencapsulation 1988, 5, 147.

vent evaporation,17-19 including the use of multiple emulsions.17,20-22 Microcapsules are of great interest to a wide range of industries including pharmaceuticals, printing, perfumery, cosmetics, and agrochemicals.23 In previous work by this group, Loxley and Vincent24 prepared microcapsules by the controlled phase separation of poly(methyl methacrylate) (PMMA) within the droplets of an oil-in-water emulsion. The shell-forming polymer was dissolved in a solvent mixture comprising a volatile solvent (dichloromethane) and an involatile nonsolvent (hexadecane). A sufficient volume of good solvent is present to ensure that the polymer is (just) dissolved. This solution is then dispersed into a water/stabilizer mixture to produce an oil-in-water emulsion. The volatile solvent is then removed. This leads to a change in the droplet composition, and the polymer phase separates as small droplets within the emulsion droplets (a coacervate phase). The polymerrich droplets should then migrate to the oil/water interface, where they coalesce to form a polymer shell surrounding the oil core, as is depicted schematically in Figure 1. This (14) Kawashima, Y.; Niwa, T.; Handa, T.; Takeuchi, H.; Iwamoto, T.; Itoh, K. J. Pharm. Sci. 1989, 78, 68. (15) Khopade, A. J.; Caruso, F. Biomacromolecules 2002, 3, 1154. (16) Shi, X.; Caruso, F. Langmuir 2001, 17, 2036. (17) Watts, P. J.; Davies, M. C.; Melia, C. D. Crit. Rev. Ther. Drug Carrier Syst. 1990, 7, 235. (18) Arshady, R. J. Controlled Release 1991, 17, 1. (19) Pekarek, K. J.; Jacob, J. S.; Mathiowitz, E. Nature 1994, 367, 258. (20) Heya, T.; Okada, H.; Ogawa, Y.; Toguchi, H. Int. J. Pharm. 1991, 72, 199. (21) Heya, T.; Okada, H.; Tanigawara, Y.; Ogawa, Y.; Toguchi, H. Int. J. Pharm. 1991, 69, 69. (22) Crotts, G.; Park, T. G. J. Controlled Release 1995, 35, 91. (23) Benita, S. Microencapsulation: Methods and Industrial Applications; Dekker: New York, 1996. (24) Loxley, A.; Vincent, B. J. Colloid Interface Sci. 1998, 208, 49.

10.1021/la048561h CCC: $27.50 © 2004 American Chemical Society Published on Web 11/19/2004

Oil Core-Polymer Shell Microcapsules

Langmuir, Vol. 20, No. 26, 2004 11375

Figure 1. Schematic representation of the mechanism of shell formation (redrawn from ref 24).

Figure 2. Three particle morphologies: (a) Core-shell microcapsule; (b) acorn; (c) two separate droplets.

