Layer-by-layer coated emulsion microparticles as storage and delivery tool

Current Opinion in Colloid & Interface Science 17 (2012) 281–289

Contents lists available at SciVerse ScienceDirect

Current Opinion in Colloid & Interface Science journal homepage: www.elsevier.com/locate/cocis

Layer-by-layer coated emulsion microparticles as storage and delivery tool Elena M. Shchukina, Dmitry G. Shchukin ⁎ Max-Planck Institute of Colloids and Interfaces, D14424 Potsdam, Germany

a r t i c l e

i n f o

Article history: Received 1 February 2012 Accepted 15 June 2012 Available online 30 June 2012 Keywords: Layer-by-layer assembly Emulsion Pickering emulsion Capsule Polyelectrolyte multilayer Ultrasonication Drug delivery Controlled release

a b s t r a c t Microencapsulation is an imperative technology in pharmacy, food industry and medicine. However, the current level of the development requires not only the fabrication of the emulsion systems, but also their functionalization in order to impart it multifunctional properties. One of the most perspective approaches to attain additional functionality to the emulsion particles is the use of the layer-by-layer modification of their surface. This technique permits the step-wise adsorption of various components (polyelectrolytes, nanoparticles, proteins, enzymes, etc.) as the layer growth is governed by their electrostatic, hydrogen bonding, hydrophobic, etc. forces and allows the formation of multilayer shells with nanometer (thickness) precision. The proposed review surveys the layer-by-layer approach for modification of both polymer and Pickering emulsions with polyelectrolyte or nanoparticle multilayers together with the demonstration of the application examples of the modified emulsion systems, where the emulsion particles play simultaneously the role of the template for layer-by-layer assembly as well as of the inner load. © 2012 Elsevier Ltd. All rights reserved.

1. Introduction The use of the emulsions is enormous. They are known from many hundreds of years and employed nowadays in food industry, paintings, fertilizers, medicine, chemical industry and many other activity areas of the mankind. The main feature of the emulsion is the possibility to disperse immiscible liquids and, at the same time, separately dissolving immiscible reagents in each liquid allowing their interaction at the liquid-liquid interface. However, the current level of the development in the food industry and medicine requires not only the fabrication of the emulsion systems, but also their functionalization in order to impart it multifunctional properties. Microencapsulation is an imperative technology in pharmacy and medicine. This is particularly important for the novel drug-delivery concepts where the lipophilic drug has to be (i) dispersed on micro or submicro level, (ii) delivered to the damaged part of the body and (iii) released in a controlled way. The application of the emulsion systems here is indispensable, because they can (i) dissolve and disperse lipophilic drug, (ii) protect it by the emulsion shell and (iii) release the drug upon shell destruction. All these three steps need the additional functionalities of the emulsion particle. The dissolution/dispersion of the lipophilic drug can be achieved by tuning the oil phase of the emulsion; magnetic properties and specific guest-host interactions can be used for the targeted delivery while the most difficult characteristic to be achieved for the emulsion-based delivery system is the controllable release. A potential benefit of using emulsions as delivery systems is that they can be fabricated entirely ⁎ Corresponding author. Tel.: +49 331 567 9781; fax: +49 331 567 9202. E-mail address: [email protected] (D.G. Shchukin). 1359-0294/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.cocis.2012.06.003

from natural food grade ingredients (lipids, proteins, polysaccharides) using simple processing operations (homogenization, mixing). They could therefore be utilized in the development of pharmaceutical products or functional foods designed to combat diet‐related diseases, such as obesity, heart disease, cancer, and hypertension. There are several ways for triggering the release of the encapsulated drug employing either external or internal factors. First is the release triggered by the changes in pH (e.g. at pH~ 6.8 in the tumor interstitium, or at pH~ 5.0 in endosomes) or by action of enzymes in the emulsion shell [1]. Delivery systems designed to control the digestion, release, and absorption of encapsulated lipophilic components within the gastrointestinal tract are being developed for a variety of applications within the pharmaceutical, biopharmaceutical, and food industries [2]. They can be used to control the release of drugs and other bioactive components at specific locations within the gastrointestinal tract, such as the mouth, stomach, small intestine, or colon. Whilst such ‘spontaneous’ release is the most desirable since it functions solely inside the body and remotely without further intervention, it can still be slow. In order to produce more controlled, rapid, and complete release of the drug from a carrier, local intervention techniques can use temperature change (hyperthermia) [3], light (photodynamic therapy) or mechanical disruption (e.g. by ultrasound) to open the shell of delivery container [4]. One of the most perspective approaches to attain additional functionality to the emulsion particles is the use of the layer-by-layer (LbL) modification of their surface. This technique permits the step-wise adsorption of various components (polyelectrolytes, nanoparticles, proteins, enzymes, etc.) as the layer growth is governed by their electrostatic attraction and allows the formation of multilayer shells with nanometer

282

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

(thickness) precision [5,6]. Besides electrostatic attraction, the other forces like hydrogen bonding or hydrophobic interactions can be utilized for layer-by-layer assembly. The possibility of tailoring different functionalities, impregnating inorganic and organic substances both inside emulsion volume and in the shell, controlled release of encapsulated material provided continuous scientific and industrial interest for employing layer-by-layer approach for delivery of different compounds. This improved mechanic stability was achieved due to the differences in the thickness, electrical charge and packing of the interfacial layers surrounding the droplets. The LbL adsorption technique is widely used since early 90-s and further, the more detailed description of this approach can be found in reviews elsewhere [7,8]. The main goal of the proposed review is the demonstration of the recent achievements of the layer-by-layer assembly approach for the functionalization of the emulsions. The review surveys the LbL modification (by LbL deposition directly on the droplets of dispersed phase) of both polymer and Pickering emulsions with polyelectrolyte or nanoparticle multilayers together with the demonstration of the application examples of the modified emulsion systems. Herein, the emulsion particles play simultaneously the role of the template as well as of the container load. 2. LbL functionalization of polymer emulsions A usual preparation method for LbL coated emulsion carriers involves several steps (Fig. 1) [9]. To stabilize the dispersed phase of initial emulsion, the oil phase (dodecane) was doped by small amount of cationic surfactant dioctadecyldimethylammonium bromide (DODAB). The colloidal stability of initial emulsion was achieved due to concentrated monolayer of strongly positively charged DODAB (z-potential was about +90 mV) at the surface of each droplet. Then, the subsequent LbL deposition was performed from concentrated (20 mg/mL) aqueous salt-free solutions of polyelectrolytes. The creamed upper layer of the strongly positively charged initial emulsion was added dropwise to 40 mL of the oppositely charged polyelectrolyte solution (poly(sodium 4-styrenesulfonate, PSS) upon continuous stirring at approximately 700 rpm. After this layer was completely brought into polyelectrolyte solution, the mixture was stirred for an additional hour to accomplish the binding of polyelectrolyte at the surfaces of the initial emulsion droplets and to ensure their overcharging. The second

