at SciVerse ScienceDirect

Food Hydrocolloids 27 (2012) 119e125

Contents lists available

Food Hydrocolloids

journal homepage: www.elsevier .com/locate/ foodhyd

Microcapsule production by an hybrid colloidosome-layer-by-layer technique

Francisco J. Rossier-Miranda*, Karin Schroën, Remko BoomFood and Bioprocess Engineering Group, Wageningen University, PO Box 8129, 6700 EV Wageningen, The Netherlands

a r t i c l e i n f o

Article history:Received 2 September 2010Accepted 15 August 2011

Keywords:Silica microparticlesPickering emulsionProteinPectinMicrocapsuleScaffold

* Corresponding author. Tel.: þ31 317 485 096; faxE-mail address: francisco.rossier@wur.nl (F.J. Rossi

0268-005X/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.foodhyd.2011.08.007

a b s t r a c t

Although many different methods for microencapsulation are known only some of them had beenapplied at industrial scale, due to complexity, lack of mechanical strength of the resulting capsules, andthe costs related to their production. One of such methods is the electrostatic layer-by-layer (LbL)adsorption, which produce shells from oppositely charged polymers. The thickness of those shells can betuned with nanometric precision, but to build enough strength for practical applications requires theadsorption of an impractical number of layers. We present here a method to produce strong micro-capsules combining the assembly of a protein/pectin shell via electrostatic LbL adsorption with theadsorption of bigger charged colloidal particles. Those colloidal particles do not need any pretreatment tomodify their wettability, as would be the case for a standard colloidosome route. In this way strongencapsulates with porous walls are obtained, which can be used as easy to load scaffolds. The pores inthe walls can be closed through subsequent adsorption of more layers of protein and pectin. Since theassembly scheme occurs at pH 3.5 we expect the produced microcapsules to act as an effective deliverysystem in food products, protecting their contents from the acidity of the stomach and dissolving laterat the small intestine. The proteins and pectins used as basic building blocks are food-grade andinexpensive.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The attention devoted to microencapsulation has increased inrecent years, responding to the important technological questionon how to better protect and deliver active molecules to, forexample, precise locations at the human body. An abundant andalways increasing number of techniques can be found in literature(Gouin, 2004), usually with claims about applicability in variousindustries (Cayre, Noble, & Paunov, 2004; Hsu et al., 2005),including food and pharma. However, many of these techniquescannot be applied in the food industry, since they involve non-foodgrade materials, or materials that are simply too expensive to becommercially interesting for low-added value products like foods.

From the many available methodologies, two routes for micro-encapsulation have attracted a special attention, given the precisecontrol they yield over capsule size and permeability. The first oneis known as assembly by layer-by-layer (LbL) adsorption (Decher,1997; Sukhorukov et al., 1998) and the second one as assemblyfrom Pickering emulsions; whose resulting capsules are calledcolloidosomes (Dinsmore et al., 2002).

: þ31 317 482 237.er-Miranda).

All rights reserved.

LbL microcapsules are assembled from pairs of polymers thatcan interact with each other, usually by an electrostatic interaction.These polymers are adsorbed sequentially in layers on a template,which can be a solid particle or a soft body such as an oil droplet.The size distribution of the resulting microcapsules is determinedby the size and monodispersity of the template used. The shellthickness, along with its permeability (Qiu, Donath, & Möhwald,2001) and mechanical properties (Vinogradova, 2004), is deter-mined by the number of adsorbed layers.

Colloidosomes are assembled from colloidal particles which arelocked at the interface of oil-in-water or water-in-oil emulsiondroplets (Rossier-Miranda, Schroën, & Boom, 2009). Generally, it iseasier to assemble colloidosomes from water-in-oil emulsions(Yow & Routh, 2009), since the particles have to travel a shorterdistance to the interface, reducing the concentration of particlesneeded and producing shells with a better organization. Thetemplate (i.e., the oil droplet) once more determines the size of theresulting colloidosomes; however in contrast to the LbL systemsthe resulting shell has distinctive pores between the particles inthe shell. Those pores define the colloidosome’s permeability(Wang, Duan, & Möhwald, 2005). Further, the shape and nature ofthe colloidal particles co-determine the mechanical properties ofthe shell (Croll, Stöver, & Hitchcock, 2005; Duan et al., 2005; Noble,Cayre, Alargova, Velev, & Paunov, 2004).