method has recently been extended by Atkin et al. for the preparation of microcapsules with aqueous cores from a water-in-oil emulsion.25 Analysis of the three interfacial tensions between the various different pairs of the three phases (oil, water, and polymer) allows the necessary conditions for microcapsule formation to be determined. Torza and Mason26 investigated the equilibrium morphology adopted by the droplets of three immiscible liquids. These morphologies are depicted schematically in Figure 2. As Loxley and Vincent24 showed, to make microcapsules, it is important that the oil/water interfacial tension is not reduced significantly by the emulsion stabilizer. This is in order to reduce the propensity for the oil/water interface to form (the oil/water interface is absent in the microcapsules in part a of Figure 2 but present in the other two parts). In the current study, polystyrene has been used as the shell-forming polymer, and the experimental factors governing microcapsule formation have been investigated. Continuous, in situ particle sizing measurements have been performed during both the emulsification stage and the subsequent evaporation of the volatile solvent. The mass of polymer used to form the shell and the rate of solvent removal have been systematically varied in order to determine the effect of these factors on shell thickness and shell integrity. The influence of these parameters on the release of a model active ingredient (AI) has also been studied. The active ingredient (AI) chosen was soluble in the oil core and sparingly soluble in the aqueous release medium. Dilution of the microcapsule dispersion with further water provided a driving force for release of the AI from the microcapsule. The amount released is related to the partition coefficient of the AI between the inner oil phase and the continuous water phase and the volume ratio of these two phases. However, if the AI also has a significant solubility in the polymer shell, this will lead to retention of some of the AI within the shell matrix and result in 97% unless otherwise stated. Deionized water (Purite) was used for the aqueous phase. Preparation of Microcapsules. The method used for the preparation of core-shell particles has been previously described in detail by Loxley and Vincent.24 Briefly, the required mass of polymer (1.8-8.0 g) and the active ingredient (4-nitroanisole, 57 mg) were dissolved in a mixture of dichloromethane (70.54 g) and hexadecane (3.88 g). Acetone (4 g) was added to this solution (to aid the subsequent emulsification process). An aqueous solution of 2 wt % poly(vinyl alcohol) was prepared, and 80 mL of this solution was placed in a 200 mL jacketed glass vessel, thermostated at 20 °C. This aqueous solution was sheared at 10 000 rpm using a Silverson stirrer. The nonaqueous phase was slowly added (over a 60 s period) to the aqueous solution to form an oil-in-water emulsion. Emulsification was continued for 1 h, after which the emulsion was added to 120 mL of 2% PVA solution. Formation of the polymer shell was brought about by removal of the dichloromethane. Variation of the evaporation rate used was studied. Rapid evaporation was effected by using a rotary evaporator, and slow evaporation, by stirring overnight at 40 °C, with residual dichloromethane being removed afterward by using rotary evaporation. In all cases, the final volume of the dispersion was adjusted to 200 mL. In some cases, oils other than hexadecane were used. Further control over the release kinetics may be possible by cross-linking the polystyrene, after the shell has formed, to reduce its porosity. Various methods of cross-linking the polystyrene shells with divinylbenzene (DVB) were investigated. The first route tried was to add the DVB after forming the microcapsules, assuming that it became solubilized in the polystyrene shells. The addition of a suitable initiator, dissolved in either the oil core (benzoyl peroxide, 0.1 g) or the aqueous continuous phase (V50, 0.1 g), effected polymerization of the DVB in the shell. It was found that adding the DVB to the oil phase (containing the dissolved polystyrene) prior to emulsification and initiating the cross-linking reaction after forming the microcapsules (by addition of V50 to the aqueous phase) gave the greatest control over release and was used for the cross-linking data presented in this study. Ternary Phase Diagram. The ternary phase diagram of polystyrene-hexadecane-dichloromethane was determined by mixing accurately weighed amounts of the component phases into glass phials. The solutions were stirred using a magnetic stirrer at ambient temperature to aid evaporation of the volatile dichloromethane. When the solution showed the first sign of turbidity (visually), due to polymer phase separation, the sample was reweighed. The loss in mass was attributed solely to the evaporation of the dichloromethane. In this way, the composition at this point on the phase boundary was established. Techniques for Characterization of the Shell. The morphology of the microcapsules was investigated using a Hitachi S-2300 scanning electron microscope (SEM). One drop of the microcapsule dispersion to be investigated was placed on a stainless steel SEM stub and allowed to air-dry overnight. Elevated temperatures and reduced pressures were avoided for this step in order to minimize the loss of core oil by evaporation, since this could lead to deflation or some other distortion of the microcapsules. The dried sample was gold-coated in an Edwards