encapsulation step was done in an aqueous solution of positively charged polyelectrolyte (poly(diallyldimethylammonium chloride), PDADMAC or poly(allylamine hydrochloride), PAH) in the same manner (Fig. 1). The further repetition of the alternating adsorption steps leads to the formation of containers with desired shell thickness depending on the particular final demand. Centrifugation even at low rpm values cannot be considered as a reasonable alternative to accelerate the separation because of simultaneous deformation and strong coalescence of the disperse phase droplets in spite of strong electrostatic repulsion. The further repetition of the alternating coating steps leads to the formation of capsules with desired shell thickness depending on the particular final demand. The oil core may be composed of different types of natural or artificial fats or oils like fish-oil soya-bean oil, triglycerides, hydrophobic vitamins (like vitamin B12) or drugs dissolved in oil etc. depending on the demands of the container application. Hence, the proposed approach to prepare loaded delivery systems via emulsion encapsulation can be envisaged is a general one that depends neither on the nature of the used oil nor on the specific behavior of an oil-soluble capsule load. Biopolymers such as proteins in alternation with bioemulsifiers can be effectively used for the emulsion encapsulation as was shown on diverse food-relevant systems. The influence of the outer biopolymer layer on the electrical properties and physical stability of lipid droplets coated with biopolymers β-lactoglobulin (anionic) and chitosan (cationic) was studied by McClemens et al. [10]. The droplets, which had an outer protein coating, changed from positive to negative when the pH was increased from 3 to 7. As a result, these emulsions were unstable to droplet aggregation at intermediate pH values because of the low net charge on the droplets near the isoelectric point of the adsorbed proteins. The droplets, which had an outer chitosan coating, changed from highly positive to zero when the pH was increased from 3 to 7. These emulsions were stable to droplet aggregation from pH 3 to 6, but highly unstable at higher pH values because of the low net charge on the droplets at neutral pH. Emulsions, which had an outer alginate or pectin coating, changed from slightly negative to highly negative when the pH was increased from 3 to 7. These emulsions were unstable to droplet aggregation at pH 3 because of the low net negative charge on the droplets, but could be improved if higher anionic polysaccharide levels were used.

Fig. 1. Schematic representation of several steps during L-b-L polyelectrolyte emulsion encapsulation [9].

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

If a delivery system remains stable under acid conditions but breakdown at neutral pH, then an outer coating of chitosan could be used. On the other hand, if a delivery system remains stable under neutral conditions but breakdown at acid pH, then an outer coating of alginate or pectin could be used. The presence of an outer coating of charged polysaccharides (dietary fibers) was shown to delay lipid digestibility, which has important consequences for the design of controlled delivery systems for nutraceuticals or pharmaceuticals. An electrostatic layer-by-layer deposition technique was used to prepare corn oil-in‐water emulsions (3 wt.% oil) that contained droplets coated by (1) lecithin, (2) lecithin–chitosan, or (3) lecithin–chitosan– pectin [11]. Pancreatic lipase (1.6 mg mL−1) and/or bile extract (5.0 mg mL−1) were added to each emulsion, and the particle charge, droplet aggregation, and free fatty acids released were measured. In the presence of bile extract, the amount of fatty acids released per unit amount of emulsion was much lower in the emulsions containing droplets coated by lecithin–chitosan (38 μmol mL−1) than those containing droplets coated by lecithin (250 μmol mL−1) or lecithin–chitosan– pectin (274 μmol mL−1). In addition, there was much more extensive droplet aggregation in the lecithin–chitosan emulsion than in the other two emulsions. Lipase activity was reduced in the lecithin–chitosan emulsion as a result of the formation of a relatively thick cationic layer around each droplet, as well as the formation of large flocks. The β-lactoglobulin-coated droplets were unstable at all pH, thermal treatment conditions, and NaCl content used in this study, which can be attributed to the fact that they were prepared at pH 7 and then adjusted to pH 4 before further experimentation, which promoted extensive droplet aggregation because of the low droplet charge [12]. The β-lactoglobulin–pectin coated droplets in the secondary emulsions were stable to droplet aggregation and creaming when held at 30–90 °C, at 100 mM NaCl and at pH 3–5. The β-lactoglobulin–pectin– chitosan coated droplets remained stable over a wider range of pH (3–6), but they were less stable to thermal treatment (60 °C). The improved stability of the LbL coated emulsions to droplet aggregation can be attributed to the ability of the multilayered interfaces to increase the repulsive colloidal interactions between the droplets (e.g., electrostatic and steric) and/or to increase the resistance of the interfacial membrane to rupture. The structure of nanolaminated biopolymer coatings surrounding lipid droplets was studied to find out its influence on emulsion physical stability and in vitro digestibility by pancreatic lipase [13]. Caseinate (Ca) was used as an amphoteric emulsifier, pectin (P) was used as an anionic polyelectrolyte, and chitosan (C) was used as a cationic polyelectrolyte. The electrostatic layer-by-layer deposition approach was used to prepare multilayer emulsions containing lipid droplets coated by: (1) the same coating composition but different layer order (Ca–P–C and Ca–C–P); (2) the same outer layer but different coating compositions (e.g., Ca–P, Ca–P–C–P, and Ca–C–P). The stability of the emulsions to pH changes (3 to 7) depended strongly on the order of biopolymers within the nanolaminated coatings and on the nature of the outer coating. The lipid droplets in all of the multilayer emulsions were largely digested by lipase within 30 min when monitored using an in vitro digestion model. The results suggest that encapsulation of lipids by chitosan does not inhibit their in vivo digestibility. Emulsion can be produced with electrostatic layer-by-layer deposition technologies to have cationic, thick multilayer interfacial membranes that are effective at inhibiting the oxidation of fatty acids. This study investigated the stability of spray-dried multilayer emulsion upon reconstitution into an aqueous system. The lecithin and multilayered lecithin and chitosan emulsions were spray-dried with corn syrup solids (1–20 wt.%) [14]. The lecithin–chitosan multilayer interfacial membrane remained intact on the emulsion droplets upon reconstitution into an aqueous system. Reconstituted lecithin-chitosan emulsions were more oxidatively stable than reconstituted lecithin emulsions. A minimum of 5 wt.% corn syrup solids was needed to microencapsulate the secondary emulsion droplets. Maximum oxidative