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125120

Two aspects are important for colloidosome assembly. The firstone is the wettability of the particles to be used: ideally, to seatefficiently in the interface, they have to be more or less equallywetted by both fluids. This can restrict the kind of particle that canbe included in a colloidosome shell, although some times theparticles can be chemically modified to tune their wettability(Velev, Furusawa, & Nagayama, 1996) or the adsorption can befacilitated with the use of surfactants (Shah, Kim, & Weitz, 2010).The second one is the mechanism used to lock the particlestogether. Particles can be fused together by heating the system nearthe melting or glass temperature of the colloidal particles to inducesintering (Salari, Van Heck, & Klumperman, 2010), by introducinga second phase that glues the particles together (Ao, Li, Zhang, &Ngai, 2011), or by covalent cross-linking from the oil or thewater phase (Thompson et al., 2010; Walsh, Thompson, Armes, &York, 2010).

An issue common to any encapsulation method is to establisha route to load the active component into the microcapsule(Johnston, Cortez, Angelatos, & Caruso, 2006). For colloidosomesand LbL microcapsules the loading can take place through thedroplet that is used as a template: oil droplets may be pre-loadedwith hydrophobic components, or with water droplets holdinghydrophilic substances, or both. Loading can also be completed bydiffusion of the to-be-loaded molecule into the microcapsule fromthe surrounding solution (Kim et al., 2007; Zhu, Srivastava, &McShane, 2005).

Some methodologies have aimed to complement both routes.For example Gu and coworkers assembled colloidosomes wherethe colloidal particles adsorbed on the oil droplet template were, inturn, smaller oil droplets, which could result useful to compart-mentalize different active ingredients in the same microcapsule(Gu, Decker, & McClements, 2007). Silica nanoparticles (25 nm) hasbeen used in the LbL assembly scheme of inorganic shells of0.64 mm (Caruso, Caruso, & Möhwald, 1998), but such smallmicrocapsules are intrinsically more stable and require fewadsorption steps to survive the core removal. In this paper, we takea different approach to make microcapsules, complementing theclassic route to assemble colloidosomes with LbL deposition. Thisallow us to build colloidosome-like shells without the need tochange the wettability of the colloidal particles used, and withoutapplying harsh methods to lock the colloidal particles together. Inour work, the presence of large (non-porous) particles in the LbLsystem drastically reduces the permeability of the shell, andincreases the mechanical stability without the need for numerousadsorption steps, as was noted earlier with the assembly of proteinfibrils as reinforcement material for a similar LbL system (Rossier-Miranda, Schroën, & Boom, 2010; Sagis et al., 2008). Here, wecreated first a scaffold capsule that can be loaded after the assemblyof the shell, whose pores can be closed later through the depositionof more layers. The preparation of porous shells which pores can be

Fig. 1. Scheme for the assembly of a 7-layered LbL-colloidosome. (A) A positively charged temthe rinsing of any unadsorbed excess. (B) Negatively charged pectin and positively charged prmicroparticles is adsorbed to reinforce the protein-pectin shell, creating an open but sturdypores between the silica particles. (E) The template droplet can be removed to obtain a ho

closed after the loadingwith the active component has been proveduseful for the encapsulation of small molecules e.g. albumin(Georgieva et al., 2002) or dextran (Sukhorukov, Antipov, Voigt,Donath, & Möhwald, 2001). To reduce the variability inherent tomaterials from food origin we chose hexadecane as model oil andsilica microparticles as model colloidal particles. Nevertheless, thesame assembly principles are valid to food-grade oils and chargedcolloidal particles from food origin, as for example protein micro-particles (Sa�glam, Venema, de Vries, Sagis, & van der Linden, 2011).

2. Experimental

2.1. Materials

The following materials were used as received: n-hexadecane(Merck, synthesis grade), sodium chloride (Merck, analyticalgrade), hydrochloric acid (Merck, analytical grade), sodiumhydroxide (Merck, analytical grade), formic acid (Aldrich, reagentgrade), whey protein isolate (WPI, Davisco Food International), andhigh methoxyl pectin (HMP, CP Kelco). Two kinds of silica particleswere obtained from Microparticles Berlin GmbH, one modifiedwith amino groups, with a diameter of 0.5 mm (SiO2NH2), andone modified with carboxyl groups, with a diameter of 0.6 mm(SiO2COOH).

