Frozen Cyclohexane-in-Water Emulsion as a Sacrificial Template for the Synthesis of Multilayered Polyelectrolyte Microcapsules

DOI: 10.1021/la901020j ALangmuir XXXX, XXX(XX), XXX–XXX

pubs.acs.org/Langmuir

©XXXX American Chemical Society

Frozen Cyclohexane-in-Water Emulsion as a Sacrificial Template for the

Synthesis of Multilayered Polyelectrolyte Microcapsules

Sachin Khapli,*,† Jin Ryoun Kim,†, ‡ Jin Kim Montclare,†, § Rastislav Levicky,†, ‡

Maurizio Porfiri,†, # and Stavroula Sofou†, ‡

†Center for Co-operative Bioactive Systems, ‡Department of Chemical and Biological Engineering, §Department ofChemical andBiological Sciences and #Department ofMechanical andAerospaceEngineering, Polytechnic Institute of

New York University (NYU-POLY), 6 Metrotech Center, Brooklyn, New York 11201

Received March 23, 2009. Revised Manuscript Received May 20, 2009

This paper reports the application of frozen cyclohexane-in-water emulsions as sacrificial templates for thefabrication of hollow microcapsules through layer-by-layer assembly of polyelectrolytes, poly(styrenesulfonate sodiumsalt), and poly(allylamine hydrochloride). Extraction of the cyclohexane phase from frozen emulsions stabilized with 11polyelectrolyte layers by compatibilization with 30% v/v ethanol leads to the formation of water-filled microcapsuleswhile preserving the spherical geometry. Themajority of microcapsules (>90%) are prepared with intact polyelectrolytemembranes as measured by their deformation induced by osmotic pressure. This work provides a new route for thesynthesis of hollow multilayered microcapsules under mild operating conditions.

Layer-by-layer (LbL) assembly of polyelectrolytes on dissolvablecolloidal templates is a versatile tool for the synthesis of micro-capsules.Themethodwas first reportedbyDonath et al.1,2 and laterfollowed by numerous studies that extended its utility from poly-meric microcapsules to a variety of building blocks such asnanoparticles,2 dendrimers,3 and biomolecules including DNA,4

polysaccharides,5 proteins,6 and lipids.7 Microcapsules have beenpreparedwith control over size ranging from 0.1 to 10 μmaswell ascontrol over wall thickness from 10 to 100 nm.1,2,8,9 With judiciousselection of the building blocks, such as stimuli-responsive biomo-lecules, these microcapsules can be designed to respond to variousphysicochemical stimuli, including pH,7,10-12 temperature,13 andionic strength.13,14 The tailored responsiveness makes these micro-capsules ideal candidates for a variety of biomedical applications

including drug carriers for controlled release,15-17 confined bior-eactors,18 microencapsulation,19-21 and model structures for artifi-cial cells.22,23

Despite these potential applications, there are few templatesavailable for the incorporation of biomolecules as building blockswhile maintaining their intrinsic properties. Most of the colloidaltemplates have limited utility due to challenges related to incom-plete dissolution of the template material24 and harsh chemicaltreatments for template removal.25 As an alternative, an oil-in-water type of emulsion can be considered as a dissolvabletemplate. Emulsions have been conventionally utilized for thesynthesis of microcapsules; however, there are few reports of LbLassembly of polyelectrolytes on these templates. This is mainlydue to the lack of effective techniques for the removal of excesspolyelectrolytes during LbL assembly with the stability of theemulsion preserved.

In particular, among the few available literature reports,McClements et al.27 have shown that the stability of emulsionsused in the food industry against freeze and thaw procedures canbe improved by the addition of multilayers of proteins andpolysaccharides across the oil-water interface. Their experimen-tal protocols are based on the addition of just enough polyelec-trolyte during each layer deposition step and can lead to theflocculationand agglomeration of a fractionof emulsion droplets.Similar protocols have been adopted by M€ohwald et al. for

*Corresponding author (e-mail: [email protected]).(1) Donath, E.; Sukhorukov, G. B.; Caruso, F.; Davis, S. A.; M€ohwald, H.