283

stability of both the powder and the reconstituted secondary emulsions was observed in samples with 5% and 20% corn syrup solids. Tuna oil can be microencapsulated by spray-drying oil-in-water emulsions containing corn syrup solids and oil droplets surrounded by multilayer interfacial membranes (lecithin: chitosan) [15]. Spray-drying produced powdered emulsions consisting of smooth spheroid powder particles (diameter = 5–30 μm) containing small tuna oil droplets (diameter b1 μm) embedded within a carbohydrate wall matrix. The structure of the microcapsules was unaffected by drying temperature (165–195 °C). The powders had relatively low moisture contents (b3%), high oil retention levels (>85%) and rapid water dispersibility (b 1 min). Tuna oil-in-water emulsion droplets coated by lecithin and chitosan produce cationic emulsion droplets that are more oxidatively stable than emulsions coated by lecithin alone. The improved oxidative stability of the emulsion droplets is likely due to its cationic nature that can repel prooxidative metals and possibly form a thicker interfacial region that could decrease interactions between lipids and water soluble prooxidants. Three 5% w/w oil in-water emulsions (5 mM phosphate buffer, pH 6.0) were prepared using the layer-by-layer electrostatic deposition method that had different interfacial compositions: (i) primary emulsion (β-lactoglobulin); secondary emulsion (β-lactoglobulin– i-carrageenan); (iii) tertiary emulsion (β-lactoglobulin–i-carrageenan– gelatin) [16]. The primary, secondary and tertiary emulsions were subjected to from one to three freeze–thaw cycles (−20 °C for 22 h, +40 °C for 2 h). The primary and secondary emulsions were highly unstable to droplet aggregation and creaming after three freeze–thaw cycles, whereas the tertiary emulsion was stable, which was attributed to the relatively thick biopolymer layer surrounding the oil droplets. In addition, the droplets coated with a two-component layer appeared to be more resistant to breakdown after a single freeze–thaw cycle than those coated with a one-component layer. In a previous study, using SDS, chitosan and pectin, the same authors also found that three-component interfacial layers were much more stable to freezing and thawing than one or two component layers [17]. These results suggest that the freeze–thaw stability of oil-in-water emulsions can be improved by engineering the characteristics of the interfacial layers surrounding the oil droplets using the LbL electrostatic deposition technique. Stability and control of mass transport across the interface are important issues for coated microdroplets. The oil droplets with caseinate as an emulsifier (particle size around 10 μm) were successively coated with casein, pectin and whey proteins [18]. The results indicate that the inner layers merge and the packing density of the interface increases. Force volume imaging is applied to probe the lateral distribution of mechanical properties of the droplet. The droplet size of the primary emulsion is increased by adding the first pectin layer from 10.18 to 11.64 μm. By adding further layers, the droplet size was not significantly affected. The droplets covered by six layers showed value of 11.61 μm. The same tendency was observed for the specific surface area. The primary emulsion had a specific surface area of 1.7 m 2/g, which shrank to values close to 1.4 for further layers. Whatever approach is chosen for shell assembly, extensive flocculation can occur in the system due to wrongly judged concentration of the adsorbing polyelectrolyte. The ranges of polyelectrolyte concentration ensuring the flocculation-free encapsulation are calculated taking into account the particle size distribution measured for emulsified droplets of sunflower oil [19]. Polycation (poly(allylamine hydrochloride)) and polyanion (poly(sodium 4-styrenesulfonate)) are taken in the theoretically projected concentrations to perform layer-by-layer assembly of a multilayer shell on the surface of oil droplets preliminary stabilized with a protein emulsifier (bovine serum albumin). The proposed theoretical model and experimental observations demonstrate the possibility to control the gravitational separation process of suspensions where each oil droplet is encapsulated in the polyelectrolyte

284

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

multilayer shell by varying the shell's thickness and mass density. Combining the theory and experimental data, the mass density of a polyelectrolyte multilayer shell assembled in a layer-by-layer fashion can be obtained. Prevention of oxidative degradation in oil-in-water emulsion has been achieved by encapsulating of each droplet of dispersed phase in antioxidant multilayer coating shell [20], which is one of the first demonstrations of the attaining surface protection to the emulsion core. The fabrication comprised placing a surface-active ionic emulsifier at the oil/water interface followed by stepwise alternate adsorption a biocompatible polyelectrolyte and antioxidant layers. Uncoupled polyelectrolyte macromolecules and antioxidant were thoroughly removed from formulation, thus the protection was entirely attributed to the droplets' shell (Schematic representation of encapsulated oil droplets comprising either shell or core antioxidant is shown in Fig. 2). The experiments were performed using linseed oil, the richest source of highly unstable omega-3 alpha linolenic essential fatty acid. Bovine serum albumin was exploited as an anionic emulsifier. The biodegradable coating shell was formed of poly-L-arginine (PARG) and dextran sulfate multilayers applied as a polycation and a polyanion, respectively. Tannic acid (TA) known as a natural antioxidant and possessing antimicrobial properties was used as a protective remedy. Oil microdroplets coated with TA-containing shell displayed physical, chemical and mechanical stability in aqueous phase and over freeze-drying process. Polyelectrolyte interfacial membrane with TA layer(s) sandwiched between PARG layers inhibited peroxidation of encapsulated lipids. Antioxidant activity of TA incorporated into the droplet's shell was compared with that of tocopherol admixed into the oil phase prior to multilayer encapsulation (Fig. 2). The water dispersed emulsion encapsulated in multilayer shell comprising tannic acid did not oxidize over 15 days of storage at 37 °C and remained intact in solution of pro-oxidant Fe2+ added in concentration 10 times exceeding the physiological concentration of iron in human blood serum. The tannic acid-containing shell scavenged pro-oxidant from the surrounding medium, thus preventing initiation of the chain reaction of lipid peroxidation in the oil core. This mechanism of antioxidant activity was found to be advantageous to preserve the intact form of polyunsaturated fatty acids compare to action of chain breaking antioxidant in the oil phase as shown by the example of mixed tocopherols. The protection provided by the antioxidant in the shell can be sufficient on its own and not requiring extra antioxidant compounds added either to the water or oil phase. Current methods for the automation of the LbL encapsulation process utilize conventional macro-scale reactors which are time consuming non-continuous processes requiring bulky and expensive equipment. These reactors not only extend the duration of the LbL process, but also impart problems such as non-uniformity and aggregation of microcapsules requiring further centrifugation, washing and re-suspension steps. Droplet-based microfluidics (a comparatively recent branch of microfluidics) involves the generation and/or manipulation of discrete liquid droplets inside microchannels. Droplet