2.2. Methods

2.2.1. LbL-colloidosomes preparationThe microcapsules were prepared following the layer-by-layer

electrostatic adsorption method, schematized in Fig. 1. A softtemplate was prepared by emulsifying 1% v/v n-hexadecane ina solution of WPI 0.1% w/w in formate buffer at pH 3.5 and 10 mMNaCl, using an Ultra-Turrax homogenizer (IKA T18 basic with an S18N-19Gdispersion tool) at 24,000 rpmduring 5min. The template oildroplets were separated from the serum by mild centrifugation, at94 RCF during 1 h, and draining. The recovered cream, made fromconcentrated oil droplets covered with WPI, was washed threetimes with buffer to remove any excess of non-adsorbed WPI. Thepositively charged concentrated creamwas then ready for additionof the next layer. The concentration ofmicrocapsuleswas kept equalto 1% v/v in all the following adsorption steps to reduce the chanceof bridging flocculation. Since at the set pH WPI and SiO2NH2 arepositively charged, and HMP and SiO2COOH are negatively charged,it is possible to adsorb them sequentially following the sameprocedure described above. In this way various hybrid LbL-colloidosomes were assembled with 2 layers (WPIeSiO2COOH), 3layers (WPI/HMPeSiO2NH2), 4 layers (WPI/HMPeSiO2NH2eHMP),5 layers (2�(WPI/HMP)eSiO2NH2), and 7 layers (2�(WPI/HMP)eSiO2NH2eHMP/WPI), and the degree of closing and positioning ofcomponents was investigated. The concentrations used in each

plate is prepared by emulsification, obtaining oil droplets stabilized with protein afterotein can be adsorbed sequentially. This step can be repeated at will. (C) A layer of silicascaffold. (D) More layers of pectin and protein can be sequentially adsorbed to close thellow shell.

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125 121

layer were determined from the size distribution and number oftemplates available in each adsorption cycle, in order to avoidbridging flocculation. For example, for a sample of 1% v/v ofmicrocapsuleswith a diameter of 20 mmthe concentrations found tobe optimal were 0.1% w/w forWPI and HMP for the layers adsorbedbefore the adsorption of the silica particles. The concentration ofsilica particles used was 0.3% w/w, close to the theoretical concen-tration needed for the coverage of the surface by random packing.The concentration of WPI and HMP used in subsequent layers was0.3% w/w, accounting for the increase in surface resulting from theadsorption of the silica particles. Once the capsule was complete,the template oil droplet was removed by freeze drying in a ChristEpsilon 2-6D (Osterode, Germany) under the same conditionsdescribed elsewhere (Sawalha, Fan, Schroën, & Boom, 2008).

2.2.2. Characterization of LbL-colloidosomesAfter each adsorption cycle, the size distribution of the micro-

capsules was determined by light scattering, using a MalvernMastersizer 2000. Samples with a total concentration of 1% w/whexadecane were loaded to the sample dispersion unit (Hydro SMsmall volume, Malvern) prepared with 50 ml of 10 mM formatebuffer (pH 3.5), until an obscuration of 20 units was reached. Eachsample was measured 3 times with pauses of 15 s between eachmeasurement. If the sample size distribution remained constantafter the adsorption of a new layer an average size distribution wascalculated from the three measurements. If a change in sizedistribution was detected the sample was checked with lightmicroscopy, to determine if the change was caused by weak clus-tering, which could be corrected during the next adsorption step.Similarly, the z-potential of the sample was also determined aftereach adsorption cycle by light scattering using a Malvern ZetaSizerNano, from samples of the same concentration as indicated above.The average z-potential was calculated from 5 measurements ona sample. For SEM, freeze-dried microcapsules were placed ontobrass holders with double-sided sticky carbon tape (EMS, Wash-ington, U.S.A.). The microcapsules were attached to the stickylayer with pressurized air. The specimen holders with the LbL-colloidosomes were sputter coated with 10 nm Platinum ina dedicated preparation chamber (CT 1500 HF, Oxford Instruments,Oxford UK). The analysis was performed with a field emissionscanning electron microscope (JEOL 6300 F, Tokyo, Japan) at roomtemperature at a working distance of 8 mm, with SE detection at3.5 kV. Images were digitally recorded (Orion 6 PCI, E.L.I. sprl.Belgium) and optimized and resized by Photoshop CS.