Angew. Chem., Int. Ed. 1998, 37(16), 2202–2205.(2) Caruso, F.; Caruso, R.; M€ohwald, H. Science 1998, 282, 1111–1114.(3) Kim, B.-S.; Lebedeva, O. V.; Koynov, K.; Gong, H.; Caminade, A.-M.;

Majoral, J.-P.; Vinogradova, O. I. Macromolecules 2006, 39, 5479–5483.(4) Wang, Z.; Qian, L.; Wang, Z.; Yang, F.; Yang, X. Colloids Surf. A:

Physiochem. Eng. Aspects 2008, 326, 29–36.(5) Yu, L.; Gao, Y.; Yue, X.; Liu, S.; Dai, Z. Langmuir 2008, 24, 13723–13729.(6) An, Z.; Tao, C.; Lu,G.;M€ohwald, H.; Zheng, S.; Cui, Y.; Li, J.Chem.Mater.

2005, 17, 2514–2519.(7) An, Z.; M€ohwald, H.; Li, J. Biomacromolecules 2006, 7, 580–585.(8) Shenoy, D. B.; Antipov, A. A.; Sukhorukov, G. B.; M€ohwald, H. Bioma-

cromolecules 2003, 4, 265–272.(9) Schuler, C.; Caruso, F. Biomacromolecules 2001, 2, 921–926.(10) Mauser, T.; Dejugnat, C.; M€ohwald, H.; Sukhorukov, G. B. Langmuir

2006, 22, 5888–5893.(11) Shutava, T.; Prouty,M.; Kommireddy, D.; Lvov, Y.Macromolecules 2005,

38, 2850–2858.(12) Tong, W.; Gao, C.; M€ohwald, H. Macromolecules 2006, 39, 335–340.(13) Gao, C.; Leporatti, S.; Moya, S.; Donath, E.; M€ohwald, H.Chem.;Eur. J.

2003, 9(4), 915–920.(14) Lebedeva, O. V.; Kim, B.-S.; Vasilev, K.; Vinogradova, O. I. J. Colloid

Interface Sci. 2005, 284, 455–462.(15) Liu, X.; Gao, C.; Shen, J.; M€ohwald, H. Macromol. Biosci. 2005, 5, 1209–

1219.(16) Borodina, T.; Markvicheva, E.; Kunizhev, S.; M€ohwald, H.; Sukhorukov,

G. B.; Kreft, O. Macromol. Rapid Commun. 2007, 28, 1894–1899.(17) Zhao, Q.; Zhang, S.; Tong, W.; Gao, C.; Shen, J. Eur. Polym. J. 2006, 42,

3341–3351.(18) Kreft, O.; Prevot,M.;M€ohwald, H.; Sukhorukov, G. B.Angew. Chem., Int.

Ed. 2007, 46, 5605–5608.

(19) Caruso, F.; Yang, W.; Trau, D.; Renneberg, R. Langmuir 2000, 16, 8932–8936.

(20) Caruso, F.; Trau, D.; M€ohwald, H.; Renneberg, R. Langmuir 2000, 16,1485–1488.

(21) Shchukin, D. G.; Patel, A. A.; Sukhorukov, G. B.; Lvov, Y. M. J. Am.Chem. Soc. 2004, 126, 3374–3375.

(22) Tiourina, O. P.; Radtchenko, I.; Sukhorukov, G. B.; M€ohwald, H. J.Membr. Biol. 2002, 190, 9–16.

(23) Katagiri, K.; Caruso, F. Adv. Mater. 2005, 17(6), 738–743.(24) Busse, K.; Kressler, J.; Knorr, J.; Bornemann, S.; Arnold,M.; Thomann, R.

Macromol. Mater. Eng. 2001, 286(6), 355–361.(25) Kreft, O.; Georgieva, R.; Baumler, H.; Steup, M.; Muller-Rober, B.;

Sukhorukov, G. B.; M€ohwald, H. Macromol. Rapid Commun. 2006, 27, 435–440.(26) Sukhorukov, G. B.; Volodkin, D. V.; Gunther, A. M.; Petrov, A. I.;

Shenoy, D. B.; M€ohwald, H. J. Mater. Chem. 2004, 14, 2073–2081.(27) Guzey, D.; McClements, D. J. Adv. Colloid Interface Sci. 2006, 128-130,