Fig. 2. Schematic representation of encapsulated oil droplets protected by shell antioxidant and core antioxidant together with the comparative antioxidant activity of the shell antioxidant tannic acid and encapsulated core antioxidant tocopherol [20].

microfluidics can play an important role in miniaturizing LbL technique by imparting benefits of time and reagent reduction, high monodispersity of capsules and automating the entire LbL encapsulation procedure into one continuous process. Previously, Trau et al. utilized microfluidics to report the deposition of four alternative layers of polyelectrolytes (PSS/PAH–FITC) on mineral oil droplets [21], where the droplets were generated by a flow focusing geometry and travelled through various bifurcation regions for actual deposition. Non-adsorbed polyelectrolytes were removed by exploiting the Zweifach–Fung effect while the colloidal droplets remained in the main channel. However, for all the above-mentioned literature, the system size and interface complexities are proportional to the number of polyelectrolytes being deposited, as each layer of polyelectrolyte requires its own microfluidic circuitry components (e.g. pumps and channels), thereby enlarging the overall system. The next step in automation of the LbL assembly on the emulsion microparticles was demonstrated by Trau et al. in Ref. [22]. Here, the droplets generated in microdevice are also guided and diverted to the downstream direction smoothly by repeated unit rows of fabricated micropillars. By miniaturizing this “pinball” concept, the authors achieved six layers of polyelectrolyte deposition on oil droplets in less than 3 min by guiding discrete droplets through parallel laminar streams of two polyelectrolytes/polymers and a washing solution. Fig. 3 represents a schematic overview of the device. It consists of two main parts: (i) T-junction for the generation of oil droplets and (ii) main channel consisting of a single row of micropillars arranged in a zigzag configuration to guide the movement of oil droplets. The oil droplets were generated by shearing a mineral oil stream (9 μL h −1) with a continuous aqueous stream (150 μL h −1). The droplets travelled through a small side channel to enter the main channel. The main channel consisted of 3 inlet and 3 outlet ports for injecting three different solutions (polyvinylpyrrolidone solutions, washing solution and polyacrylic acid solution). The micropillar row angle was empirically determined to be 30° with the channel wall for achieving smooth droplet guiding. The design approach using micropillars has the following advantages: (a) at the T-junction, if any small satellite droplets are generated, they pass through the gaps between the pillars (40 μm) and get screened out and are collected through the leftmost outlet port of the first polyelectrolyte acting as an automatic filter to sort out satellite droplets from the targeted size droplets. (b) The zigzag arrangement of the micropillars allows twice the incubation time for polyelectrolyte deposition as that of the washing step, which provides sufficient residence time for polyelectrolyte deposition and (c) since, the number of zigzag turns of the arrangement of micropillars decide the number of layers that get deposited on a droplet, input interface components remain constant irrespective of the number of polyelectrolyte layers to be encapsulated. The collected droplets exhibited uniform fluorescence indicating successful polyelectrolyte deposition. Presented design approach not only provides a faster and more efficient alternative to conventional LbL deposition techniques, but also achieves the highest number of polyelectrolyte multilayers reported thus far using microfluidics. Fiber reinforced capsules were produced using electrostatic layer-by-layer adsorption of oppositely charged proteins and polysaccharides [23]. Alternating layers of positively charged whey protein isolate (WPI) fibrils and negatively charged high methoxyl pectin (HMP) were adsorbed onto oil droplets dispersed in an aqueous phase at a pH of 3.5. Using alternating layers of WPI fibrils and HMP gave a shell with a structure of a fiber-reinforced nanocomposite. The inner oil phase can be removed by dispersing the capsules in an appropriate solvent for the oil and freeze drying. After the removal of the oil, the capsules can be inflated by dispersing them in an aqueous solution containing the component to be encapsulated. Using emulsion droplets as a template allows the size and size distribution to be tightly controlled simply by the energy input in the initial homogenization step and the concentration of the primary emulsifier. The strength of

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

285

Fig. 3. Left: schematic view of the microfluidic system designed for LbL assembly on the emulsion particles with the expanded view of a single unit of pillars in zigzag arrangement. Right: LbL modified oil droplets (size 30 μm): (a) — bright field image, (b) — fluorescence image [22].

the shell can be influenced by varying the number of layers and the size of the capsules. Other examples of the use of the emulsion (bioemulsion) droplet as a template core for layer-by-layer assembly of polyelectrolytes followed by subsequent decomposition of the emulsion particle and formation of polyelectrolyte capsules are shown in several publications [24,25]. Ref. [24] describes a novel method for the encapsulation of DNA and proteins in hollow polyelectrolyte microcapsules prepared on human erythrocytes. Capsules were obtained by removing the cell after polyelectrolyte multilayer formation by means of NaOCl treatment. Opening and resealing of the capsule shell occur during drying and rehydration of microcapsules in the presence of the compound to be encapsulated. These capsules are rather permeable for low as well as for high molecular weight species. However, upon adsorption of extra polyelectrolyte layers the permeability decreased remarkably. The assembly of six supplementary layers of polyallylamine hydrochloride and polystyrene sulfonate rendered the capsule almost impermeable for fluorescein. Ref. [25] reports the application of frozen cyclohexane-in-water emulsions as sacrificial templates for the fabrication of hollow microcapsules through layer-by-layer assembly of poly(styrenesulfonate sodium salt) and poly(allylamine hydrochloride). Extraction of the cyclohexane phase from frozen emulsions stabilized with 11 polyelectrolyte layers by compatibilization with 30% v/v ethanol leads to the formation of water-filled microcapsules while have prepared with intact polyelectrolyte membranes as measured by their deformation induced by osmotic pressure.