2.2.3. Reflectometry experimentsAdsorption experiments were performed in a stagnation flow

cell (a.k.a. reflectometer) as described elsewhere (Dijt, Cohen Stuart,

Fig. 2. LbL-colloidosomes assembled from SiO2COOH particles adsorbed on a single layer oftemplate (scale bar represents 10 mm) (b) SEM picture after removal of the oil template (c)

Hofman, & Fleer, 1990), with 0.1% w/w solutions of WPI or HMP.Silicon oxide stripes, with an oxide layer thickness of 80 nm, wereprepared and hydrophobized with dimethyldichlorosilane asdescribed in thework of Schroën et al. (Schroën, Cohen Stuart, VoortMaarschalk, Padt, & Riet, 1995). Following each adsorption stepthere was a washing step with formate buffer. Both the adsorptionand the washing time were fixed at 15 min each, within which theadsorbed amount reached its maximum value. The refractive indexincrements used were 0.15 for the HMP (Corredig & Wicker, 2001)and 0.18 for the WPI (Dijk & Smit, 2000).

3. Results and discussion

Hybrid LbL-colloidosomes were assembled from SiO2 particlesadsorbed on oil droplets that were previously covered with WPIand HMP layers or combinations thereof. Fig. 2a shows a typicalmicrograph of LbL-colloidosomes assembled from SiO2 micropar-ticles adsorbed on a single layer of WPI. These LbL-colloidosomesare stable in solution, provided that the particle concentration isabove the saturation concentration, and below the depletionconcentration (McClements, 2005). However, they are not stableenough to survive the template removal; Fig. 2b shows a colloido-some that has collapsed after freeze drying. Nevertheless, the close-up in Fig. 2c shows that a single layer ofWPI seems to ‘glue’ some ofthe SiO2 microparticles together.

The amount of protein adsorbed at the oil/water interfacegenerally corresponds approximately to monolayer coverage(Tcholakova, Denkov, Sidzhakova, Ivanov, & Campbell, 2003), afterequilibrium has been established. Protein that is not adsorbedproperly is most likely removed during the washing steps(Tcholakova, Denkov, Ivanov, & Campbell, 2002). The amount ofmaterial that can be adsorbed in a single adsorption cycle istherefore limited, and not sufficient to stabilize the colloidosomeduring freeze drying. We thus decided to use the layer-by-layertechnique to increase the amount of material in the interface.

Fig. 3 shows a typical charge inversion plot found for theinvestigated LbL systems. When the charge inversion is complete,we observe rather constant z-potentials around þ56 mV fora SiO2NH2 layer, and �14 mV for an HMP layer. It is interesting tonote that the z-potential of a layer ofWPI adsorbed on an oil droplet(þ39 mV) differs significantly from that of a WPI layer adsorbed onan HMP layer (around þ7 mV), and from the z-potential of WPIdissolved in the formate buffer (around þ17 mV). At pH 3.5 all themain proteins ofWPI (b-lactoglobulin, a-lactalbumin, and albumin)are well under their isoelectric points, and therefore positivelycharged. When in solution, WPI proteins are in the globular formand expose only part of their charged groups; however, whenadsorbed on the oil interface they are expected to reorganize and

WPI, (a) Micrograph of microcapsules as prepared in liquid, without removal of the oilDetail of silica particles glued together by a layer of WPI.

Fig. 3. z-potential of adsorbed layers. SiO2COOH on a single layer of WPI (>), SiO2NH2

on a WPI/HMP sequence and covered with a single layer of HMP (,), and SiO2NH2 ona double WPI/HMP sequence and covered with an HMP/WPI sequence (D).

Fig. 4. LbL-Colloidosomes. (a) SiO2NH2 microparticles adsorbed on a single (WPI/HMP)sequence. (b) SiO2NH2 microparticles adsorbed on a double (WPI/HMP) sequence.

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125122

expose their most hydrophobic portions to the oil, and the mosthydrophilic (and therefore charged ones) to the continuous phase(Basheva, Gurkov, Christov, & Campbell, 2006), therewith causingthis difference in z-potential.

When adsorbed on an HMP layer, part of the charge of WPI iscompensated by the electrostatic interaction with HMP. Althoughthe remaining charge is not as high as that of a WPI layer on theinterface of an oil droplet, it is sufficient to adsorb the next layer ofHMP without neutralizing the charge of the LbL-colloidosome. Wealso found that adsorption of a layer of SiO2COOH on a layer of WPIis not enough to cause charge inversion, probably due to a lowsurface coverage. This should however not be a problem foradsorption of subsequent layers, since charge inversion is nota prerequisite for assembly of LbL systems (Neff, Wunderlich,Klitzing, & Bausch, 2007).