11376

Langmuir, Vol. 20, No. 26, 2004

S150A sputter-coater. The chamber was evacuated to a pressure of ∼0.8 kPa, and a sputtering current of 20 mA was applied for 4 min, giving a gold coating with a thickness of ∼10 nm. To determine the polymer shell thickness, the microcapsules were fractured prior to gold-coating by applying direct pressure with a clean, round-tipped glass rod. To achieve brittle fracture of the capsules, the process was repeated under liquid nitrogen on freshly dried samples. An atomic force microscope (AFM) from EXFO Burleigh Instruments was used to characterize the pore size in the microcapsule shell. Optical microscopy was performed using a Nikon Optiphot microscope fitted with Nikon 320 and 340 objective lenses. Study of Particle Size during Emulsification and CoreShell Particle Formation. In situ particle size analysis was performed using a Lasentec M200L instrument. This instrument uses the focused-beam reflectance measurement (FBRM) technique, in which particles or droplets passing the sapphire window of the probe reflect a laser beam. The time over which reflectance occurs shows a direct dependence on particle size. The length scale recorded will vary depending on which part of the droplet or particle the beam strikes (since the wavelength of the laser used is much smaller than the particle size). In this way, a “chordlength” distribution of the particles is obtained; this relates to, but is not a true measure of, the actual particle size distribution. Measurements were made throughout the emulsification and evaporation stages, to study the size variation during the transition from oil-in-water emulsion droplets to microcapsules. Studies were performed as a function of the mass of polymer used to form the shell (3.8, 5, or 7 g) and the polymeric stabilizer (PVA) concentration (1 or 2 wt %). Release Studies. The active ingredient used in this study was 4-nitroanisole. This has a maximum solubility in hexadecane of 1.7% (w/w); it also has some solubility (589 mg L-1) in the aqueous release medium.27 This provides a driving force for release. The concentration of 4-nitroanisole in hexadecane used was 1.5% (w/w), which is below its maximum solubility. Release studies were performed by pouring a sample of the microcapsule dispersion (15 mL) into dialysis tubing, which was placed in distilled water (485 mL). 4-nitroanisole has a high UV molar extinction coefficient in water, which allows determination of low concentrations in the release medium by UV-visible spectroscopy. By measuring the UV absorbance of the release medium as a function of time, the release profile of the active ingredient could be determined.

Dowding et al.

Figure 3. Experimentally determined ternary phase diagram for polystyrene + dichloromethane + hexadecane. The composition used to prepare microcapsules (when no acetone was present) is shown by the symbol [. I and II indicate the singlephase and two-phase regions, respectively. The axes are marked in weight percent.

Results and Discussion Ternary Phase Diagram. The phase behavior of the three-component mixture of polystyrene + hexadecane + dichloromethane is presented in Figure 3. The low solubility of polystyrene in hexadecane makes it suitable for use as the shell polymer for hexadecane core microcapsules. The ternary phase diagram also provides information regarding the point at which phase separation will occur and gives an indication of the likely composition in the region where the solid polymer shell forms. For example, for a 3 g polystyrene system, phase separation is likely to occur at the following composition: 3 g of polystyrene, 3.88 g of hexadecane, and ∼28 g of dichloromethane. This represents a 60 wt % loss of dichloromethane (from the initial composition) before phase separation of the polymer begins to occur. Particle Size Variation during Emulsification and Core-Shell Particle Formation. Using the FBRM technique, in situ measurements of changes in particle size were monitored throughout the process, that is, during emulsification, while stirring overnight at 40 °C to effect dichloromethane removal, and during the final period of rotary evaporation. The data obtained for the 3.8 g polystyrene system are presented in Figure 4. The results (27) Gross, P. M.; Sauker, J. H.; Garman, M. J. Am. Chem. Soc. 1933, 55, 650.

Figure 4. Average length- and volume-weighted chord lengths obtained during emulsification (0-1 h), during evaporation (119 h), and after rotary evaporation (19-20 h) for a system containing 3.8 g of polystyrene.

obtained for other masses of polystyrene investigated (1.5, 5, and 8 g of polystyrene) were similar in form. It is possible to apply different weightings to the data obtained using the FBRM method, to give more emphasis to a particular region of the size distribution. A lengthweighted distribution gives greater sensitivity to smaller particles, whereas a volume-weighted distribution gives greater sensitivity to larger particles. The data in Figure 4 show that the particle size varies during the formation of the microcapsules. During the initial emulsification step (0-1 h), the oil phase is dispersed into droplets, leading to the observed reduction in chord length over this first period. Over the next period (1-19 h), during which the system is stirred and the temperature is maintained at 40 °C, slow evaporation of the dichloromethane occurs. Rather unexpectedly, the average size of particles does not decrease, as the droplets release their dichloromethane, but, if anything, the average droplet size appears to actually increase over this period (although, in the final rotary evaporation period, to remove the final traces of dichloromethane, there is an, albeit small,

