Polymer Shell Microcapsules by Internal Phase Separation

K KOOIMAN , M BOHMER , M EMMER , H VOS , C CHLON , W SHI , C HALL , S DEWINTER , K SCHROEN , M VERSLUIS. Journal of Controlled Release 2009 ...
0 downloads 0 Views 138KB Size
Download PDF
5278

Langmuir 2005, 21, 5278-5284

Oil Core/Polymer Shell Microcapsules by Internal Phase Separation from Emulsion Droplets. II: Controlling the Release Profile of Active Molecules Peter J. Dowding, Rob Atkin, Brian Vincent,* and Philippe Bouillot School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, United Kingdom Received November 28, 2004. In Final Form: March 30, 2005 Microcapsules with oil cores and solid polymer shells have been prepared by precipitation of the polymer from the internal phase of an oil-in-water emulsion. The dispersed phase consists of a polymer, a good solvent for the polymer (dichloromethane), and a poor solvent for the polymer (hexadecane). Removal of the good solvent results in phase separation of the polymer within the emulsion droplet, leading to the formation of a polymeric shell surrounding the poor solvent. A UV-active organic molecule is added to the oil phase prior to emulsification. Provided this molecule has some water solubility, the release profile of the molecule from the capsule can be determined. While the microcapsule size was kept approximately constant, the influence of a wide range of factors on the release profile has been studied. These include the type and molecular weight of the shell-forming polymer, the molecular weight of the active ingredient molecule, the shell thickness, the use of copolymers or polymer blends to form the shell, the effect of cross-linking the shell or heating the capsule to temperatures above the Tg value of the polymer after the shell has been formed, and the effect of changes in the pH of the release solution in the case when a weak polyelectrolyte is used as the shell polymer. The differences in behavior are discussed in terms of the properties of the polymer shell, in particular the thickness, the polymer/release molecule interaction, and the free volume/porosity. Variation of these parameters allows one to control both the final release yield and the rate of release for time periods between a few hours and days.

Introduction The ability to encapsulate active ingredients such as drugs, pesticides, herbicides, enzymes, and dyes has many applications. In general, encapsulation is performed in order to provide protection and/or controlled release.1 Microcapsules, which consist of a liquid core and a solid semipermeable shell, can be formed by a wide variety of methods, including interfacial polymerization reactions involving emulsions2-5 and microemulsions,6-8 spray drying,9,10 phase separation,11,12 solvent extraction,13,14 layer-by-layer addition,15,16 solvent evaporation,17,19 and the use of multiple emulsions.17,20-22 It is important that the microcapsules do not react adversely with the active ingredient or the release medium. Their utility for any given application is deter* To whom correspondence should be addressed: te: +44 (0)117 9288160; fax +44 (0)117 9250612; [email protected]. (1) Arshady, R. Microspheres, Microcapsules & Liposomes, Plenum: New York, 1998. (2) Janssen, L.; Tenijenhuis, K. J. Membr. Sci. 1992, 65, 59-73. (3) Lambert, G.; Fattal, E.; Pinto-Alphandary, H.; Gulik, A.; Couvreur, P. Pharm. Res. 2000, 17, 707-712. (4) Fallouh, N. A.; Roblot-Treupel, L.; Fessi, H.; Devissaguet, J. Ph.; Puisieux, F. Int. J. Pharm. 1986, 28, 125-132. (5) Fresta, M.; Cavallaro, G.; Giammona, G.; Wehrli, E.; Puglisi, G. Biomaterials 1996, 17, 751-758. (6) Watnasirichaikul, S.; Davies, N. M.; Rades, T.; Tucker, I. G. Pharm. Res. 2000, 17, 684-690. (7) Hong, K.; Park, S. Mater. Chem. Phys. 1999, 58, 128-131. (8) Daubresse, C.; Grandfils, C.; Jerome, R.; Teyssie, P. J. Colloid. Interface. Sci. 1994, 168, 222-229. (9) Mathiowitz, E.; Bernstein, H.; Dor, Ph.; Turek, T.; Langer, R. J. Appl. Polym. Sci. 1992, 45, 125-134. (10) Shi, X. Y.; Tan, T. W. Biomaterials 2002, 23, 4469-4473. (11) Ruiz, J. M.; Busnel, J. P.; Benoit, J. P. Pharm. Res. 1990, 7, 928-934. (12) Vachon, M. G.; Nairn, J. G. J. Microencapsul. 1995, 12, 287305. (13) Leelarasamee, N.; Howard, S. A.; Malanga, C. J.; Ma, J. K. H. J. Microencapsul. 1988, 5, 147-157. (14) Kawashima, Y.; Niwa, T.; Handa, T.; Takeuchi, H.; Iwamoto, T.; Ito, Y. Chem Pharm Bull. 1989, 37, 425-429.