3. Ultrasonic fabrication of the LbL functionalized emulsions Acoustic cavitation (the formation, growth and collapse of bubbles) provides the primary mechanism for sonochemical effects. Cavitation produces intense local heating, high pressures and very short lifetimes; these transient, localized hot spots drive high-energy chemical reactions. Measurements of the sonoluminescence spectra [26] and rates of thermo-sensitive chemical reactions of metal carbonyl substitution [27] confirmed approximately 5000 K as temperature inside a cavitation microbubble, pressures of about 1000 atm and heating and cooling rates at the interface of the cavitation microbubble above 10 10 K/s. Hence, cavitation can create extreme physical and

chemical conditions at the gas/liquid interface of the microbubble maintaining low temperature of the bulk solution. The mechanism responsible for ultrasonic microcapsule formation is a combination of two acoustic phenomena: emulsification and cavitation. When ultrasonically-prepared emulsions are used as drugs carriers, they promise some advantages. Firstly, drugs loading can be easily realized by dissolution in oil phase before sonication and different kinds of drugs can be loaded in one microsphere at the same time. Secondly, the loaded drugs can be stably carried and delivered due to the protection of protein or other biopolymer shell. Thirdly, the surface of emulsion shell can be easily modified to further functionalize these emulsion carriers. The yield of microspheres strongly depends on the temperature profile of the solution during irradiation. Optimization of the initial temperature must be made for each specific experimental configuration. In the preparation of protein microspheres by ultrasonic process, the sizes of microspheres are affected by the sonication variables, such as energy input sonication time [28]. It is found that the mean sizes of the ultrasonically prepared emulsions decrease with the increase of sonication time and sonication amplitude [29]. The dispersity of emulsion size during ultrasonic fabrication is caused by the uneven distribution of acoustic energy in the ultrasonic vessel and the region with intensive energy is restricted to areas close to the sound emitting surface of the ultrasonic probe producing the emulsion particles with the minimal size. Far from this area the acoustic energy decreases sharply, which results in the formation of big microspheres. A new targeted delivery system was developed by depositing magnetic nanoparticles on protein containers which were prepared by sonicating oil in a protein solution [30]. The deposition was conducted by using layer by layer technique and monitored by zeta potential measurement. Such prepared samples have directed movement in an external magnetic field. The hydrophobic dye (5,10,15,20 tetraphenylporphyrin (TPP)), as a model of drug, was loaded in the containers by dissolution in the oil phase before sonication. The containers loaded with dye are stable and can sustain the deposition treatment without loss of dye due to the protection of protein nanoshells. Typical procedure for the ultrasonic formation of LbL modified emulsions is shown in Fig. 4. Five milliliters of 1 wt.% human serum

286

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

Fig. 4. Schematic illustration of sonochemical preparation of magnetic protein emulsion. An emulsion is formed from an oil phase containing drug and water phase containing human serum albumin which assembles at the interface of droplet. Then, polyelectrolyte multilayers and magnetic nanoparticles were adsorbed [30].

albumin solution were put in a cylindrical vessel attached to a sonicating probe. The albumin solution was overlayered with silicon oil. The tip of the sonicator probe was brought to the interface between the two phases. The system was sonicated for several minutes at 20 kHz frequency. Then, the coating of nanoparticles on protein containers was realized by layer by layer deposition. The protein containers were firstly modified with the deposition of poly(allylamine hydrochloride), then by layer of poly(styrene sulfonate) and again poly(allylamine hydrochloride). Negative Fe3O4 nanoparticles were deposited on the emulsion container by adding the Fe3O4 colloids to the emulsion solution. The nanoparticles were adsorbed for 20 min and the excess of the nanoparticles was removed by centrifugation. The original protein containers present a net charge of −22 mV at pH 7.4 due to the highly charged protein shell resulting from both acid and basic surface residues. After deposition of poly(allylamine hydrochloride), the charge of the container reverses to positive and amounts to +49.6 mV. The coating of poly(styrene sulfonate) layer leads to the reduction of the potential, but not to charge reversal. A possible explanation of this phenomenon is that during deposition of the first poly(allylamine hydrochloride) layer, the poly(allylamine hydrochloride) molecules deeply penetrate into the soft protein shell forming a thick layer with an excess of positive charge inside, which cannot be completely compensated by the following poly(styrene sulfonate) layer. Deposition of magnetic nanoparticles also does not lead to charge

exchange, which indicates that only part of the container surface is covered by the negatively charged magnetite. The outermost layer of poly(allylamine hydrochloride) is added to prevent the flocculation of magnetic containers. To verify the magnetic property of the prepared containers, their movement in an external magnetic field was tested (Fig. 5). Fig. 5A shows the initial solution containing the dispersed magnetic containers. When a magnet is placed on one side of the vial, as shown in Fig. 5B, the brown solution becomes transparent after 30 min, and the containers are collected on the wall of the vial close to the magnet forming brown precipitates there (Fig. 5C). This result demonstrates that the prepared containers can be moved by an external magnet in the desired direction. The magnetic containers are easily redispersed in the solution by gentle agitation after removal of magnet. When a magnet is placed on the vial again, the brown precipitates forms again, which indicates that the prepared magnetic containers can be redirected without loss of the magnetic property due to the protection by the outermost poly(allylamine hydrochloride) layer. Another example of the magnetic emulsions prepared by LbL assembly is the deposition of oligochitosan as the polycation and sodium alginate as the polyanion on the surface of oil-in-water emulsion droplets containing the superparamagnetic ferroferric oxide nanoparticles and drug molecules (dipyridamole) as cores [31]. Here, the drug molecules were directly encapsulated into the interior of droplets without etching the templates and refilling with the desired guest molecules. The cumulative release ratio of dipyridamole from the oligochitosan/sodium alginate multilayer-encapsulated magnetic hybrid emulsion droplets was up to almost 100% after 31 h at pH 1.8. However, the cumulative release ratio decreases to only 3.3% at pH 7.4 even after 48 h. The synthesis of stable and functional capsules coated with chemically reduced lysozyme was demonstrated using high-intensity ultrasound in aqueous solution [32]. The lysozyme-coated capsules are stable for several months and retain the enzymatic (antimicrobial) activity of lysozyme. Two layers of poly(styrene sulfonate)/poly(allylamine hydrochloride) were assembled on the lysozyme shell. Lysozyme is a positively charged protein at pH 7 and the uncoated microbubbles had a positive potential of +40 mV. Upon poly(styrene sulfonate) and poly(allylamine hydrochloride) adsorption the surface potential of the coated particles was reversed to −30 and +40 mV, respectively. 600 nm, uniform, stable and monodisperse polyglutamate/polyethyleneimine (PEI)/polyacrylic acid (PAA) and polyglutamate/PEI/PAA/silver nanocontainers loaded with hydrophobic dye 5,10,15,20-tetraphenylporphin dissolved in toluene were fabricated combining ultrasonic technique and layer-by‐layer protocol [33]. Aqueous polyglutamate nanocontainers filled with water-insoluble liquid (toluene) were made by applying high intensity ultrasound to a two-phase system of aqueous polyglutamate solution and nonaqueous liquid. LbL assembly of a biocompatible