Oil Core-Polymer Shell Microcapsules

Figure 5. (a) SEM micrograph of core-shell particles (5.0 g of polystyrene) showing cracking in the particle surface. (b) SEM micrograph of cryofractured core-shell particles (3.8 g of polystyrene), showing particle deformation. (c) SEM micrograph of cryofractured core-shell particles (3.8 g of polystyrene), showing “smashed” particles. (d) SEM micrograph of coreshell particles embedded in resin and sliced using a microtone knife (3.0 g of polystyrene).

decrease in particle size). The phase diagram in Figure 4 suggests that shell formation will occur when ∼60% of the dichloromethane has evaporated. This should equate to a reduction in radius of ∼35%. The fact that an increase is observed probably indicates that some coalescence and/ or Ostwald ripening is occurring for the droplets during the evaporation stage. Loxley and Vincent24 reported a similar phenomenon for PMMA microcapsules prepared using a very similar method and suggested that coalescence of droplets prior to the solid microcapsule shell forming could lead to the observed nonreduction in radius during removal of the volatile solvent from the oil droplets. Shell Morphology. As long as the concentration of the AI in the oil core remains relatively high, the diffusion of the AI through the shell material will control the net release rate. However, the presence of cracks or pores in the shell will reduce the barrier to diffusion and alter the kinetics of release.28 SEM studies (Figure 5a) have shown that cracks in the polymer shells are indeed present. An atomic force microscope (AFM) was used to further investigate the surface of the microcapsule and, in particular, the cracks that were revealed by the SEM. The advantage of this method is that it does not require the surface to be dried prior to study. For these studies, microcapsules were prepared from an oil (hexadecane + dichloromethane + 2.0 g of polystyrene)/water emulsion which had been formed using a paddle stirrer (580 rpm), to produce larger particles (∼100 µm) suitable for AFM study. The volatile solvent evaporation process was performed at two different rates, rapid evaporation using a rotary evaporator and slow evaporation by simply stirring the emulsion overnight, followed by a short period of rotary evaporation. Cracks in the surface of the particles were also seen in the AFM images (not shown here) for both the fast and slow solvent removal routes. However, the size of the cracks was dependent on the evaporation rate: when the solvent was removed quickly, larger cracks or pores were evident. Unfortunately, it was not really possible to obtain a pore size distribution from these AFM images. (28) Athanasiou, K. A.; Nierderauer, G. G.; Agrawal, C. M. Biomaterials 1996, 17, 93.

Langmuir, Vol. 20, No. 26, 2004 11377

Figure 6. Effect of the dichloromethane evaporation rate (for a 3.0 g polystyrene system) on the release profile of 4-nitroanisole: (O) evaporation by rotary evaporator; (2) sample stirred for 24 h; (9) sample stirred for 24 h, followed by rotary evaporation. The inset shows the release data over the first 10 h. The lines are drawn to guide the eye.

It is important to correlate the amount of polymer used (1.75-8 g) in forming the microcapsules with the resultant shell thickness. The most direct method for determining the shell thickness is by fracturing the core-shell particles (preferably along the equatorial plane) and measuring the thickness directly from SEM micrographs.24 To this end, some initial experiments were performed using cryofracturing. This method involves placing a sample of particles between two SEM stubs, which are then placed in liquid nitrogen. The two stubs are then broken apart, fracturing the particles, which are allowed to warm to room temperature, prior to gold-coating. Unfortunately, simple hemispheres of core-shell particles were not observed. Indeed, deformation of particles was noted (shown in Figure 5b), suggesting some flexibility of the polymer shell. In some cases, “smashed” particles were also observed (Figure 5c). To overcome these problems, the microcapsules were freeze-dried and embedded in a polymer resin, which was then cured. Thin slices of this solid matrix were then produced using a microtone knife. A typical micrograph, for the system with 3.0 g of polystyrene, is shown in Figure 5d. It may be seen that the shell is not uniform in thickness, but the average thickness (taken over several such micrographs) is ∼75 nm and represents ∼25% of the microcapsule radius. When the mass of polystyrene was increased to 5.0 g, the shell thickness also increased (to ∼90 nm). However, it was difficult to obtain a sufficient number of reasonable images to obtain accurate values for the shell thickness of the other microcapsules studied, with different polystyrene amounts present. Suffice to say, qualitatively, the shell thickness did appear to increase with the amount of polystyrene present, as expected. Factors Influencing Release Rate. (a) Rate of Volatile Solvent Removal and Crack Size. Figure 6 shows how the method (and hence the rate) of evaporation of the dichloromethane affects the release rate profile for 4-nitroanisole. As discussed above, rapid removal of the volatile solvent (by rotary evaporation) leads to the creation of larger cracks or pores in the microcapsule wall than does slower evaporation (stirring overnight). There is a good correlation between the crack size and the 4-nitroanisole release profiles. Rapid removal of dichloromethane, under reduced pressure using a rotary evaporator, results in the quickest release rate, with ∼80% release within ∼8 h. If the rate of evaporation is reduced (by stirring overnight at room temperature and pressure),