mined primarily by the release profile of the active ingredient.23 For a given core liquid and release medium, the release profile is determined by the physical size of the microcapsule, the solubility of the active ingredient in the polymer shell, and the release medium and the properties of the microcapsule wall, primarily its thickness and permeability. The need to achieve good control over the time-release profile, for both sustained release and targeted release applications, is a primary driving force for the continued development of new microcapsules. Many studies have investigated the influence of shell thickness and porosity on the release profile for hydrophilic polymer shells. By monitoring the conductivity of the aqueous release medium, Jalsenjak and Kondo24 were able to determine the time-release profile of NaCl from gelatinacacia microcapsules formed by external coacervation. For similarly prepared microcapsules, Senjkovic and Jalsenjak25 measured the apparent diffusion coefficient of sodium phenobarbitone, as a function of the microcapsule size. It was found that, for these ethyl cellulose microcapsules, the diffusion coefficient deceased as the microcapsule size was decreased. For polyamide microcapsules, formed by interfacial polymerization, Mathiowitz and Cohen26 have (15) Sukhorukov, G. B.; Brumen, M.; Donath, E.; Mohwald, H. J. Phys. Chem. B. 1999, 103, 6434-6440. (16) Shi, X.; Caruso, F. Langmuir 2001, 17, 2036-2042. (17) Park, S. J.; Kim, S. H. J. Colloid. Interface. Sci. 2004, 271, 336341. (18) Arshady, R. J. Controlled Release 1991, 17, 1-21. (19) Pekarek, K. J.; Jacob, J. S.; Mathiowitz, E. Nature 1994, 367, 258-260. (20) Heya, T.; Okada, H.; Ogawa, Y.; Toguchi, H. Int. J. Pharm. 1991, 72, 199-205. (21) Heya, T.; Okada, H.; Tanigawara, Y.; Ogawa, Y.; Toguchi, H. Int. J. Pharm. 1991, 69, 69-75. (22) Crotts, G.; Park, T. G. J. Controlled Release 1995, 35, 91-105. (23) Benita, S. Microencapsulation: Methods and Industrial Applications, Dekker: New York, 1996. (24) Jalsenyak, I.; Kondo, T. J. Pharm. Sci. 1981, 70, 456-57. (25) Senjkovic, R.; Jalgenjak, I. J. Pharm. Pharmacol. 1981, 33, 279282.

10.1021/la0470838 CCC: $30.25 © 2005 American Chemical Society Published on Web 05/05/2005