Fig. 5. LbL assembled magnetic/polyelectrolyte/protein emulsions in aqueous solution and their targeted movement under an external magnetic stimuli [30].

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

polyethyleneimine/polyacrylic acid shell was performed immediately after formation of the toluene loaded polyglutamate shell. The composite nanocontainers are more stable than the original polyglutamate nanocontainers without polyelectrolyte coating. They can be preserved at 2–5 °C for at least 1 month, while the initial polyglutamate nanocontainers are stable only for several hours at 2–5 °C. Any agglomeration was not observed during LbL deposition of the oppositely charged polyelectrolytes. The loaded hydrophobic material can be released by switching the polyglutamate/polyelectrolyte shell permeability through variation of environmental conditions such as pH value and ionic strength. Changing of the initial pH value has a drastic influence on the yield of polyglutamate containers. Containers are formed at the pH maintained between 6.0 and 7.0, whereas the sonochemical treatment at pH 4.0 does not yield any containers. The optimal pH is about 7.0, decreasing to pH = 6 increases the container size and polydispersity. Teng et al. demonstrated the use a lipophilic drug (rifampicin) and vegetable oil to fabricate a bio-friendly drug carrier [34]. Rifampicin is a semisynthetic antibiotic, which is derived from a form of rifamycin that interferes with the synthesis of RNA and is used to treat bacterial and viral diseases. The shell of the containers is made of polyglutamate/polyethyleneimine/polyacrylic acid multilayers. The hydrophobic drug can be released by switching the polyglutamate/polyelectrolyte shell permeability varying pH value of the media. The polydispersity index of the resulting oil-loaded polyglutamate containers is 0.15. They are stable for at least 4 months at 4 °C. 4. LbL functionalization of Pickering emulsions Pickering emulsions (or colloidosomes) are emulsions stabilized by solid particles localized at the oil-water interface. They are often more stable than conventional emulsions, as the particles at the oil-water interface exert steric hindrance and, in some cases, electrostatic repulsion to prevent emulsions from coalescing. In many cases, particle adsorption at the interface is deemed irreversible as the adsorption energy is much higher than the thermal energy, rendering extremely high emulsion stability. Generally, it is easier to assemble colloidosomes from water-in-oil emulsions, since the particles have to travel a shorter distance to the interface reducing the concentration of particles needed and producing shells with a better organization. For colloidosomes the loading can take place through the droplet that is used as a template: oil droplets may be pre-loaded with hydrophobic components, or with water droplets holding hydrophilic substances, or both. Since particle stabilized droplets resemble core shell architectures, they have a high potential to be applied in the field of active molecule encapsulation. An economically attractive aspect is the simplicity of the fabrication procedure of such “particle stabilized capsules”. In principle, only the components (solid, oil, water) need to be mixed and the application of high shear rates generates capsules with adjustable size. In comparison to surfactant based capsule production

287

no need of subsequent purification from excess surfactants is required since by choosing the right process-parameters all solids self-assemble on the available oil-water interface. The use of Pickering emulsions for making microcapsules has been explored extensively involving interlocking interfacially anchored particles using various methods to afford the so-called colloidosomes. Although such interlocking can fix the particles at the interface, interstitial pores between particles remain and need to be dealt with when designing slow or controlled release systems. Therefore, the application of the layer-by-layer assembly approach for Pickering emulsions will not only stabilize the emulsion particles due to the electrostatic repulsion, but also close the interstitial pores of the emulsion nanoparticulated shell thus providing its controlled permeability and release of the materials dissolved in the oil core. Despite on such an attractive potential, there are only few examples in the literature describing the combination of the formation of colloidosomes and LbL assembly. A hybrid method to prepare microcapsules was presented in Ref. [35] combining a layer-by-layer adsorption of whey protein isolate and high methoxyl pectin and deposition of silica colloidal particles for reinforcement. The stability and porosity of the capsules can be altered by adsorbing more layers, but even after deposition of 2 layers of whey protein isolate and 2 layers of high methoxyl pectin and a silica layer, the capsules were still somewhat porous. This creates the possibility to load the capsules after their assembly with an active component. It was shown that the remaining pores can be closed by an additional high methoxyl pectin/whey protein isolate sequence. The presented method allows the creation of a Pickering capsule that can be loaded and then closed at will, using inexpensive food-grade materials, while all process steps are operated at room temperature. Ref. [36] describes the formation, stability and properties of oil-in-water emulsions consisting of large oil droplets with small oil droplets adsorbed to their surfaces. These “colloidosomes” were formed by mixing an oil-in-water emulsion containing relatively large anionic droplets (corn oil, 600 mm, β-lactoglobulin–pectin coated, pH 4) with another oil-in-water emulsion containing relatively small cationic droplets (corn oil, 200 mm, β-lactoglobulin coated, pH 4). The z-potential of the particles went from negative (−30 mV) to positive (+25 mV) as the concentration of small cationic droplets was increased, indicating that they adsorbed to the surface of the large anionic droplets until they eventually saturated the surface. Colloidosomes were produced entirely from food-grade components (corn oil, whey protein and pectin) and may therefore have applications in products that are consumed by humans, e.g., beverages, foods, pharmaceuticals and supplements. The affinity of weak polyelectrolyte coated oxide particles to the oil-water interface can be controlled by the degree of dissociation and the thickness of the weak polyelectrolyte layer [37]. Thereby the oil in water (o/w) emulsification ability of the particles can be enabled. To demonstrate this, weak polyacid poly(methacrylic acid sodium salt) and the weak polybase poly(allylamine hydrochloride) were selected for the surface modification of oppositely charged

Fig. 6. Cryo SEM images of dodecane droplets stabilized with silica-poly(allylamine hydrochloride) particles. Corresponding pH values of emulsions are (A) 8.5, (B) 9.1, and (C) 9.8. Length of unlabeled scale bars equals 500 nm [37].