Fig. 4 shows typical examples of LbL-colloidosomes preparedfrom oil droplets covered by one (a) or two (b)WPI/HMP sequentiallayers. In both cases, the silica particles are sufficiently lockedtogether to maintain the spherical shape after template removal,but the coverage is much higher with a double WPI/HMP layer. It isexpected that the low surface coverage of the single WPI/HMPlayered system was caused by the template removal by freezedrying. Since the oil has to be removed as a gas at low pressure,much volume has to escape from the capsule; this may havecreated buildup of internal pressure, which ruptured part of thewall. The system with two WPI/HMP layers is more robust andwithstood the pressure buildup. Detailed images of the walls ofboth capsules are shown in Fig. 5.

The interstitial phase obtained after a singleWPI/HMP sequencewith silica particles (Fig. 5a) seems considerably stronger than thatby a single WPI layer (Fig. 2c) but many open spaces remainbetween the particles (Fig. 5a). The available WPI/HMP amount isapparently not sufficient to close the capsule. A double WPI/HMPsequence forms a continuous membrane that deforms withoutbreaking upon adsorption of the particles, which remain embeddedin this matrix (Fig. 5b).

Since the deposition of the interstitial (cement) layer is of greatinfluence on the construction of the capsules, we decided toemulate this part of the capsule formation, by adsorbing HMP and

WPI sequentially on a flat, hydrophobized silicon oxide surface; theadsorbed mass of polymers in each step was followed by reflec-tometry. Fig. 6 shows a typical curve for the adsorption of twoWPI/HMP sequences. The amount of adsorbedWPI in the first layer(1.4 mg/m2) is in the normal range for proteins like b-lactoglobulinat oil/water interfaces (Tcholakova et al., 2002), and the adsorbedamount of HMP in the second layer is also close to values of pectinadsorbed on proteins at an interface, as reported elsewhere(Ganzevles, Fokkink, Vliet, & Cohen Stuart, 2008). However, themass adsorbed in the following layers showed different behavior,the mass of WPI adsorbed in the third layer is more than 6 timeshigher than the first one, and the adsorbed amount in the fourthlayer is 6 times higher compared with the second one. This addi-tional adsorbedmass in the third and fourth layers explains, at leastin part, the big differences that were observed in stability of thecapsules with the single layer, and the single and double sequence

Fig. 5. Details of SiO2NH2 particles and supporting WPI/HMP sequences. (a) Silicamicroparticles adsorbed on a single WPI/HMP sequence. (b) Silica microparticlesadsorbed/embedded on/in a double WPI/HMP sequence.

Fig. 6. Adsorbed amount during multilayer formation on hydrophobized silicon oxideby the sequential electrostatic LbL adsorption of WPI and HMP.

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125 123

of WPI/HMP. The non-linear deposition of polyelectrolyte has beenobserved experimentally and confirmed theoretically (Lavalle et al.,2004; Porcel et al., 2006), especially when biopolymers are used.When one layer is adsorbed, the polyelectrolyte can diffuse into theprevious layer (depending on polyelectrolyte size, stiffness, andcharge density) and form contacts in and on the growing layer,therewith increasing the total space available for adsorption.

Apart from the adsorption of WPI and HMP, the deposition ofthe colloidal particles is important. As can be seen in Fig. 2c, thesilica particles that are adsorbed on a single WPI layer seem to beglued by their equatorial area, instead of adsorbed by their polarareas, and the WPI between particles forms strands that seems tobe stretched. We expect that this is caused by the excess adsorptionof the silica particles. In Fig. 2c, only a small amount of WPI ispresent in a very thin film on the droplet. After the first particles

have adsorbed, new particles may have only limited contact withthe WPI film, due to steric hindrance of the previously adsorbedparticles. There is thus a balance between hydrodynamic forcesdragging the particles away, and the elastic force from the WPI,attaching it to the capsule. For some particles the drag will be toolarge, while others will be retained. At the end, the capsule iscovered by those particles that were retained with just enoughWPI(Abu-Sharkh, 2006; Jeon, Panchagnula, Pan, & Dobrynin, 2006).This explains the stretched strands between the particles, and theroughness of the capsule.

In the system with a single WPI/HMP sequence (Fig. 5a) muchmore WPI and HMP is present, as is shown by Fig. 6. The amount ofbiopolymer present to fill the interstitial spaces between theparticles is much larger. At the same time, the surface charge is nothigher (Fig. 3), since the WPI and HMP partially compensate eachother’s charge. As a result, the amount of biopolymer relative to theamount of charge is higher. The amount of particles that adsorbs isdetermined by the overall surface (z) potential: the surface chargeis neutralized and then reverted to its equilibrium charge (due tothe charge of the particle that are not covered by the poly-electrolytes; Fig. 3). Thus, there is more WPI/HMP phase present tofill the interstitial space between the particles, and the layer isdenser than with a single WPI layer.