Oil Core/Polymer Shell Microcapsules

shown that the rate of azobenzene release decreases with increasing shell thickness and also with the addition of a cross-linking molecule. It was also demonstrated that the properties of the capsule varied with the species of amine used to form the shell. A decrease in release rate with shell thickness has also been reported by Yadav et al.27 for release from polyurea microcapsules. Shulkin and Sto¨ver28,29 have shown that, for capsules formed by the interfacial reaction between anhydride copolymers and water-soluble polyamines, a thick, porous shell was more effective at reducing the rate of release than a thin, dense shell. In the present study, we present release data for microcapsules formed by internal phase separation.30 The effects of a wide variety of factors on the release profile have been investigated. A number of different polymers have been used to form the capsule shell. For polystyrene microcapsules, we previously demonstrated that increasing the shell thickness reduces the rate of release.31 A similar study but with various different polymers, including copolymers and polymer blends, has been conducted in this work. The effect of increasing the molecular weight of the polymer used to form the shell has been investigated. Several methods of reducing the rate of release, after the capsule has been formed, have also been studied. These include the effects of post-cross-linking the shell to reduce its porosity and of heating the microcapsules above the Tg of the polymer shell. The use of different molecular mass model active ingredients has also been investigated. It should be noted that similar chemistry has recently been used by us to prepare microcapsules with aqueous cores and polymer shells.32 Materials and Methods Materials. The following chemicals were obtained from Aldrich and were used without further purification unless otherwise stated. All purities were >97%. Polystyrene (average molecular mass ≈ 280 000 g mol-1), poly(vinylphenyl ketone) (PVPK, average molecular mass ) 1350 g mol-1), poly(methyl methacrylate) (PMMA, average molecular mass ) 120 000 g mol-1), poly(isobutyl methacrylate) (PIBMA, average molecular mass ) 120 000 g mol-1), poly(2-vinylpyridine) (PVP, molecular mass ) 40 000 and 320 000 g mol-1), poly(2-vinylpyridine-costyrene) copolymers (with 10 and 80 wt % styrene) (PVP-co-PS), poly(ethyl methacrylate-co-methacrylate) (PEMA-co-MA, molar ratio of monomers 1:0.16), poly(vinyl alcohol) (PVA, molecular mass ) 95 000 g mol-1, 96% hydrolyzed), dichloromethane, 4-nitroanisole, 2-propylpyridine, acridine, Sudan 2, Sudan 3, and n-hexadecane purified by repeated passage over an alumina column prior to use. Deionized water (Purite) was used for making the aqueous phase. Preparation of Microcapsules. The method used for the preparation of core-shell microcapsule particles by internal coacervation has been described in detail previously by Loxley and Vincent30 and in a recent paper by us.31 Briefly, the required mass of polymer (1.8-8.0 g) and the model active ingredient were dissolved in a mixture of dichloromethane (70.5 g) and hexadecane (3.88 g), along with acetone (4 g) to aid emulsification. An aqueous solution of 2 wt % poly(vinyl alcohol) was prepared. A portion (80 mL) of this solution was placed in a 200 mL jacketed (26) Mathiowitz, E.; Cohen, M. D. J. Membr. Sci. 1989, 40, 27-41. (27) Yadav, S. K.; Kartic, C.; Khilar, A. K.; Suresh, A. J. Membr. Sci. 1997, 125, 213-218. (28) Shulkin, A.; Sto¨ver, H. D. H. J. Membr. Sci. 2002, 209, 421432. (29) Shulkin, A.; Sto¨ver, H. D. H. J. Membr. Sci. 2002, 209, 433444. (30) Loxley, A.; Vincent, B. J. Colloid. Interface. Sci. 1998, 208, 4962. (31) Dowding, P. J.; Atkin, R.; Vincent, B.; Bouillot, P. Langmuir 2004, 20, 11374-11379. (32) Atkin, R.; Davies, P.; Hardy, J.; Vincent, B. Macromolecules 2004, 37, 7979-7985.

