pubs.acs.org/Langmuir © 2009 American Chemical Society
Silica-Shell/Oil-Core Microcapsules with Controlled Shell Thickness and Their Breakage Stress Michael O’Sullivan,†,§ Zhibing Zhang,‡ and Brian Vincent*,† †
‡
School of Chemistry, University of Bristol, Bristol BS8 1TS, United Kingdom, and Department of Chemical Engineering, University of Birmingham, Birmingham B15 2TT, United Kingdom. § Current address: BP Institute, University of Cambridge, Cambridge CB3 0EZ, United Kingdom Received February 19, 2009. Revised Manuscript Received March 25, 2009
The encapsulation of one material by another, to form core-shell particles (microcapsules), has many applications, principally the containment, protection, and distribution of an active material. This work describes the development of core-shell particles with silicone oil cores and solid silica-like shells of controlled thickness. Oligomeric polydimethylsiloxane (PDMS) emulsions are employed as the core templates for the formation of the solid shells. The core templates are prepared by the surfactant-free, condensation polymerization of diethoxydimethylsilane (DEODMS) that leads to the formation of monodisperse silicone oil/water emulsions. Solid silica-like, composite shells were formed through condensation of tetraethoxysilane (TEOS) and DEODMS onto the core templates. The shell thickness may be controlled by manipulation of relative TEOS and DEODMS concentrations or by quenching the shell development step. It is possible to incorporate a dye into the core prior to shell formation, which does not seem to permeate the shell. The coated PDMS particles were subjected to a controlled compression stress using a micromanipulation technique. The capsule breaking force was found to be proportional to the shell thickness, as quantified using scanning electron microscopy (SEM) ultramicrotomy.
Introduction The formation of liquid core-polymer shell particles as microcapsules for the protection and release of active species has been reviewed by Yow and Routh.1 Polymer shells are particularly useful in the context of sustained release of the active species over time. Other applications require triggered release, where the active is suddenly released into the surroundings. For this purpose, solid shells, which offer no release prior to breakage but then break under an applied stress, are required. Although solid polymers could be used in this context, leakage is often a problem, particularly with lower molecular weight active molecules. Inorganic solids (such as silica or calcium carbonate), in principle, offer a better barrier to such leakage. The liquid core may be aqueous or nonaqueous, depending on the nature of the active species. With regard to aqueous cores with polymer shells, Atkin et al.2 have reviewed methods of preparing such systems and described a novel internal phase-separation procedure for preparing them. The preparation of microcapsules with silica shells and aqueous cores has been described by O’Sullivan.3 In this paper, however, we concentrate on microcapsules with oil cores, in particular silicone oil cores, and silicabased shells. Goller and Vincent4 and Zoldesi and co-workers5,6 have previously described how silicone oil droplets may be encapsulated with silica-based shells. This paper is an extension of their work. Monodisperse oligomeric polydimethylsiloxane (PDMS or silicone oil) droplets, dispersed in water, may be prepared by a
nucleation and growth mechanism.7,8 This process is analogous to silica dispersion formation.9,10 To form the emulsions, the monomer diethoxydimethylsilane (DEODMS) is shaken with an aqueous or aqueous-ethanol solution of ammonium hydroxide.7 If triethoxymethylsilane (TEOMS) is used in conjunction with DEODMS, as a monomer mixture, a dispersion of cross-linked microgel particles is produced; the nature of the dispersed phase may range from liquid to amorphous solid, depending on the extent of cross-linking.11 Final emulsions or microgel dispersions, in pure water, may be obtained by dialysis. Goller and Vincent4 showed how such silicone oil droplets or microgel particles could be encapsulated with silica shells using a two-step procedure. In the first step, a thin layer of silica is deposited on the droplets by the acid-induced hydrolysis of calcium silicate and precipitation of the resulting silica at the oil/water interface. In the second step, tetraethoxysilane (TEOS) is added to the aqueous phase to build up the silica shell thickness. It was originally thought that if the first step was omitted then the TEOS would simply diffuse into the PDMS droplets, without shell formation. However, subsequent work by Zoldesi and co-workers5,6 showed that this was not the case and that solid shells may be formed around PDMS droplets directly in one step when TEOS is added directly to the aqueous phase. However, for this to occur, there must be some unreacted DEODMS remaining in the oil droplets; that is, the condensation reaction of the DEODMS must not be allowed to go to completion. Indeed, Zoldesi and co-workers5,6 showed that the thickness of the silica-based coating is inversely proportional to the time of formation allowed for the droplets.