11378

Langmuir, Vol. 20, No. 26, 2004

Figure 7. Effect of variation in the mass (grams) of polystyrene present, and hence the shell thickness, on the release profile of 4-nitroanisole: (O) 1.75 g; (2) 3.8 g; (9) 5 g; ([) 8 g. The lines are drawn to guide the eye. The inset shows the release of 4-nitroanisole over a longer time scale (up to 150 h).

Dowding et al.

Figure 8. Effect of cross-linking on the release profile of 4-nitroanisole (for a 3.8 g polystyrene system): (O) un-crosslinked polystyrene; (×) cross-linked polystyrene (10 wt % DVB on polystyrene). The lines are drawn to guide the eye. Table 1. Oil Phases Used to Form the Microcapsulesa

the rate of release is reduced, and 80% release is not reached until ∼25 h. These results clearly reflect the crack sizes in the shell walls. If, after stirring overnight, the sample is rotary evaporated for 1 h, 80% release is not achieved until ∼40 h. This suggests that the final rotary evaporation step is important in removing the final traces of dichloromethane form the microcapsules. Probably some dichloromethane may still be present in the polymer wall after stirring, allowing slightly faster diffusion of the AI through the shell wall. The removal of the remaining good solvent by rotary evaporation reduces this diffusion rate somewhat. (b) Effect of Shell Thickness. The effect of an increase in shell thickness, brought about by increasing the total mass of polymer present in the system, on the release profile for polystyrene microcapsules is shown in Figure 7. As the width of the microcapsule shell is difficult to measure accurately (as described above), the thickness is described in terms of the mass of polymer used to form the shell. The fastest release is obtained with the smallest amount of polystyrene present (1.75 g); that is, for the thinnest shells, 80% release is reached within 5 h. If the mass of polystyrene is increased to 3.8 g, 80% release is not attained until 24 h, reflecting the corresponding increase in shell thickness. No difference in the release profile for the 3.8 and 5 g polystyrene systems can be detected for the first 24 h of the release experiment. However, the inset of Figure 7 shows the release profile up to 150 h. On this scale, a slight reduction in the rate of release for the 5 g system is observed. The slowest release was seen determined for the 8 g polystyrene system (i.e., the thickest-walled microcapsules), where 80% release is not attained until ∼80 h. It may also be seen from the inset of Figure 7 that increasing the shell thickness also leads to a reduction in the final released amount (yield). The most likely explanation for this has to do with the distribution of the 4-nitroanisole between the continuous phase, the oil cores, and the polymer sheath. If the polymer sheath retains a significant quantity of 4-nitroanisole, then less will diffuse out the thicker the shell wall. (c) Cross-Linking of the Polymer Shell. The effect of cross-linking the polystyrene shell with DVB is shown in Figure 8. It can be seen that cross-linking the polymer shell reduces the rate of release, consistent with the production of a more homogeneous, less porous polymer shell. Cross-linking also has a strong effect on the maximum yield. This would suggest that a greater amount

oil

hexadecane

butyl ether

density/(g dm-3) dipole moment/D Kw/o Kp/o

0.773 0 7.8 × 10-3 5.2 × 10-3

0.764 1.2 6.0 × 10-3 4.4 × 10-3

ethyl isovalerate 4-heptanone 0.864 1.9 2.3 × 10-3 1.2 × 10-3

0.817 2.7 1.8 × 10-3 9.6 × 10-4

a K w/o is the partition coefficient of 4-nitroanisole between water and oil; Kp/o is the partition coefficient of 4-nitroanisole between polystyrene and oil.