Langmuir, Vol. 21, No. 12, 2005 5279 glass vessel, thermostated at 20 °C. A Silverson stirrer was used to shear this aqueous solution at 10 000 rpm for 1 h. The multicomponent oil phase was added to the aqueous solution over a 60 s period. An oil-in-water emulsion resulted. After stirring, the emulsion was poured into 120 mL of the PVA solution. The polymer shell formed when the dichloromethane was removed. This was accomplished by stirring overnight at 40 °C, with any residual dichloromethane being subsequently removed by rotary evaporation. This method has been found to lead to the least porous shells.31 The final volume of the dispersion was adjusted to 200 mL with water. In some cases, after formation of the microcapsules, their shells were post-cross-linked. This was achieved for PS and PVPK by swelling the shells with a difunctional monomer (divinylbenzene or ethylenediamine, respectively), and then absorbing into the shell a suitable initiator and heating to effect the cross-linking reaction. For PVP-based microcapsules this was achieved by adding 1,6-dibromohexane to the system and heating to effect cross-linking (by hexane chains) between the N-atoms of the pyridine rings, by elimination of HBr.33 The size of the microcapsules was determined by dynamic light scattering. Although they were somewhat polydisperse, their average diameter was in all cases ∼2 µm. Release Studies. The active ingredients used in this study are listed in Table 1. The active ingredient must be soluble in the oil phase (hexadecane) and also sparingly soluble in the aqueous release medium, as this provides the driving force for release. The O/W partition coefficients are given in Table 1. Unless otherwise stated, the mass of the active ingredient dissolved initially in the oil phase was 57 mg. Release studies were performed by adding a fixed concentration of the core-shell particle dispersion (15 mL) to cleaned dialysis tubing, which was then placed in distilled water (485 mL). All the release molecules used have a high molar UV-visible extinction coefficient. Thus, by measuring the absorbance of the release medium, at a suitable wavelength, as a function of time, the release profile of the active ingredient can be determined. By allowing the experiments to run for sufficient time, the release yield could be determined. This is defined as the mass of the active ingredient released from the microcapsule suspension divided by the mass of the active ingredient initially dissolved in the oil phase.

Results and Discussion Variation in Polymer Shell Type. In previous work from this group, microcapsules with shells formed with PMMA30 and PS31 were prepared by precipitating the polymer from the internal phase of an appropriate oil/ water emulsion. In the present work, microcapsules with PVPK, PEMA-co-MA, and PIBMA shells have also been prepared by this method. Data for the release of 4-nitroanisole from all five capsule types are presented in Figure 1. It is clear that the type of polymer used to form the shell influences both the release kinetics and the release yield. By comparing Figure 1 panels a and b, one may see that these two parameters are coupled: generally, the faster the initial release rate, the higher the yield. Hence, PS and PIBMA have the highest rate and yield, PEMA-coMA the next highest, and PMMA and PVPK have the lowest rate and yield. It is difficult to rationalize these differences explicitly, since the release of 4-nitroanisole from the microcapsules will be governed by a number of factors. These include the size of the microcapsule, the thickness and porosity of the polymer shell, and the relative solubilities of 4-nitroanisole in, and the respective total volumes of, the oil core, the polymer shell, and the aqueous medium. However, for the experiments presented in Figure 1, several of these parameter were kept approximately constant, namely, the size of the micro(33) Loxley, A.; Vincent, B. Colloid Polym. Sci. 1997, 275, 11081114.

5280

Langmuir, Vol. 21, No. 12, 2005

Dowding et al. Table 1. Model Active Ingredients

active ingredient

structure

MW

O/W partition coefficient

2-propylpyridine

121.18

86.3

4-nitroanisole

153.14

13.4

acridine

179.22

87.7

Sudan 2

276.34

Sudan 3

352.4

capsules and the thickness of the polymer wall (by using the same weight of polymer in each case) and also the masses of the oil core, the polymer shell, and the final amount of the water phase. Since the partition coefficient of 4-nitroanisole between the oil phase and the aqueous phase is fixed, then the only variables remaining, for the data shown in Figure 1 are the partition coefficient of 4-nitroanisole between the polymer shell and the oil core and the porosity of the polymer shell. Clearly, if the shells are truly solid (i.e., no micropores or larger pores), then the higher the partition coefficient between the shell polymer and the oil core, the lower the yield will be and

Figure 1. Effect of polymer type on the release profile of 4-nitroanisole (for a 3.0 g shell): (O) PS, (2) PIBMA, (9) PEMAco-MA, ([) PMMA, and (×) PVPK; (A) over 80 h; (B) first 5 h only, to show the initial release rate more clearly.