*Corresponding author. E-mail:
[email protected]. (1) Yow, Y. N.; Routh, A. F. Soft Matter 2006, 2, 940–949. (2) Atkin, R; Davies, P.; Hardy, J; Vincent, B. Macromolecules 2004, 37, 7979– 7985. (3) O’Sullivan, M. Ph.D. Thesis, 2007. (4) Goller, M. I.; Vincent, B Colloids Surf. A 1998, 142, 281–285. (5) Zoldesi, C. I.; Imhof, A. Adv. Mater. 2005, 17, 924–928. (6) Zoldesi, C. I.; vanWalree, C. A.; Imhof, A. Langmuir 2006, 22, 4343–4352.
7962 DOI: 10.1021/la9006229
(7) Obey, T. M.; Vincent, B. J. Colloid Interface Sci. 1994, 163, 454–463. (8) Neumann, B.; Vincent, B. Langmuir 2004, 20, 4336–4344. :: (9) Stober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62–69. (10) van Blaaderen, A.; van Geest, J.; Vrij, A. J. Colloid Interface Sci. 1992, 154, 481–501. (11) Goller, M.; Obey, T. M.; Teare, D. O. H.; Vincent, B. Colloids Surf., A 1997, 123-124, 183–193.
Published on Web 04/29/2009
Langmuir 2009, 25(14), 7962–7966
O’Sullivan et al.
Article
In this paper, we extend the Zoldesi method by allowing the DEODMS reaction to go to completion but then subsequently adding a mixture of TEOS and DEODMS to the PDMS emulsion droplets. In this way, the shell thickness can be better controlled in an a priori manner. In essence, the Zoldesi method may be described as an in situ method, and the procedure introduced here as a postaddition method, with regard to the location of the DEODMS monomer. We also investigate the relationship between the shell thickness and the compressive stress required to break the capsules. To this end, a direct micromanipulation method was employed that has been developed by Zhang and co-workers.12,13 We compare results for core-shell particles prepared by the two methods, that is, the in situ and postaddition methods.
Experimental Section PDMS Droplets. Diethoxydimethylsilane, DMDES (97%, Aldrich), and all other monomers were purified by passing them down a neutral alumina column prior to use. Emulsions were typically prepared in 2% v/v dilutions of NH3 (0.362 mol dm-3, 35 wt %, Fisher) in deionized water. DMDES (0.12 mol dm-3) was added to NH3 solution in a 30 mL glass vial and, for reproducibility, shaken with a Gallenkamp Spinmix instrument at maximum setting for 1 min. Core-Shell Particles: In Situ (Zoldesi) Method. PDMS emulsions were formed and aged for a predetermined time before the dropwise addition of TEOS (0.018 mol dm-3, 98%, Aldrich) while stirring for 1 h with a magnetic follower. In the current work, using this method, it was found that the core-shell particles would sediment on standing and subsequently coagulate. To prevent this, the core-shell particles were centrifuged three times at 500g RCF in a Sorvall Legend T centrifuge, and the supernatant replaced each time with an aqueous, 0.1 wt % nonionic surfactant solution (INUTEC SP1, ex Orafti Non-Food14). As well as the surfactant providing steric stabilization for the particles, this centrifugation procedure also served to remove any residual oligomeric material and ammonium hydroxide from the continuous phase. Core-Shell Particles: Postaddition (New) Method. PDMS emulsions were formed and aged for 4 days, with occasional agitation to prevent formation of a creamed phase, if necessary. DEODMS (final concentrations ranging from 0.005 to 0.035 mol dm-3) and TEOS (0.018 mol dm-3) were then added dropwise to the emulsions with magnetic stirring, which was continued subsequently for 1 h. The dispersions were then left to stand for a minimum of 3 days for the shell formation reaction to go to completion. At the end of this period, a centrifugation cycle, replacing the supernatant with aqueous surfactant solution, was carried out as described above for the Zoldesi core-shell particles. Particle Size and Shell Thickness Determination. Particle size determination using dynamic light scattering proved to be problematic due to the presence of secondary growth silica particles in the continuous phase. Hence, scanning electron microscopy (SEM) was used; this also enabled the shell thickness to be estimated. Dispersions were cyclically centrifuged into ethanol, with successive supernatant removal and replacement with fresh solvent. After three cycles, the sediment was scraped into microtome vials (Agar Scientific) and dried at 60 °C. A low viscosity (Spurr) resin (Agar Scientific) was added to the dried samples and cured at 60 °C for 12 h.15 The resin blocks were then sliced with a Diatome Ultra 45 diamond knife mounted on an (12) Zhang, Z.; Saunders, R.; Thomas, C. R. J. Microencapsulation 1999, 16, 117–124. (13) Sun, G.; Zhang, Z. Int. J. Pharm. 