227–248.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j

B DOI: 10.1021/la901020j Langmuir XXXX, XXX(XX), XXX–XXX

Article Khapli et al.

encapsulation of toluene droplets with polyglutamate/polyelec-trolyte layers.28 Recently, M€ohwald et al. have used the LbLmethod to coat dodecane droplets with poly(styrenesulfonatesodium salt) (PSS)/poly(diallyldimethylammonium chloride)(PDADMAC) polyelectrolyte layers using creaming-based se-paration of excess polyelectrolytes.29 They have performed well-defined multilayer assembly on oil droplets; however, removal ofthe oil cores to create hollow microcapsules has not beenattempted. So far, there has been only one report of using anemulsion-based sacrificial template for the fabrication of micro-capsules by the LbLmethod: Tjipto et al. have used liquid crystalemulsion of a nematic liquid crystal, 5CB (40-pentyl-4-cyanobi-phenyl), and removed the liquid crystal core by ethanol extractionto yield hollow microcapsules.30

In this paper, we present a novel template in the form of frozencyclohexane droplets dispersed in water. LbL assembly of PSS/PAH (poly(allylamine hydrochloride)) multilayers is performedat the interface of cyclohexane-in-water emulsions below thefreezing point of cyclohexane. Cyclohexane is chosen as the oilphase because of its low boiling point (81.0 �C) and unusuallyhigh melting point31 (6.5 �C). Thus, oil droplets can be frozenbelow 6.5 �C, increasing the stability of the emulsion againstdroplet coalescence. Freezing also reduces the loss of volatilecyclohexane phase by suppressing evaporation. After the oildroplets have been frozen, this template can be handled just likeany other solid templates for LbL assembly, enabling utilizationof microfiltration and centrifugation for removal of excess poly-electrolyte.Cyclohexane, being a volatile compound, can be easilyremoved by subsequent evaporation, lyophilization, or ethanol-assisted compatibilization. In this approach, the cyclohexane isremoved through molecular diffusion across the polyelectrolytemultilayer membrane, enabling nearly complete removal of thetemplate. In addition, this procedure does not require any harshchemical treatments.

Experimental Section

Materials.Cyclohexane (>99%), hexadecane (>98%), etha-nol (95%), poly(4-styrenesulfonate sodium salt) (PSS, MW75 kDa, 30 wt % solution), poly(allylamine hydrochloride)(PAH, MW 56 kDa), and fluorescein isothiocynate (FITC) werepurchased from Sigma-Aldrich Co. FITC-labeled PAH wassynthesized following a reported procedure.19 All chemicals wereused as purchased without further purification.Milli-Q deionizedwater (specific electric conductivity=18.2 MΩ 3 cm) was used toprepare all aqueous solutions.

Emulsification. Cyclohexane-in-water emulsions were prepar-ed by sonicating cyclohexane (2 mL) with PSS solution (10 mL,containing 0.1 M NaCl; [PSS] = 0.12 mg mL-1) using a tipsonicator (Branson Sonifier 150, operating at a power level of10 W for 6� 30 s cycles with sufficient cooling between). Emul-sions were immediately diluted by the addition of 0.1 M NaClsolution (40 mL) at room temperature. Hexadecane-in-wateremulsions were also prepared following the same protocol asabove.

Multilayer Coating of Emulsions. Emulsions, as preparedabove, were coated with extra polyelectrolyte layers to impartadditional stability prior to the freezing step. First, a PSS layerwas added with the minimum possible concentration of theemulsifier (0.12mgmL-1) that gave emulsionswithout significant

creaming tendency. This concentration was derived by system-atically increasing the emulsifier concentration, [PSS], in a seriesof experiments and visually observing the resulting emulsions forcreaming stability. Next, the PAH layer was deposited by theaddition of PAH solution immediately after dilution of theemulsion with 0.1 M NaCl solution (40 mL). The concentrationof PAH during this step was maintained at 0.5 mg mL-1. Thisemulsionwas allowed to creamovernight, 48mLof the subnatantcontaining excess of PAH solution aswell as PAH/PSS complexeswas removed, and the emulsion was replenished with an equalvolume of 0.1 M NaCl solution filtered through 0.22 μm filter.After redispersion, the emulsion was further divided into fivealiquots of 10 mL each and allowed to stand for 15 min. Thecreamy layer that developed on top during this procedure wasdiscarded. Next, the PSS layer was formed by the addition of PSSsolution to make the concentration of PSS=0.75 mg mL-1.