108

9.3

the slower the release rate will be, as diffusion will be retarded by a stronger 4-nitroanisole/polymer interaction (the relative free volume of the solid polymer, which will depend on molecular weight as well as polymer type, will also play a role but a minor one relatively). If there is a significant percentage of pore volume present in the shells (particularly of connected, open pores or channels), then although the (equilibrium) yields for the different polymer shells would not be changed much, there would be much faster release rates. In fact, one would not expect to see much difference in the initial release rates between the different polymers (in contrast to Figure 1b), as diffusion from the core to the outside would now occur mainly through the pores, rather than the polymer itself. Variation in the Nature of the Release Molecule. The release of a series of model active ingredients (2propylpyridine, 4-nitroanisole, acridine, Sudan 2, and Sudan 3, in order of increasing molecular weight; see Table 1) has been investigated through microcapsules with polystyrene shells, having a fixed shell thickness (with 3.8 g of polymer). Molecular structures and the oil/water partition coefficients are given in Table 1. Release data for propylpyridine, 4-nitroanisole, and acridine are presented in Figure 2. The first point to note is that the yields obtained, after long times, are not related to the oil/water partition coefficients listed in Table 1. Rather, they must be related to the partition coefficients of the release molecules between the polymer (polystyrene) and the oil. The fact that with propylpyridine the yield is ∼100% indicates that this molecule is not retained in the shell. Furthermore, it would seem that acridine (with a yield of only 60%) is more strongly retained in the polystyrene shell than 4-nitroanisole (yield 80%). It is of interest that using a higher initial amount of acridine leads to a higher yield. This must reflect the fact that the polymer shell becomes saturated in this release molecule for both initial concentrations in the oil. Thus, for the higher initial concentration system, a higher percentage of the acridine molecules will be released into the water. Again, since the initial release rates for the different release molecules (see Figure 2b) do differ somewhat, this suggests that continuous, open macropores/channels do not play a significant role in these systems. Figure 3 shows the release profiles for the somewhat larger Sudan 2 and Sudan 3 molecules (see Table 1). The shape of the profiles is now rather different than those for

Oil Core/Polymer Shell Microcapsules

Langmuir, Vol. 21, No. 12, 2005 5281

Figure 4. Effect of variation of the thickness (mass) of the shell on the release profile of 4-nitroanisole: for PVPK, (9) 3.8 g, (2) 5 g, and ([) 8 g; for PMMA, (0) 2.5 g, (]) 3.0 g, and (O) 3.8 g.

Figure 2. Release studies from polystyrene- (3.8 g) based microcapsules, for the following model release compounds: (A) (O) propylpyridine (57 mg), (4) 4-nitroanisole (57 mg), ([) acridine (57 mg), and (9) acridine (67 mg). (B) First 5 h only, to show the initial release rate more clearly. (The amounts in parentheses refer to the mass dissolved in the oil phase initially.)

Figure 3. Release studies from polystyrene- (3.8 g) based microcapsules, for the following model release compounds: (O) Sudan 2 (57 g) and ([) Sudan 3 (57 g). (The amounts in parentheses refer to the mass dissolved in the oil phase initially).