In case of the system with a double WPI/HMP bilayer (Fig. 5b),there is even more WPI/HMP present, while the surface has thesame surface (z) potential. Therefore, while a similar amount ofparticles will adsorb (dictated by the overall surface potential),there is more WPI/HMP phase present to fill the interstitial spaces;therefore these walls are less porous.

Even though the systems with the double WPI/HMP sequenceretain their shape better, they are still somewhat porous. Thiswould enable the loading of the capsules with active components,after their assembly. A prerequisite is that one needs to be able toclose the remaining pores after the loading. As a proof of principle,we attempted to close the scaffolds with extra layers of WPI/HMP(Fig. 7). It is clear that a single HMP layer is not enough to closea capsule having a single WPI/HMP bilayer, but when starting from

Fig. 7. LbL-Colloidosomes. (a) SiO2NH2 microparticles adsorbed on a single WPI/HMPsequence and covered with a single HMP layer. (b) SiO2NH2 microparticles adsorbed ona double WPI/HMP sequence and covered with an HMP/WPI sequence.

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125124

a systemwith a doubleWPI/HMP bilayer, the adsorption of an extraHMP/WPI sequence can close the microcapsule pores, as shown inFig. 7b. Thus, it can be concluded that closure of the porous scaf-folds is possible as long as the primary capsule has sufficiently lowporosity.

4. Conclusions

A hybrid method to prepare microcapsules was presented,combining a layer-by-layer adsorption of whey protein isolate andhigh methoxyl pectin, and deposition of silica colloidal particles forreinforcement. The stability and porosity of the capsules can bealtered by adsorbing more layers, but even after deposition of2 layers of WPI and 2 layers of HMP and a silica layer, the capsuleswere still somewhat porous. This creates the possibility to load thecapsules after their assembly with an active component. It was

shown that the remaining pores can be closed by an additionalHMP/WPI sequence.

In summary, the presented method allows the creation ofa primary capsule that can be loaded and then closed at will, usinginexpensive food-grade materials, while all process steps areoperated at room temperature. We therefore expect this method tobe of interest for application in the food and pharma industries,over all considering that the range of pH in which the exploitedelectrostatic interactions are present would make these shells aptfor protecting encapsulates from the acidity of the stomach, to laterbe dissolved in the neutral environment of the small intestine.

Acknowledgments

F.J.R.M.was supported byProgrammeAlban, the EuropeanUnionProgramme of High Level Scholarships for Latin America, scholar-ship no. E05D060840CL, and by VLAG, Wageningen University. Theauthors thank MicroNed for supporting this research, Adriaan vanAelst for his helpwith the SEM analyses, Remko Fokkink for his helpwith the reflectometry experiments, and Jolet and Riëlle de Ruiterfor preliminary experiments.

References

Abu-Sharkh, B. (2006). Structure and mechanism of the deposition of multilayers ofpolyelectrolytes and nanoparticles. Langmuir, 22, 3028e3034.

Ao, Z., Li, Z., Zhang, G., & Ngai, T. (2011). Colloidosomes formation by controlling thesolvent extraction from particle-stabilized emulsions. Colloids and SurfacesA: Physicochemical and Engineering Aspects, 384(1-3), 592e596.

Basheva, E. S., Gurkov, T. D., Christov, N. C., & Campbell, B. (2006). Interactions in oil/water/oil films stabilized by b-lactoglobulin; role of the surface charge. Colloidsand Surfaces A: Physicochemical Engineering Aspects, 282-283, 99e108.

Caruso, F., Caruso, R. A., & Möhwald, H. (1998). Nanoengineering of inorganic andhybrid hollow spheres by colloidal templating. Science, 282, 1111e1114.

Cayre, O. J., Noble, P. F., & Paunov, V. N. (2004). Fabrication of novel colloidosomemicrocapsules with gelled aqueous cores. Journal of Materials Chemistry, 14,3351e3355.

Corredig, M., & Wicker, L. (2001). Changes in the molecular weight distribution of threecommercial pectins after valve homogenization. Food Hydrocolloids, 15, 17e23.

Croll, L.M., Stöver, H. D. H., &Hitchcock, A. P. (2005). Composite tectocapsules containingporous polymer microspheres as release gates.Macromolecules, 38, 2903e2910.