Freezing of Emulsions.Freezingof emulsionswas carried outin a 4 �C refrigerator. Ten milliliters of emulsion cooled to 10 �Cwas added into 0.1MNaCl (40 mL) solution at 1 �C and allowedto stir for 5 min in an ice bath. Stirring was then stopped, and theemulsion was kept in the ice bath for another 15 min. Nomacroscopic aggregates were observed during this cooling stepwith the multilayer coated emulsion.

Layer by Layer Assembly. Ten milliliters of PSS-coatedfrozen emulsion was centrifuged at 50g for 4 min at T=0 �C(Beckman Coulter, Avanti J-E centrifuge). The subnatant (9 mL)was discarded, and the floating turbid layer was redispersed in anequal volume of ice-cooled 0.1MNaCl solution atT=1 �C. Thisprocedure was performed twice, followed by the addition ofPAH solution to make the final PAH concentration, [PAH] =0.5mgmL-1. The frozenemulsionwas then transferred intoan icebath kept in a 4 �C refrigerator. For each layer deposition step, theadsorption time was 20 min, and the emulsion was stirred aftereach 5min time interval to counter gravity-induced concentrationgradients. The same polyelectrolyte concentration (0.5 mgmL-1)was used for both PSS and PAH deposition steps. All solu-tions were stored in an ice bath for the entire duration of theexperiment.

Lyophilization of Frozen Cyclohexane in Water Emul-

sion. Freeze-drying of emulsions was carried out using a Lab-conco lyophilizer (model 775320). Samples were prepared bypouring small droplets of the emulsion (droplet volume=1 μL)into a glass vial filled with liquid nitrogen and storing the frozenglobules at-20 �C for 12 h.The frozen solidwas then freeze-driedfor 12 h at a base pressure of 0.021 mbar. The condensertemperature was maintained at -54 �C throughout the freeze-drying procedure. Lyophilized microcapsules were redispersedinto 0.1 M NaCl solution by sonication.

Compatibilization of Frozen Cyclohexane in Water

Emulsion. After deposition of 11 polyelectrolyte layers, 10 mLof frozen emulsion was centrifuged at 50g for 4 min and 9 mL ofsubnatant was replaced with an equal volume ofMillipore water.This procedure was repeated twice. After the second centrifuga-tion cycle, frozen droplets were redispersed in 30% v/v ethanolsolution in Millipore water and stored in an ice bath for 15 min.Microcapsules were obtained after room temperature compatibi-lization for 12 h and were further purified by three washings withMillipore water.

Fluorescence Microscopy. Images were acquired on anOlympus IX70 inverted fluorescence microscope equipped with10�, 20�, 40�, and 70� objectives and optical filters for FITC/RITC. Images were captured with a CCD camera and analyzedwith MetaMorph software (version 7.5).

ConfocalMicroscopy.Confocal imageswere acquiredwith aLeica TCS SP2 Confocal system using a 100� oil-immersionobjective in the fluorescencemode. Leica confocal software (LeicaLite version 2.61) was used for analyzing the images.

Zeta Potential Measurements. Zeta potentials of frozenemulsions were measured at 3 �C by a Zetasizer Nano ZS 90

(28) Teng, X.; Shchukin, D. G.; M€ohwald, H. Adv. Funct. Mater. 2007, 17,1273–1278.(29) Grigoriev, D. O.; Bukreeva, T.; M€ohwald, H.; Shchukin, D. G. Langmuir

2008, 24, 999–1004.(30) Tjipto, E.; Cadwell, K. D.; Quinn, J. F.; Johnston, A. P. R.; Abbott, N. L.;

Caruso, F. Nano Lett. 2006, 6, 2243–2248.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j

DOI: 10.1021/la901020j CLangmuir XXXX, XXX(XX), XXX–XXX

Khapli et al. Article

instrument (Malvern Instruments Ltd., U.K.). Each measure-ment was repeated three times and for five different samples.Reported values represent the average of these measurements.Corresponding error bars represent the average peakwidths of thezeta potential distributions.