the smaller molecules shown in Figure 2. The first fact to note is that for both Sudan molecules the final yield is ∼100%, indicating that neither species is ultimately retained in the polystyrene shells of the microcapsules. Also, the smaller of the two species (Sudan 2) is released faster. The unexpected result is the two-step profile, apparent in both cases, with an initial slow step, followed by a more rapid second step, before finally coming to equilibrium, at times much longer than for the smaller molecules (Figure 2). These results could be indicative of the presence of small pores in the shell, comparable in size to, though probably somewhat bigger than, the Sudan molecules. Diffusion from the shell into the aqueous phase

is controlled essentially by the concentration of molecules at the periphery of the shell; these are replaced by molecules diffusing across the shell, and these in turn are replaced by molecules diffusing into the shell from the internal oil phase. A constant release rate (zero-order kinetics) implies a constant concentration of molecules at the shell periphery. It may well be that after some time this concentration at the periphery increases, because somehow the pores have been enlarged/opened up by the molecular transfer process itself across the shell. This would account for the increase in release rate in the second step. Effect of Shell Thickness. For microcapsules prepared by internal coacervation, the shell thickness is controlled by varying the mass of polymer dissolved in the oil phase of the emulsion. It was reported in a previous paper by us31 that increasing the shell thickness decreased both the release rate and yield of 4-nitroanisole for PS-based microcapsules. Figure 4 shows that increasing the mass of polymer used, and thus the shell thickness, for PVPKand PMMA-based microcapsules also results in a reduced release rate and yield. Other studies have reported similar findings.26,27 Increasing shell thickness leads to a greater amount of active ingredient being retained within the shell, and hence a reduction in the release yield. The effect of variation in the mass of polymer used to form the shell on the average size of the microcapsules formed was investigated by dynamic light scattering. The particles with the lowest shell thickness were approximately 2 µm in diameter. In general, there was an increase (up to 25%) in the capsule size when a greater mass of polymer was used to form the microcapsule shell. Effect of Cross-Linking the Polymer Shell. To obtain further control over the release profile, an attempt was made to reduce any free volume or pores in the shell by cross-linking the shell after the microcapsules had been formed.26 The release profiles, for PS microcapsules crosslinked with divinylbenzene and for PVPK microcapsules cross-linked with ethylenediamine, are presented in Figure 5. It is apparent that cross-linking of the polymer shell does indeed reduce the rate of release. It would seem that there is also a reduction in the final yield, even though the amount released still appears to be rising slowly after 80 h (when the experiment ceased). This implies that less 4-nitroanisole is being retained in the shell for the crosslinked systems. So this would suggest that some of the 4-nitroanisole, in the unmodified shells, does actually reside in the small pores thought to be present. Upon cross-linking, the volume of these pores is reduced significantly.

5282

Langmuir, Vol. 21, No. 12, 2005

Figure 5. Effect of cross-linking on the release profile of 4-nitroanisole: (O) PS, (×) PS cross-linked (10 wt %) with DVB, (2) PVPK, and ([) PVPK cross-linked (10 wt %) with ethylenediamine. A 3.8 g portion of polymer was used to form the shell in each case.

Figure 6. Effect of heating the shell polymer above its Tg value on the release profiles (at room temperature) of 4-nitroanisole from microcapsules with various polymer shells: ([, ]) PVPK (Tg ) 58 °C); (b, O) PIBMA (Tg ) 55 °C); (2, 4) PEMA-co-MA (Tg ) 48 °C). Solid symbols, system not heated; open symbols, system heated to 10 °C above the Tg value of the polymer.

Effect of Heating above the Tg Value of the Polymer Shell. Microcapsules with PVPK, PIBMA, and PEMA-co-MA shells were heated for 2 h at a temperature 10 °C greater than the glass transition temperature (Tg) of the polymer. The hypothesis is that, above Tg, the extra mobility afforded to the polymer chains will lead to better packing of the polymer chains in the shell. This morphological change should be retained upon cooling, thereby hindering release from the microcapsule. It is possible, of course, that there might be some commensurate changes in the shell thickness, which would also change the release rate. These results are presented in Figure 6. For the three polymer shell types studied, heating the microcapsules above the Tg of the polymer does indeed lead to a significant reduction in the release rate of 4-nitroanisole in all three cases. This reduction in release rate was not, however, accompanied by a reduction in the final release yield, so that the uptake of 4-nitroanisole into the polymer shell is not greatly affected by altering the packing of the polymer chains. Thus, heating microcapsules above the Tg of the shell polymer is an excellent method for reducing release rates, as this modification process is relatively straightforward and effective. Poly(vinylpyridine) Shells: Effects of pH and PVP Molecular Weight. Poly(vinylpyridine) (PVP), in aqueous solution, is a weakly basic polymer, becoming positively charged below about pH 5 due to protonation of the N-atoms in the pyridine rings. It is expected, therefore,

Dowding et al.