Decher, G. (1997). Fuzzy nanoassemblies: toward layered polymeric multi-composites. Science, 277, 1232e1237.

Dijk, J. A. P. P. v., & Smit, J. A. M. (2000). Size-exclusion chromatographyemultianglelaser light scattering analysis of â-lactoglobulin and bovine serum albumin inaqueous solution with added salt. Journal of Chromatography A, 867, 105e112.

Dijt, J. C., Cohen Stuart, M. A., Hofman, J. E., & Fleer, G. J. (1990). Kinetics of polymeradsorption in stagnation point flow. Colloids and Surfaces, 51, 141e158.

Dinsmore, A. D., Hsu, M. F., Nikolaides, M. G., Marquez, M., Bausch, A. R., &Weitz, D. A. (2002). Colloidosomes: selectively permeable capsules composedof colloidal particles. Science, 298, 1006e1009.

Duan, H., Wang, D., Sobal, N. S., Giersig, M., Kurth, D. G., & Möhwald, H. (2005).Magnetic colloidosomes derived from nanoparticle interfacial self-assembly.Nano Letters, 5(5), 949e952.

Ganzevles, R. A., Fokkink, R., Vliet, T. v., & Cohen Stuart, M. A. (2008). Structure ofmixed b-lactoglobulin/pectine adsorbed layers at air/water interfaces: a spec-troscopy study. Journal of Colloid and Interface Science, 317, 137e147.

Georgieva, R., Moya, S., Hin, M., Mitlöhner, R., Donath, E., Kiesewetter, H., et al.(2002). Permeation of macromolecules into polyelectrolyte microcapsules.Biomacromolecules, 3(3), 517e524.

Gouin, S. (2004). Microencapsulation: industrial appraisal of existing technologiesand trends. Trends in Food Science & Technology, 15, 330e347.

Gu, Y. S., Decker, E. A., & McClements, D. J. (2007). Formation of colloidosomes byadsorption of small charged oil droplets onto the surface of large oppositelycharged oil droplets. Food Hydrocolloids, 21(4), 516e526.

Hsu, M. F., Nikolaides, M. G., Dinsmore, A. D., Bausch, A. R., Gordon, V. D., Chen, X.,et al. (2005). Self-assembled shells composed of colloidal particles: fabricationand characterization. Langmuir, 21, 2963e2970.

Jeon, J., Panchagnula, V., Pan, J., & Dobrynin, A. V. (2006). Molecular dynamicssimulations of multilayer films of polyelectrolytes and nanoparticles. Langmuir,22, 4629e4637.

Johnston, A. P. R., Cortez, C., Angelatos, A. S., & Caruso, F. (2006). Layer-by-layerengineered capsules and their applications. Current Opinion in Colloid & Inter-face Science, 11, 203e209.

Kim, J. W., Fernández-Nieves, A., Dan, N., Utada, A. S., Marquez, M., & Weitz, D. A.(2007). Colloidal assembly route for responsive colloidosomes with tunablepermeability. Nano Letters, 7(9), 2876e2880.

F.J. Rossier-Miranda et al. / Food Hydrocolloids 27 (2012) 119e125 125

Lavalle, P., Picart, C., Mutterer, J., Gergely, C., Reiss, H., Voegel, J.-C., et al. (2004).Modeling the buildup of polyelectrolyte multilayer films having exponentialgrowth. The Journal of Physical Chemistry B, 108(2), 635e648.

McClements, D. J. (2005). Theoretical analysis of factors affecting the formation andstability of multilayered colloidal dispersions. Langmuir, 21, 9777e9785.

Neff, P. A., Wunderlich, B. K., Klitzing, R. v., & Bausch, A. R. (2007). Formation anddielectric properties of polyelectrolyte multilayers studied by a silicon-on-insulator based thin film resistor. Langmuir, 23, 4048e4052.

Noble, P. F., Cayre, O. J., Alargova, R. G., Velev, O. D., & Paunov, V. N. (2004). Fabri-cation of “hairy” colloidosomes with shells of polymeric microrods. Journal ofthe American Chemical Society, 126(26), 8092e8093.

Porcel, C., Lavalle, P., Ball, V., Decher, G., Senger, B., Voegel, J.-C., et al. (2006). Fromexponential to linear growth in polyelectrolyte multilayers. Langmuir, 22,4376e4383.

Qiu, X., Donath, E., & Möhwald, H. (2001). Permeability of ibuprofen in various poly-electrolyte multilayers. Macromolecular Materials and Engineering, 286, 591e597.