Deformation of Microcapsules Induced by Osmotic Pres-

sure. Microcapsules were incubated with PSS solution contain-ing 1.6 wt % PSS for 10 min to induce deformations in thepolyelectrolyte shells. Confocal fluorescencemicroscopywas thenperformed to visualize the deformations. Osmotic pressure ofthe PSS solution was measured with an osmometer, based onthe technique of freezing point depression (Advanced Instrume-nts Inc., model 3250). Osmolarity reading of the osmometerwas converted into osmotic pressure using van’t Hoff equation,π=RTC, whereC is the osmolarity of the solution (expressed inmol/m-3) and R is the universal gas constant.

Results and Discussion

We performed LbL assembly of PSS/PAH on frozen cyclo-hexane droplets in water at 4 �C with centrifugation after eachlayer deposition step to remove nonadsorbed polyelectrolyte. Upto 11 layers of PSS/PAH were deposited, and ethanol-assistedcompatibilization was used for removal of the cyclohexane phaseto obtain hollow microcapsules. In what follows, we report adetailed discussion of each step.Emulsification. Cyclohexane was dispersed at a volume

fraction of 20% in 0.1 M NaCl solution containing PSS as anemulsifier. Emulsions, as prepared above, were polydisperse withdroplet sizes ranging from 500 nm to 3 μm (from opticalmicroscopy). The emulsions very fairly stable against creaming,as evidenced by lack of a visible creamy layer even after about 4 hof standing at room temperature. However, when emulsions wereallowed to stand overnight, it was observed that creaming led tothe formation of a small fraction of droplets with significantlylarger diameters (>10 μm). This observation suggested that evenwith a negatively charged polyelectrolyte layer (zeta potential =-80 mV at 3 �C and pH 6.4 in 0.1 M NaCl solution), dropletagglomeration was possible when the droplets were forced intoclose proximity within the creamy layer. Droplet coalescence wassubsequently avoided by incubating the emulsions at roomtemperature on a rotisserie mixer overnight, which effectivelysuppressed the formation of a creamy layer. In contrast, whenhexadecane was dispersed in water under identical conditions,the resulting emulsions were found to be stable against creamingfor at least 96 h. Because the aqueous solubility of hexadecane32

(<36 ppb) is considerably smaller compared to that of cyclohex-ane32 (55 ppm), this experimental finding suggested that Ostwaldripening33-36 might be responsible for the lowered stability of thecyclohexane-in-water emulsion at room temperature.Preparation of Emulsions for Freezing. Freezing of emul-

sions with a single polyelectrolyte layer of PSS caused a smallfraction of the droplets to aggregate. This aggregation waspossibly induced by a loss of the polyelectrolyte layer triggeredby an abrupt change in the solvation of the PSS phenyl rings incyclohexane when the solvent was frozen. Hence, for enhancedstability, two additional layers (PAH/PSS) were deposited at the

cyclohexane-water interface before freezing according to themethod of McClements et al.37,38

Figure 1 shows the schematic of our method involving twophases of LbL assembly for fabrication of microcapsules. Duringphase 1, multilayer assembly was performed on liquid cyclohex-ane droplets at room temperature following the methods ofMcClements et al.38 In this approach, just enough polyelectrolyteis added to saturate the droplet surfaces. However, as thismethodis prone to droplet aggregation and flocculation, it was only usedto obtain secondary and tertiary emulsions coatedwith PSS/PAHand PSS/PAH/PSS multilayers, respectively. Tertiary emulsionswere then frozen (phase 2) and modified with standard LbLassembly, including separation of excess polyelectrolyte followingeach layer deposition step by centrifugation. Typically, oil-in-water emulsions are not stable against freezing, because thedominant crystallization mechanisms for dispersed oil dropletsin the absence of impurities are surfactant-induced nucleation39

and homogeneous nucleation.40 This leads to the presence ofsupercooled droplets41 and, consequently, the destabilization ofemulsion by interdroplet heterogeneous nucleation.42 However,freezing of the cyclohexane-in-water emulsions did not lead to thisphenomenon, even with a single PSS layer at the cyclohexane-water interface. We hypothesize that this enhanced stability maybe attributed to the inherent molecular order of liquid cyclohex-ane as evident from its unusually highmelting point.31Due to thishigh molecular order, homogeneous nucleation might be favor-able in cyclohexane-in-water emulsions.LbL Assembly of Polyelectrolytes on Frozen Cyclohex-