Figure 7. Effect of the polymer molecular mass and the pH on the release profile of 4-nitroanisole from PVP-based microcapsules (5% cross-linked): 40 000 g mol-1, (O) pH 3 and (4) pH 9; 320 000 g mol-1, (b) pH 3 and (2) pH 9.

that microcapsules prepared with PVP would swell at low pH values, due to the charging of the polymer chains, leading to an increase in release rate. The release of 4-nitroanisole from PVP-based microcapsules has been studied for two PVP molecular masses, 40 000 and 320 000 g mol-1, at two pH values for the aqueous phase, 3 and 9. Note that, as stated earlier in the Materials and Methods section, the PVP polymer shells had been crosslinked after their formation, by reaction with 5 wt % dibromohexane. The results are shown in Figure 7. The differences in the four experiments, observed over the full period of study (10 h), do not appear to be very significant, and effectively one curve may be drawn through the data points. However, an inset has been included that shows the data over the first 100 minutes of the experiment. Here some distinct trends may be observed, although again the differences are not great. One point to note straight away from the longer-time data set is that the release rates are somewhat higher compared to the other polymers used (compare Figures 1 and 7) and the release yield approaches 100% in all cases, suggesting that very little, if any, of the 4-nitroanisole is retained in the PVP shells. PVP is more hydrophilic than any of the other polymers studied, and the shells might well contain some water for this reason, making them more permeable to 4-nitroanisole. It can be seen from Figure 7 (inset) that increasing the molecular weight of the shell polymer (prior to crosslinking with dibromohexane in situ in the shell) does lead to a small reduction in the release rate, at least at short times. If this small difference is real (i.e., outside the experimental error) then it maybe that there are more chain entanglements in the case of the higher molecular weight PVP, reducing the pore free volume somewhat and leading to a slightly slower release rate. The effect of pH is also not great either (see Figure 7 inset) but is in the direction expected; that is, decreasing the pH from 9 to 3 leads to a small increase in the release rate, as a result of some swelling of the shell, associated with the charging of the pyridine moieties at the lower pH value. The fact that this pH effect is small is consistent with the fact that there is probably a significant amount of water in the PVP shells even at the higher pH value, because of its more hydrophilic nature compared to the other polymers studied in this work. Copolymer Microcapsules of PVP and PS. As may be seen by comparing Figures 1 and 7, the release rate for 4-nitroanisole is significantly faster from PVP-based microcapsules than PS-based ones. (The release yields

Oil Core/Polymer Shell Microcapsules

Langmuir, Vol. 21, No. 12, 2005 5283

Figure 9. Release profiles for 4-nitroanisole from pure PVPKbased microcapsules and PVPK/PVP blends: (×) pure PVPK; (b, O) 0.5 g of PVP; (9, 0) 1 g of PVP. Total polymer used in each case ) 3 g. Solid symbols, pH 3; open symbols, pH 9.

Figure 8. Influence of the amount of styrene in the shells for poly(vinylpyridine-co-styrene)-based microcapsules on the release of 4-nitroanisole: (O) 0%, (9) 10%, and (2) 80%. (A) pH 3; (B) pH 9.