Figure 9. Influence of the core oil type on the release rate and yield of 4-nitroanisole (for a 3.0 g polystyrene system): (O) hexadecane; (9) butyl ether; (2) ethyl isovalerate; ([) 4-heptanone. A fixed mass of core oil (3.88 g) was used in the preparation of the microcapsules. The lines are drawn to guide the eye.

of 4-nitroanisole is retained in the cross-linked shell than in the un-cross-linked ones. (d) Nature of the Oil Core. In all the work presented above, hexadecane was used as the oil phase. Three other oils were tried, and microcapsules were prepared successfully from all of them. These oils and some of their properties are listed in Table 1. Their densities are not significantly different. The partition coefficients of 4-nitoanisole between water and the various oils (Kw/o) and between polystyrene and the various oils (Kp/o) were very simply determined by adding a known quantity of water or polystyrene to a given volume of a solution of 4-nitroanisole in the oil concerned and determining the change in its concentration in the oil phase after equilibration. It can be seen that both sets of partition coefficients decrease in the order of increasing polarity of the oil (as reflected in its dipole moment), and therefore increasing

Oil Core-Polymer Shell Microcapsules

relative solubility of 4-nitoanisole in the oil concerned (for example, the solubility of 4-nitroanisole increases from 1.7 wt % in hexadecane to 10 wt % in butyl ether). The release profiles of 4-nitroanisole from microcapsules produced with the four oils are shown in Figure 9. Both the release rate and the yield decrease in the following order: hexadecane > butyl ether > ethyl valerate > 4-heptanone. This is the same order of increasing polarity, and decreasing partition coefficients, shown in Table 1. The oil with the largest water/oil partition coefficient (i.e., hexadecane) should have the greatest yield, and this is the case. The fact that the initial release rates are not that significantly different for the four oils suggests that the higher partition coefficients of the less polar oils between the shell polymer and the oil phase play an important role in retarding diffusion of the 4-nitroanisole out of the polymer shells in their case. Conclusions Oil core microcapsules having a polystyrene shell, suitable for the controlled release into water of an organic active ingredient, have been prepared by phase separating the polymer from the internal phase of an o/w emulsion. For such a system, correct choices of the shell polymer, volatile solvent, nonvolatile core oil, and w/o emulsion stabilizer are critical to the successful formation of microcapsules. Their release characteristics are dependent on the method of microcapsule preparation. The rate of evaporation of the good solvent influences the porosity of

Langmuir, Vol. 20, No. 26, 2004 11379

the polymer shell and hence the release characteristics. Increasing the shell thickness was found to reduce the release rate and the release yield. The reduction in the final yield with thicker shells is due to a greater amount of the active ingredient being retained in the shell. The reduced rate of release is due to a slower rate of diffusion of AI through the pores in the thicker polymer shell. Crosslinking the polymer also reduces the release rate and yield. Finally, it has been shown that the nature of the core oil is important in controlling the release rate and yield. The greater the polarity of the core oil, the more soluble the active ingredient, and the lower the final yield in the aqueous phase. The effect of the oil phase polarity on the (initial) release rate was found to be less significant, due to the competing retention of the active ingredient in the polymer shell, the greater its partition coefficient between the polymer and the oil. Acknowledgment. The authors would like to thank the former Astra Zeneca Company (now Syngenta) for initial financial support, in particular for P.B. and P.D.. We would also like to thank P&G and EPSRC (GR/ R90086/01) for subsequent financial support through the IMPACT Faraday Partnership for R.A.. Drs David Rodham (now at IMPACT) and Ian Shirley (now at Syngenta) are also thanked for very fruitful discussions during the early stages of this work. LA048561H