ane Cores. LbL assembly of PSS/PAH layers was performed byalternating the concentrations of polyelectrolytes in the frozenemulsion. Two centrifugation cycles were performed after eachlayer deposition step to decrease the residual concentration of thenonadsorbed polyelectrolyte by 99%. Because of the lowerdensity of cyclohexane (0.79 g mL-1) relative to that of theaqueous phase, frozen droplets floated on the surface. Thedroplets could be easily redispersed by gentle shaking, provided

Figure 1. Schematic of layer-by-layer assembly procedure. Phase1 refers to the encapsulation of liquid cyclohexane droplets at roomtemperature. Phase 2 refers to the deposition of polyelectrolytemultilayers in frozen emulsion. Phase 3 refers to the removal of thecyclohexane phase to obtain microcapsules.

(31) This is typical of molecules with a high degree of symmetry, e.g, benzene,wherein the entropy of the liquid is lowered by an amountRT ln(Ω), whereΩ is thesymmetry number of the molecule.(32) CRCHandbook of Chemistry and Physics, 44th ed.; Chemical Rubber Co.:

Cleveland, OH, 1963.(33) Taylor, P. Adv. Colloid Interface Sci. 2003, 106, 261–285.(34) Wooster, T. J.; Golding, M.; Sanguansri, P. Langmuir 2008, 24, 12758–

12765.(35) Mun, S.; McClements, D. J. Langmuir 2006, 22, 1551–1554.(36) Sakai, T.; Kamogawa, K.; Nishiyama, K.; Sakai, H.; Abe, M. Langmuir

2002, 18, 1985–1990.

(37) McClements, D. J. Langmuir 2005, 21, 9777–9785.(38) Guzey, D.; McClements, D. J. Adv. Colloid Interface Sci. 2006, 128-130,

227–248.(39) McClements, D. J.; Dungan, S. R.; German, J., B.; Simoneau, C.; Kinsella,

J., E. J. Food. Sci. 1993, 58, 1148–1151.(40) Skoda, W.; Van den Tempel, M. J. Colloid Sci. 1963, 18, 568–584.(41) Cramp, G. L.; Docking, A. M.; Ghosh, S.; Coupland, J. N. Food Hydro-

colloids 2004, 18, 899–905.(42) McClements, D. J.; Dungan, S. R. J. Colloid Interface Sci. 1997, 186, 17–28.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j

D DOI: 10.1021/la901020j Langmuir XXXX, XXX(XX), XXX–XXX

Article Khapli et al.

that the centrifugation was performed at low revolutions perminute (corresponding to 50g) and for short time intervals(4 min). The removal of excess polyelectrolyte was followed bythe addition of the oppositely charged species.

The volume fraction of frozen droplets was kept sufficientlylow (approximately 1%) to avoid droplet aggregation during thelayer deposition steps.The polyelectrolyte concentrationwas keptconstant at 0.5 mg mL-1 to ensure complete coverage of thedroplet surface while at the same time avoiding depletion-inducedflocculation, as predicted byMcClements.37 Zeta potentials weremeasured to confirm surface charge reversal during LbL assem-bly. Measurements were performed in 0.1 M NaCl solution at3 �C (corresponding to Debye length=0.97 nm and pH=6.4).These results are plotted in Figure 2 for the addition of eightconsecutive layers of PSS/PAH deposition. The repeated reversalof surface charge confirms the successful LbL assembly of apolyelectrolyte multilayer. The zeta potential measurements alsoindicate that the emulsion droplets are quite stable againstcoalescence. However, estimation of surface charge from the zetapotential was not possible because the exact location of the planeof shear, which is required for surface charge estimation, isdifficult to determine due to surface roughness of the coateddroplets as well as penetration of ions into the polyelectrolytelayers. A modest decreasing trend in the magnitude of zeta

potential could be due to increasing surface density of defectswithin the multilayer membrane during successive deposition ofpolyelectrolyte layers.