are slightly lower for the PS-based ones as well: 80% for PS, compared to ∼100% for PVP). To see if incorporating some styrene into a PVP-based microcapsule would reduce the release rate, some poly(vinylpyridine-co-styrene) microcapsules were prepared. The results for the initial release rates are shown in Figure 8, for two pH values, 3 and 9. As may be seen, at both pH values, the introduction of just 10 wt % styrene is sufficient to reduce the release rate significantly. In fact, increasing the styrene content to 80 wt % does not reduce the release rate by much more, if at all. It is known from earlier studies of Loxley and Vincent,33 from their stopped-flow kinetic studies of the swelling rates of PVP-co-PS microgel particles, that the introduction of just 10 mol % styrene reduces the swelling rate considerably compared to the pure PVP microgel particles, on lowering the pH from 8 to 3. Styrene is a much more hydrophobic monomer than vinylpyridine, and this must have an important effect on the packing of the polymer chains in an aqueous environment and the extent of water penetration into the polymer shells of the microcapsules. Polymer Blend Microcapsules of PVPK and PVP. For all the polymers studied, microcapsules with shells formed from PVPK produce the slowest releasing capsules and give the lowest yields for 4-nitroanisole. This is due to the strong interaction of this molecule with PVPK, and hence its high retention in the polymer shell. On the other hand, microcapsules formed with PVP have been shown to have very much faster release kinetics. As was discussed earlier, this is most likely a consequence of the hydrophilic nature of the PVP group inducing swelling in the polymer shell by water penetration. In this section the use of blends of PVP and PVPK to form microcapsule shells is investigated. Microcapsules were prepared containing 0, 16, and 33 wt % PVP (with

a total shell mass of 3 g). The results for the release profiles are shown in Figure 9, for two pH values: 3 and 9 (except for the pure PVPK microcapsules, where no pH dependence was expected or observed). Figure 9 shows that increasing the amount of PVP in the PVPK/PVP blend does indeed increase the release rate and yield considerably. There is also the (much smaller) effect seen before, that decreasing the pH from 9 to 3 also increases the release rate somewhat. The use of blends of this sort should, therefore, prove to be a very practical and straightforward way of controlling release rates and yields. Conclusions We have studied the release profiles (rates and final yields) of a number of model release “active” molecules, in particular 4-nitroanisole, from oil core/polymer shell microcapsules into water. A variety of polymers have been used to form the shell. A number of the system variables have been held constant. These include the nature of the oil used (n-hexadecane), the mean particle size (∼2 µm) and concentration, and in most cases the total mass of polymer used (and hence the shell thickness). The main factors that control the final yield seem to be the strength of the polymer/release molecule interaction (for example, the interaction is strong for 4-nitroanisole/ PVPK but very weak for 4-nitroanisole/PVP) and the thickness of the polymer shell. These factors also control the release rate, but other factors come into consideration, in particular, the free volume and porosity of the polymer shell. Increasing the polymer molecular mass seems to decrease these (albeit slightly), and reduces the release rate to some extent. Other factors, which decrease the release rate and which can be introduced after the microcapsules have been formed, are introducing crosslinks into the shell polymer and heating the polymer in the shell above its Tg value. A number of studies have also been made with poly(vinylpyridine) (PVP) as the shell polymer. This is a much more hydrophilic polymer than the other polymers used. The release rate of 4-nitroanisole is much faster for this polymer. Also, the yield is ∼100%, reflecting the very weak interaction of 4-nitroanisole with PVP. It would seem that there is some penetration of water into the PVP shell. Ways of decreasing and controlling the release rate for PVP shells include use of copolymers of PVP with a more hydrophobic monomer (e.g., styrene). However, perhaps the most straightforward way is to use a polymer blend, such as PVP and PVPK.

5284

Langmuir, Vol. 21, No. 12, 2005

Acknowledgment. We thank the former Astra Zeneca Company (now Syngenta) for initial financial support, in particular for P.B. and P.J.D. We also thank Procter & Gamble and EPSRC (GR/R 90086/1) for subsequent financial support, through the IMPACT Faraday Part-

Dowding et al.

nership, for R.A. Dr. David Rodham (now at IMPACT) and Dr. Ian Shirley (now at Syngenta) are also thanked for useful discussions during the early part of this work. LA0470838