288

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289

alumina and silica colloids. To prepare the emulsion samples, first the aqueous components were mixed and, depending on the pH, colloidal or gelated suspensions of nanoparticles in water were obtained. The isoelectric point and the pH range of colloidal stability of both particlepolyelectrolyte composites depend on the thickness of the weak polyelectrolyte layer. Highly stable emulsions can be obtained when the degree of dissociation of the weak polyelectrolyte is below 80%. The thickness of the adsorbed polyelectrolyte layer strongly influences the droplet size of dodecane/water emulsion droplets but has a less pronounced impact on the diethylphthalate/water droplets. Cryo-SEM visualization shows that the regularity of the densely packed particles on the oil-water interface correlates with the degree of dissociation of the corresponding polyelectrolyte (Fig. 6). Silica-poly(allylamine hydrochloride) particles arrange themselves in a monolayer, which partially consists of some aggregates below pH 9.2. Above this pH value, flocculation of particles takes place; consequentially, the droplet shell consists almost entirely of particle aggregates. Less-pronounced but still established is the fact that for the same emulsion pH, particles with thicker polyelectrolyte coatings are capable of creating smaller droplets. The average droplet size reaches a minimum between pH 4.5 and 5.5 (0.15b R b 0.45). There is an optimal pH window (degree of dissociation between 15 and 45%) for the emulsions where the smallest droplets can be obtained. Pickering emulsions stabilized by poly(sodium styrenesulfonate) surface-modified LUDOX CL particles were used as templates for the layer-by-layer deposition of polyelectrolytes and charged nanoparticles to form composite shells [38]. Because of the residual charges on modified CL particle surfaces, the Pickering emulsions were used as templates to initiate the LbL coating of charged species from xylene/ water interface for making microcapsules. High stability of Pickering emulsions allows one the creaming by centrifugation and water washing without noticeable emulsion coalescence facilitating the handling of emulsions/capsules during LbL coating. The microcapsules resulting from repeated LbL coating with poly(diallyldimethylammonium chloride) and poly(sodium styrene sulfonate) had porous walls due to the loose arrangement of the original nanoparticle aggregates at the oil-water interface leading to significant microcapsule rupture and low encapsulation efficiency. Microcapsules formed by coating with poly(diallyldimethylammonium chloride) and anionic LUDOX HS nanoparticles are suitable for microencapsulation of hydrophobic materials with a wide range of polarities. This improved property was attributed to the dense microcapsule walls that were formed by filling the pores of LUDOX CL particle aggregates with coating LUDOX HS nanoparticles. A range of core solvents with different polarity could be utilized to make oil-in-water emulsions using the same type of particulate emulsifiers, which could be in turn be used as templates for making microcapsules with different fills. Thus, this microencapsulation method supplements the LUDOX CT-LbL assembly technique for making microcapsules and is suitable for encapsulating various hydrophobic materials with different polarities.

5. Conclusions and outlook The application of the layer-by-layer assembly procedure for the functionalization of the emulsion surfaces was successfully proven by the modification of polymer, Pickering and ultrasonically prepared emulsions by polyelectrolyte and nanoparticles assembly. On the other hand, it is also a great challenge to develop multifunctional emulsion nanocontainers able to encapsulate active material, retain it in the inner volume for a long period, and immediately release it on demand. A lot of research work still remains, especially in better understanding of the detailed mechanism of shell permeation, controlled delivery and of the structure of the inner voids. Notwithstanding to the fact that perspectives of functionalized emulsions have already been demonstrated, the necessary up-scaling technologies for

fabrication in large quantities, especially for food industry, are not yet comprehensively developed.