The LbL assembly was further confirmed by confocal fluore-scence microscopy using PAH labeled with FITC (fluoresceinisothiocyanate). Figure 3 shows a fluorescence micrograph of acyclohexane-in-water emulsion coated with 5.5 bilayers of PSS/PAH-FITC. The presence of uniform fluorescence forming abright shell around the droplets confirms the incorporation offluorescently labeled PAH layer into the polyelectrolyte multi-layer assembly at the cyclohexane-water interface. The fluores-cence intensity of the multilayered shells increased monotonicallyafter each deposition step of PAH-FITC conjugate.Removal of Cyclohexane and Formation of Microcap-

sules.After successful LbLassembly on frozen emulsions, variousstrategies were evaluated for removal of the frozen cyclohexanecores: evaporation, lyophilization, and compatibilization withethanol as a cosolvent to obtain water-filled microcapsules.

Evaporation of cyclohexane was performed by allowing thefrozen emulsions to warm to room temperature. Evaporationunder reduced pressure led to the formation of broken andcollapsed shells as shown in Figure 4a, whereas evaporationunder atmospheric pressure led to partially shrunken hollowcapsules as shown in Figure 4b. The broken shells in Figure 4amight be due to rapid evaporation of the cyclohexane phase at theair-water interface and resultant bursting of the shells. On theother hand, the slower evaporation under ambient atmosphericconditions led to buckled microspheres, indicating that theaqueous phase did not spontaneously enter the capsules as thecyclohexane phase evaporated.

Lyophilization of multilayer-coated emulsion was then at-tempted to maintain the spherical geometry of the microspheres(Figure 4c). In this method, the emulsion was first fast-frozen bypouring small droplets into liquid nitrogen, followed by freeze-drying of the frozen droplets under vacuum. This resulted inslightly better removal of the cyclohexane phase and minimalbuckling of the microcapsules (Figure 4c). However, this methodled to the aggregation of microspheres during freezing of theaqueous phase.

Figure 3. Confocal fluorescence micrograph of cyclohexane-in-water emulsion coated with 5.5 bilayers of PSS/PAH-FITC. Thescale bar is 75 μm.

Figure 2. Zeta potential of frozen emulsions measured duringlayer-by-layer assembly.T=3 �C, [NaCl]=0.1M(correspondingto Debye length = 0.97 nm), and pH= 6.4.

Figure 4. (a) Fluorescence micrograph of broken shells after eva-poration at reduced pressure, (b) phase contrast micrograph ofbuckled shells after slow evaporation, and (c) fluorescence micro-graph ofmicrocapsules after lyophilization. The scale bar is 10 μm.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j

DOI: 10.1021/la901020j ELangmuir XXXX, XXX(XX), XXX–XXX

Khapli et al. Article

The most effective method for removal of the cyclohexanecores was realized by using ethanol as a compatibilizer. Ethanolforms a ternary azeotrope with water and cyclohexane with aboiling point of 62.1 �C (7% water, 17% ethanol, and 76%cyclohexane).32 By improving the miscibility of water and cyclo-hexane, we found that ethanol can simultaneously facilitatetransport of cyclohexane from the microcapsules into the waterphase, as well as the entry of water into the microcapsules.Ethanol was added to 30% v/v to the emulsion, and the emulsionwas allowed to warm to room temperature. After standing over-night, microcapsules sank in the aqueous phase, indicatingreplacement of the low-density cyclohexane phase with the higherdensity ethanol-aqueous phase. In a control experiment, thesame treatment but with 30%v/v of a 10:1mixture of ethanol andcyclohexane resulted in emulsion droplets that floated on water.This outcome confirmed the important role of ethanol in remov-ing the cyclohexane phase. In particular, if the aqueous phase wasalready saturated with cyclohexane and ethanol, no compatibili-zation resulted. Figure 5 shows the confocal fluorescence micro-graph of water-filled microcapsules obtained after compatibiliza-tion of a 5.5 PSS/PAH-FITC bilayer emulsion. As seen inFigure 5, the final size of microcapsules was in the 10-15 μmrange, much larger than the initial template size of 0.5-3 μm. Webelieve that this increasemay reflect aggregationof heterogeneouslycharged emulsion droplets during LbL assembly, or Ostwaldripening, or a combination of both phenomena. Homogenizationbefore freezing, to obtain a more monodisperse emulsion, mighthelp overcome these problems.