2002, 242, 307–311. (14) Tadros, Th. F.; Vandamme, A.; Booten, K.; Levecke, B.; Stevens, C. V. Colloids Surf., A 2004, 250, 133–140. (15) Spurr, A. R. J. Ultrastruct. Res. 1969, 26, 31–43.
Langmuir 2009, 25(14), 7962–7966
MT-XL microtome (Ventanna Medical Systems). The sliced surface was retained, sputter-coated with palladium-platinum alloy, and examined with a Jeol JSM 6330F FEG-SEM instrument. Because, as will be shown, the core-shell particles produced in this work are reasonably monodisperse in size, it was straightforward to select “sliced” particles whose total diameter was consistent with the diameter value determined from EM pictures of the unbroken particles. In this way, the thickness values determined were not biased by taking measurements on sliced particles where the slice had not passed reasonably close through the center of the particle. Micromanipulation Technique. Particle breaking stresses were investigated using a micromanipulation technique.12,13 The micromanipulation rig consisted of a fine glass probe that was glued to a force transducer input, which was in turn mounted on a three-dimensional micromanipulator that could be programmed to deploy the probe at a specified speed. The glass probe was formed from a 1.0 mm borosilicate capillary tube (Harvard Apparatus Ltd.) that had been heated and drawn. Half of the drawn tube was mounted on a microforge (Narishige Co., Japan), which consists of a filament arranged in front of a transverse microscope. The drawn capillary was clamped and brought into proximity with the filament, which was heated and raised to meet the capillary tip. The filament was then pulled away from the now molten tube, further drawing it into a finer tip. The probe tip was then subjected to grinding, using a micropipet grinder (Narishige Co.), to ensure an even surface. The probe array was positioned perpendicularly above the stage of an inverted microscope. Samples were deposited on sections of borosilicate glass slide, which were then placed on the stage. The feeds from the microscope-mounted camera and an additional transverse camera were combined in a split view on a monitor, which allows individual particles to be precisely positioned below the glass probe prior to an experimental run. During the experiment, the micromanipulator was lowered under computer control and the output transducer voltage was sampled using a PC data acquisition board. A maximum threshold voltage was set beyond which the array descent would be halted to prevent transducer damage by exceeding tolerance. A drop of the dispersion (1 vol %) to be studied was placed on a microscope slide. Initial experiments were attempted while the droplet persisted, but the particles were so small that their rapid Brownian motion prevented them from being accurately placed beneath the probe. All experiments thereafter were carried out on slides whose samples had been air-dried. The individual slides were placed on the microscope stage and maneuvered with the stage’s translational controls so that discrete, stationary particles could be positioned directly beneath the probe. The probe tip was then manually lowered to a safe distance above the particle, ensuring no encounter between the probe and the particle surface could occur. Probe descent was then initiated and controlled by computer. Descent proceeded until the cutoff voltage was reached. The probe was then withdrawn by computer control. The probe tip was regularly cleaned by using a fine tissue twist, moistened with acetone. At least 20 repeat runs were attempted for each examined sample.