References [1] Rillema JA, Osmialowski EC, Linebaugh BE. Phospholipase activity in 9,10-dimethyl1,2-benzanthracene-induced mammary tumors of rats. Biochim Biophys Acta 1980;617:150–5. [2] Porter CJH, Wasan KM. Lipid-based systems for the enhanced delivery of poorly water soluble drugs. Adv Drug Deliv Rev 2008;60:615–6. [3] Needham D, Dewhirst MW. The development and testing of a new temperature-sensitive drug delivery system for the treatment of solid tumors. Adv Drug Deliv Rev 2001;53:285–305. [4] Shchukin DG, Gorin DA, Möhwald H. Ultrasonically induced opening of polyelectrolyte microcontainers. Langmuir 2006;22:7400–4. [5] Decher G. Fuzzy nanoassemblies: toward layered polymeric multicomposites. Science 1997;277:1232–7. [6] Sukhorukov GB, Donath E, Davis S, Lichtenfeld H, Caruso F, Popov V, et al. Stepwise polyelectrolyte assembly on particle surfaces: a novel approach to colloid design. Polym Adv Technol 1998;9:759–67. [7] Ariga K, McShane M, Lvov YM, Ji Q, Hill JP. Layer-by-layer assembly for drug delivery and related applications. Expert Opin Drug Deliv 2011;8(5):633–44. [8] Wong JE, Richtering W. Layer-by-layer assembly on stimuli responsive microgels. Curr Opin Colloid Interface Sci 2008;13(6):403–12. [9] Grigoriev DO, Bukreeva T, Möhwald H, Shchukin DG. New method for fabrication of loaded micro- and nanocontainers: emulsion encapsulation by polyelectrolyte L-b-L deposition on the liquid core. Langmuir 2008;24:999–1004. [10] Li Y, Hu M, Xiao H, Du Y, Decker EA, McClemens DJ. Controlling the functional performance of emulsion-based delivery systems using multi-component biopolymer coatings. Eur J Pharm Biopharm 2010;76:38–47. [11] Mun S, Decker EA, Park Y, Weiss J, McClements DJ. Influence of interfacial composition on in vitro digestibility of emulsified lipids: potential mechanism for chitosan's ability to inhibit fat digestion. Food Biophys 2006;1:21–9. [12] Güzey D, McClements DJ. Influence of environmental stresses on O/W emulsions stabilized by β-lactoglobulin–pectin and β-lactoglobulin–pectin–chitosan membranes produced by the electrostatic layer-by-layer deposition technique. Food Biophys 2006;1:30–40. [13] Hu M, Li Y, Decker EA, Xiao H, McClements DJ. Impact of layer structure on physical stability and lipase digestibility of lipid droplets coated by biopolymer nanolaminated coatings. Food Biophys 2011;6:37–48. [14] Shaw LA, Decker EA, Xiao H, McClements DJ. Spray-dried multilayered emulsions as a delivery method for ω-3 fatty acids into food systems. J Agric Food Chem 2007;55(8):3112–9. [15] Klinkesorn U, Sophanodora P, Chinachoti P, Decker EA, McClements DJ. Characterization of spray-dried tuna oil emulsified in two-layered interfacial membranes prepared using electrostatic layer-by-layer deposition. Food Res Int 2006;39: 449–57. [16] Gu YS, Decker EA, McClements DJ. Application of multi-component biopolymer layers to improve the freeze–thaw stability of oil-in-water emulsions: β-Lactoglobulin–i-carrageenan–gelatin. J Food Eng 2007;80:1246–54. [17] Thanasukarn P, Pongsawatmanit R, McClements DJ. Utilization of layer-by-layer interfacial deposition technique to improve freeze–thaw stability of oil-in-water emulsions. Food Res Int 2006;39:721–9. [18] Wackerbarth H, Schon P, Bindrich U. Preparation and characterization of multilayer coated microdroplets: droplet deformation simultaneously probed by atomic force spectroscopy and optical detection. Langmuir 2009;25:2636–40. [19] Sadovoy AV, Kiryukhin MV, Sukhorukov GB, Antipina MN. Kinetic stability of water-dispersed oil droplets encapsulated in a polyelectrolyte multilayer shell. Phys Chem Chem Phys 2011;13:4005–12. [20] Lomova MV, Sukhorukov GB, Antipina MN. Antioxidant coating of micronsize droplets for prevention of lipid peroxidation in oil-in-water emulsion. ACS Appl Mater Interfaces 2010;2:3669–76. [21] Peng K, Ong W, Yobas L, Trau D. 11th Intl. Conf. on mTAS, Paris, France; 2007. [22] Kantak C, Beyer S, Yobas L, Bansala T, Trau D. A “microfluidic pinball” for on-chip generation of layer-by-layer polyelectrolyte microcapsules. Lab Chip 2011;11: 1030–5. [23] Sagis LMC, de Ruiter R, Rossier FJ, de Ruiter MJ, Schroen K, van Aelst AC, et al. Polymer microcapsules with a fiber-reinforced nanocomposite shell. Langmuir 2008;24:1608–12. [24] Kreft O, Georgieva R, Baumler H, Steup M, Muller-Rober B, Sukhorukov GB, et al. Red blood cell templated polyelectrolyte capsules: a novel vehicle for the stable encapsulation of DNA and proteins. Macromol Rapid Commun 2006;27:435–40. [25] Khapli S, Kim JR, Montclare JK, Rastislav Levicky R, Porfiri M, Sofou S. Frozen cyclohexane-in-water emulsion as a sacrificial template for the synthesis of multilayered polyelectrolyte microcapsules. Langmuir 2009;25:9728–33. [26] Ashokkumar M, Grieser F. A comparison between multibubble sonoluminescent intensity and the temperature within cavitation bubbles. J Am Chem Soc 2005;127:5326–7. [27] Suslick KS, Hammerton DA, Cline RE. Sonochemical hotspot. J Am Chem Soc 1986;108:5641–2. [28] Landfester K, Eisenblätter J, Rothe R. Preparation of polymerizable miniemulsions by ultrasound. JCT Res 2004;1:65–8. [29] Han Y, Radziuk D, Shchukin DG, Möhwald H. Stability and size dependence of protein microspheres prepared by ultrasonication. J Mater Chem 2008;18:5162–6.

E.M. Shchukina, D.G. Shchukin / Current Opinion in Colloid & Interface Science 17 (2012) 281–289 [30] Han Y, Radzuik D, Shchukin DG, Möhwald H. Sonochemical synthesis of magnetic protein container for targeted delivery. Macromol Rapid Commun 2008;29: 1203–7. [31] Mu B, Liu P, Du P, Dong Y, Lu C. Magnetic-targeted pH-responsive drug delivery system via layer-by-layer self-assembly of polyelectrolytes onto drug-containing emulsion droplets and its controlled release. J Polym Sci Part A: Polym Chem 2011;49:1969–76. [32] Cavalieri F, Ashokkumar M, Grieser F, Caruso F. Ultrasonic synthesis of stable, functional lysozyme microbubbles. Langmuir 2008;24:10078–83. [33] Teng X, Shchukin DG, Möhwald H. Encapsulation of water-immiscible solvents in polyglutamate/polyelectrolyte nanocontainers. Adv Funct Mater 2007;17: 1273–8.

289

[34] Teng X, Shchukin DG, Möhwald H. Ultrasonically fabricated polyelectrolyte nanocontainers. Langmuir 2008;24:383–9. [35] Rossier-Miranda FJ, Schroën K, Boom R. Microcapsule production by an hybrid colloidosome-layer-by-layer technique. Food Hydrocolloids 2012;27:119–25. [36] Gu YS, Decker EA, McClements DJ. Formation of colloidosomes by adsorption of small charged oil droplets onto the surface of large oppositely charged oil droplets. Food Hydrocolloids 2007;21:516–26. [37] Haase MF, Grigoriev D, Möhwald H, Tiersch B, Shchukin DG. Nanoparticle modification by weak polyelectrolytes for pH-sensitive Pickering emulsions. Langmuir 2011;27:74–82. [38] Li J, Stöver HDH. Pickering emulsion templated layer-by-layer assembly for making microcapsules. Langmuir 2010;26:15554–60.