During ethanol-assisted compatibilization, itwas also observedthat microcapsules with PAH as the outer layer exhibited agreater tendency to aggregate than those with PSS as the outerlayer. Therefore, PSS, a strong polyelectrolyte, was always kept asthe outermost layer by design. It was also observed that emulsionswith 5.5 PSS/PAH bilayers after treatment with 30% v/v ofethanol produced spherical hollow capsules while those with lessthan 5.5 bilayers yielded nonspherical geometries with evidenceof buckling of the weaker shells. This likely reflects greatersusceptibility of thinner shells to build-up of osmotic pressure

as cyclohexane is being replaced by ethanol and water. Hence, aminimum of 5.5 bilayers is recommended before ethanol-assistedcompatibilization.Osmotic Pressure Induced Deformation. The integrity of

polyelectrolyte shells after core removal was assessed by theirability to support osmotic stress. The presence of an osmoticstress led to buckling of the microcapsules as illustrated inFigure 6. In this image, spherical microcapsules were exposed toa 1.6 wt % solution of PSS, which exerts an osmotic pressure of2.03 � 105 N m-2. The rather large macromolecules of PSS(70 kDa) could not permeate through the polyelectrolyte shellsand thus retained an excess of Na+ counterions in the aqueousphase surrounding the microcapsules. In response, exchange ofwater across the polyelectrolyte shells occurred as a means toequalize the chemical potential ofwater between themicrocapsuleinterior and the surroundings. The accompanying water loss ledto invaginations of the intact shells. More than 90% of the shells(as observed by fluorescence microscopy) were deformed,consistent with the majority of microcapsules remaining intactafter removal of the cyclohexane phase. The magnitude ofminimum osmotic stress required to induce buckling ofmicrocapsules was found to be 1.8 � 104 N m-2, an order ofmagnitude lower than the literature-reported values for PSS/PAHLbL microcapsules with 10 polyelectrolyte layers and radius =2-3 μm.43 The critical pressure required for buckling of micro-capsules with similar shell thickness and elastic modulus isinversely proportional to the square of the radius.43 Microcap-sules prepared according to ourmethod are considerably larger insize (radius=5-10μm) and therefore deform readily under lowerosmotic stress.

Summary and Conclusions

This work provides a new route for the synthesis of hollowmicrocapsules by LbL assembly. Cyclohexane-in-water emul-sions with at least 1.5 bilayers of PSS/PAH can be stably frozenand used as sacrificial templates for further LbL assembly in the

Figure 5. Confocal fluorescencemicrograph ofmicrocapsules ob-tained after extraction of cyclohexane from 5.5 PSS/PAH-FITCbilayer coated emulsion by compatibilization with 30% v/v ofethanol. The scale bar is 75 μm.

Figure 6. Osmotic pressure induced deformation of spherical mi-crocapsules upon exposure to a 1.6 wt% solution of PSS (osmoticpressure of 2.03 � 105 N m-2). The scale bar is 30 μm.

(43) Gao, C.; Donath, E.; Moya, S.; Dudnik, V; M€ohwald, H. Eur. Phys. J. E2001, 5, 21–27.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j

F DOI: 10.1021/la901020j Langmuir XXXX, XXX(XX), XXX–XXX

Article Khapli et al.

frozen state. The cyclohexane cores can be subsequently removedby ethanol-assisted compatibilization of cyclohexane and water.At least 5.5 bilayers of PSS/PAH were required to yield robust,water-filled microcapsules. Due to the mild operating conditions,this method is expected to be well-suited for the use of sensitivebiomolecules as building blocks. In the future, combination offrozen-template techniques demonstrated here with improved

control over the initial droplet size and polydispersity, usingestablished techniques, should provide additional control overthe microcapsule geometry.

Acknowledgment. This work was supported by the Polytech-nic Institute of New York University (NYU POLY) through theCooperative Bioactive Systems (CBAS) initiative.

Dow

nloa

ded

by C

AL

TE

CH

on

July

31,

200

9Pu

blis

hed

on J

une

9, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/la

9010

20j