Results and Discussion Core-Shell Particles: In Situ (Zoldesi) Method. A typical SEM picture of particles prepared by this method is shown in Figure 1. In this case, the mean particle diameter is 1.0 ( 0.1 μm (the standard deviation increased the longer the delay time, prior to addition of TEOS, possibly as a consequence of some PDMS droplet coalescence), and the shell thickness is 53 ( 5 nm. As may be seen from Figure 2, the shell thickness decreased as the delay time increased, confirming the earlier results of Zoldesi and co-workers.5,6 DOI: 10.1021/la9006229
7963
Article
O’Sullivan et al.
Figure 4. Shell thickness as a function of DEODMS concentration (postaddition method).
Figure 1. Typical SEM micrograph of core-shell particles prepared by the in situ, Zoldesi method.
Figure 2. Shell thickness as a function of TEOS addition delay time. Error bars represent the standard deviation (in situ, Zoldesi method).
Figure 5. Core-shell particles whose shells were formed in the presence of (a) 0.023 and (b) 0.035 mol dm-3 DEODMS (postaddition method).
Figure 3. Optical micrograph of core-shell particles prepared by the postaddition method and dried onto a glass slide. White scale bar represents 20 μm.
Core-Shell Particles: Postaddition Method. Figure 3 shows a typical optical micrograph of core-shell particles prepared by this method. The regular packing of the particles indicates a narrow size distribution. This suggests that little or no coalescence occurs during the interfacial reaction period. The actual average size of the particles, of course, will depend 7964 DOI: 10.1021/la9006229
primarily on the initial size of the PDMS droplets used, which in this case can be varied over a wide range (up to ∼5 μm7). A series of core-shell particles was created in which the concentration (in the aqueous phase) of added DEODMS was varied from 0.005 to 0.035 mol dm-3 while that of added TEOS was fixed at 0.018 mol dm-3. Figure 4 shows that the shell thickness increases with the amount of DEODMS added, up to ∼0.023 mol dm-3. Thereafter, it decreases. The reason for the decrease may be ascertained from Figure 5, which shows two SEM micrographs for core-shell particles produced (a) with 0.023 mol dm-3 and (b) with 0.035 mol dm-3 DEODMS. In the latter case, the core-shell particles Langmuir 2009, 25(14), 7962–7966
O’Sullivan et al.
Figure 6. Shell thickness as a function of reaction quench time (postaddition method).