Rossier-Miranda, F. J., Schroën, C. G. P. H., & Boom, R. M. (2009). Colloidosomes:versatile microcapsules in perspective. Colloids and Surfaces A: Physicochemicaland Engineering Aspects, 343(1e3), 43e49.

Rossier-Miranda, F. J., Schroën, K., & Boom, R. (2010). Mechanical characterizationand ph response of fibril-reinforced microcapsules prepared by layer-by-layeradsorption. Langmuir, 26(24), 19106e19113.

Sagis, L. M. C., de Ruiter, R., Rossier-Miranda, F. J., de Ruiter, J., Schroën, K., vanAelst, A. C., et al. (2008). Polymer microcapsules with a fiber-reinforced nano-composite shell. Langmuir, 24, 1608e1612.

Salari, J. W. O., Van Heck, J., & Klumperman, B. (2010). Steric stabilization of pick-ering emulsions for the efficient synthesis of polymeric microcapsules. Lang-muir, 26(18), 14929e14936.

Sawalha, H., Fan, Y., Schroën, K., & Boom, R. M. (2008). Preparation of hollow pol-ylactide microcapsules through premix membrane emulsification e effects ofnonsolvent properties. Journal of Membrane Science, 325(2), 665e671.

Sa�glam, D., Venema, P., de Vries, R., Sagis, L. M. C., & van der Linden, E. (2011).Preparation of high protein micro-particles using two-step emulsification. FoodHydrocolloids, 25(5), 1139e1148.

Schroën, C. G. P. H., Cohen Stuart, M. A., Voort Maarschalk, K. v. d., Padt, A. v. d., &Riet, K. v. t (1995). Influence of preadsorbed block copolymers on protein

adsorption: surface properties, layer thickness, and surface coverage. Langmuir,11, 3068e3074.

Shah, R. K., Kim, J. W., & Weitz, D. A. (2010). Monodisperse stimuli-responsivecolloidosomes by self-assembly of microgels in droplets. Langmuir, 26(3),1561e1565.

Sukhorukov, G. B., Antipov, A. A., Voigt, A., Donath, E., & Möhwald, H. (2001). pH-controlled macromolecule encapsulation in and release from polyelectrolytemultilayer nanocapsules. Macromolecular Rapid Communications, 22(1), 44e46.

Sukhorukov, G. B., Donath, E., Lichtenfeld, H., Knippel, E., Knippel, M., Budde, A.,et al. (1998). Layer-by-layer self assembly of polyelectrolytes on colloidalparticles. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 137,253e266.

Tcholakova, S., Denkov, N. D., Ivanov, I. B., & Campbell, B. (2002). Coalescence inb-lactoglobulin-stabilized emulsions: effect of protein adsorption and drop size.Langmuir, 18, 8960e8971.

Tcholakova, S., Denkov, N. D., Sidzhakova, D., Ivanov, I. B., & Campbell, B. (2003).Interrelation between drop size and protein adsorption at various emulsifica-tion conditions. Langmuir, 19, 5640e5649.

Thompson, K. L., Armes, S. P., Howse, J. R., Ebbens, S., Ahmad, I., Zaidi, J. H., et al.(2010). Covalently cross-linked colloidosomes. Macromolecules, 43(24),10466e10474.

Velev, O. D., Furusawa, K., & Nagayama, K. (1996). Assembly of latex particles byusing emulsion droplets as templates. 1. Microstructured hollow spheres.Langmuir, 12(10), 2374.

Vinogradova, O. I. (2004). Mechanical properties of polyelectrolyte multilayermicrocapsules. Journal of Physics: Condensed Matter, 16(32), R1105eR1134.

Walsh, A., Thompson, K. L., Armes, S. P., & York, D. W. (2010). Polyamine-functionalsterically stabilized latexes for covalently cross-linkable colloidosomes. Lang-muir, 26(23), 18039e18048.

Wang, D., Duan, H., & Möhwald, H. (2005). The water/oil interface: the emerginghorizon for self-assembly of nanoparticles. Soft Matter, 1, 412e416.

Yow, H. N., & Routh, A. F. (2009). Release profiles of encapsulated actives fromcolloidosomes sintered for various durations. Langmuir, 25, 159e166.

Zhu, H., Srivastava, R., & McShane, M. J. (2005). Spontaneous loading of positivelycharged macromolecules into alginate-templated polyelectrolyte multilayermicrocapsules. Biomacromolecules, 6(4), 2221e2228.