produced have a reasonably narrow size distribution, with an average particle diameter of ∼3-4 μm, but the picture also indicates the presence of small silica particles. These are produced by secondary reactions occurring in the aqueous phase. It would seem that this reaction occurs, in addition to the primary reaction at the oil-water interface, at these higher DEODMS concentrations. However, these small silica particles could be readily removed by gentle centrifugation of the larger core-shell particles, removal of the supernatant (containing the much smaller silica particles), and redispersion of the core-shell particles into water or surfactant solution. It is also possible that the growth of these secondary silica particles could be reduced or even eliminated, and thicker shells could be obtained by slow addition of the monomer mixture (i.e., semibatch growth) rather than having all the monomer present initially. This avenue was not explored in this current work, however. By comparing Figures 2 and 4, it may be seen that both the in situ and the postaddition methods produce maximum shell thicknesses of ∼0.12 μm. This similarity is perhaps not surprising, since the concentration of TEOS used in both methods was fixed (0.018 mol dm-3). An alternative method of controlling the shell thickness, using the postaddition method, is to simply stop the reaction at the interface by quenching the system after a given time. This was achieved by carrying out the centrifugation and redispersion cycles into surfactant solution after some chosen time, between 15 and 69 h, rather than allowing the interfacial reaction to go to completion (which takes about 3 days). The results for the coreshell particles produced with 0.035 mol dm-3 DEODMS are shown in Figure 6. As expected, the growth of the shell is fast initially but decreases as the reaction time increases. Compression Studies. Compression studies were undertaken to investigate the compressive breaking force of the core-shell particles. Figure 7 illustrates a typical force-displacement profile obtained using the micromanipulation device. Figure 8 shows a reasonably linear correlation between shell thickness and the force required to break the shells for particles prepared by both methods (in situ and postaddition). However, the thinnest shells seem to give breaking force values somewhat higher than expected, which suggests a more rigid structure. A possible explanation could have to do with the fact that the thinnest shells were formed with less DEODMS; TEOS was the major component. So, one would expect these thinnest shells to be the most “silicalike” in structure. On the other hand, shells with a higher concentration of DEODMS might well have a less rigid structure, similar perhaps to that suggested by Zoldesi et al.,6 shown in Figure 9. Incorporation of a Dye: Shell Permeability. In order to test how impermeable the core-shell particles were to diffusion of a model active material, a dye (Sudan III) was incorporated into the Langmuir 2009, 25(14), 7962–7966
Article
Figure 7. Force-displacement data from a typical particle compression experiment for a particle produced by the postaddition method. The probe meets the particle at point A. The measured applied force increases until the particle breaks at point B. The probe meets no further resistance until C where it begins to compress particle debris. At D, the probe experiences resistance from the substrate. The displacement from point A to D serves as an estimate of particle size.
Figure 8. Breaking force as a function of shell thickness for particles prepared by both the in situ (9) and postaddition (b) methods.
Figure 9. Possible structure for a shell formed from mixtures of TEOS and DEODMS, according to Zoldesi et al.6
PDMS emulsion droplets prior to the growth of the shell. In order to achieve this, Sudan III was dissolved in chloroform to form a 1 wt % solution and then added to a PDMS emulsion, turning it red in color. The droplets tended to sediment on standing, and the appearance of this sediment was more strongly red in color than the supernatant, indicating the incorporation of both the (high density) chloroform and the dye into the droplets. Shells were then formed around the droplets using the postaddition method. The core-shell particles were then centrifuged and redispersed into aqueous surfactant (INUTEC SP1) solution. This procedure was repeated several times. There was no indication of any red color in the final supernatant on standing. This suggests that the shells are continuous in structure with no significant macropores. If the shells were deliberately broken, however, under water, some red color reappeared in the aqueous phase. DOI: 10.1021/la9006229
7965
Article
O’Sullivan et al.
Conclusions Core-shell particles have been prepared consisting of a silicone oil (PDMS) core and a shell based on a copolymer of TEOS and DMDEOS. A new preparation method, capable of greater predictive control of the shell thickness, has been developed. In addition, a study has been made of the breakage strength of the shells; a good correlation with the shell thickness has been found. The shells seem to be impenetrable to dye
7966 DOI: 10.1021/la9006229
molecules, dissolved in the cores, prior to encapsulation, suggesting that the shells are continuous in structure with no significant macropores. Acknowledgment. M.O’S. acknowledges Schlumberger Cambridge Research and the Douglas Everett Postgraduate Scholarship for funding, and Katie Min Lui for assistance with the micromanipulation experiments.
Langmuir 2009, 25(14), 7962–7966