F E A T U R E A R T I C L E

Journ

al of

Materials

Ch

emistry

ww

w.rsc.o

rg/m

aterials

Porous calcium carbonate microparticles as templates for

encapsulation of bioactive compounds

Gleb B. Sukhorukov,a Dmitry V. Volodkin,a,b Anja M. Gunther,a Alexander I. Petrov,c

Dinesh B. Shenoya and Helmuth Mohwalda

aMax-Planck Institute of Colloids and Interfaces, Potsdam/Golm 14424, GermanybMoscow State University, Department of Chemistry, Moscow 119992, RussiacInstitute of Theoretical and Experimental Biophysics, Russian Academy of Sciences,

142290 Pushchino, Moscow Region, Russia

Received 19th February 2004, Accepted 7th May 2004

First published as an Advance Article on the web 15th June 2004

The paper describes the preparation and characterisationof porous calcium carbonate microparticles with anaverage size of 5 mm and their use for encapsulation ofbiomacromolecules. The average pore size of about 30–50 nm enables size selective and time-dependentpermeation of different macromolecules. Layer-by-layeradsorption of polyelectrolytes into these particlesfollowed by core dissolution leads to formation ofinterconnecting networks (matrix-like structure) made ofpolyelectrolyte complexes. The structure can be used foraccumulation of bio-macromolecules, mainly proteins.Besides the inter-polyelectrolyte structure templated onporous CaCO3 microparticles the microgel particles(‘‘ghost’’) can also be made inside by complexing alginateand calcium. The adsorption of biomacromoleculesinside the porous calcium carbonate particles is

presumably regulated by electrostatic interactions on themicroparticle surface within pores and protein–proteininteractions. Protein adsorption into CaCO3 micro-particle voids together with layer-by-layer assembly ofbiopolymers provide a way for fabrication of completelybiocompatible microcapsules envisaging their use asbiomaterials.

Introduction to layer-by-layer (LbL) surfaceengineering of colloidal particles and novelmicroencapsulation technique

Design and engineering of functional colloidal particles ofmicron and submicron size have attracted considerable atten-tion in the last decade because of their significance in

DO

I:1

0.1

03

9/b

40

26

17

a

Gleb Sukhorukov received the degree ofMaster of Science in 1991 at the Depart-ment of Physics, Moscow State Universityand a Ph.D. in speciality biophysics atthe same Department in 1994. Then heworked as a postdoctoral researcher at theUniversity of Mainz and Max-Planck-Institute of Colloids and Interfaces. In2000 he co-founded the start-up company"Capsulution NanoScience" based onencapsulation technology developed inthe Institute. He worked in the companyas project manager till he won in 2001the Sofja Kovalevskaja Award of theAlexander von Humboldt Foundation andreturned to the Institute as group-leader.

His research field crosses Physical Chem-istry, Biophysics and Material Sciencecomprising physics and (bio)-chemistryon submicron dimensions, design of multi-functional colloidal particles and capsulesand nano-engineered biomaterials.

Dmitry Volodkin was born in Moscow,Russia in 1979. He received the degree ofMaster of Science in 2001 at the Depart-ment of Chemistry, Lomonosov MoscowState University. Currently he is a Ph.D.student at the Lomonosov Moscow StateUniversity and Max Planck Institute ofColloids and Interfaces in the groupof Dr. G. Sukhorukov. His principle

research interests are in the fieldof microencapsulation and controlledrelease of bioactive materials usinglayer-by-layer polyelectrolyte assembly.

Anja M. Gunther was born in 1975 inRostock, Germany. She studied Biochem-istry at the University Potsdam, Germany,and received her Diploma degree in Mole-cular Biology in 2001. Currently, she iscompleting her Ph.D. Thesis under thesupervision of Dr. G. Sukhorukov at theMax-Planck-Institute of Colloids andInterfaces. Her work is concerned withencapsulation of proteins and DNA inbiocompatible microcapsules.

Dmitry VolodkinGleb Sukhorukov Anja Gunther

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1 2 0 7 3T h i s j o u r n a l i s � T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

AView Online / Journal Homepage / Table of Contents for this issue

biotechnological and nanotechnological frontiers and potentialfor exploitation in biology, medicine, catalysis, ecology,nutrition etc. Up to now several techniques have beeninvestigated and applied to encapsulate/incorporate varioussubstances of interest into different micro- and nano-particles/capsules. Some of the frequently employed approaches are:liposomal (or vesicle-based) techniques, polymeric micro- andnano-particles (mainly prepared using preformed polymers—leading to the formation of matrix systems or micro-gel beads),polymeric micro- and nano-capsules prepared using techniquessuch as interfacial polymerization, self-organization etc.The reader can deal with these technologies in fine details inthe comprehensive series of books edited by Arshady1 as in thecurrent article we intend to concentrate upon development ofa novel approach based on template-assisted assembly ofultrathin polymeric films on colloidal surfaces.

The nanostructured films are constructed by the LbLsequential physisorption of oppositely charged polyelectro-lytes. This simple concept for the design of composite materialswas originally proposed by Iller, but brought to the limelight bythe systematic and extensive research carried by Decher and co-workers to fabricate ultrathin films on macroscopically flatsurfaces.2 Following this, in 1998 the technology was success-fully transferred for surface nanoengineering of micron andsub-micron sized core-particles by alternate exposure to poly-electrolytes having counter-charges.3 As a further extension,the colloidal cores were eliminated (by dissolution or decom-position) by suitable means keeping the assembled polyelec-trolyte multilayers stable. This led to the formation of hollowstructures replicating the templating particles in terms of sizeand shape. The existing shell wall of the polyelectrolyte multi-layer films assembled via the LbL technique may be composedof numerous charged species, hence it is not necessarily limitedto polyelectrolytes alone. Inorganic nanoparticles, lipids,

multivalent ions and certain biological polymers like proteinsand nucleic acids have been employed as components of theseshells. The greatest advantage of the LbL protocol is thestriking simplicity with which the shell thickness can be tunedto nanometric precision by controlling the physical chemistryand the number of adsorbed molecular layers.

On the other hand, the LbL technology adds-up theflexibility of using a broad variety of colloidal species astemplates or core materials. They could be inorganic or organicparticles of size from 20 nm to tens of microns, drug or dyemicro-/nano-crystals, compacted DNA, protein aggregates andbiological cells.4–8 When used as decomposable supports, thecores could be eliminated by dissolution or decomposition(assisted by suitable physicochemical means such as solvents,pH change, thermal treatment etc.). Weakly crosslinkedmelamine formaldehyde (MF) particles were the originallyemployed and most intensively studied templates for manu-facturing of hollow polyelectrolyte capsules. They dissolve atlow pH or in water-miscible organic solvents such as dimethylsulfoxide. Incomplete elimination of MF-oligomers (formedduring dissolution) and their biological incompatibility havestrongly limited the usability of these cores beyond theacademic level.9 Biodegradable polymeric microparticles (e.g.

homo- and co-polymers of polylactic acid) convenientlyovercome the drawbacks of MF—but pose other limitationslike polydispersity and aggregation tendency.10 Traditionalinorganic cores such as silica oxides can be completely elimi-nated by dissolution, but demand use of hazardous solventssuch as hydrofluoric acid for their dissolution.11 We havesuccessfully explored even biological cells (like red blood cells)as templates for LbL assembly wherein the core removal wasfacilitated by strong oxidizing agents.12 Each of these decom-posable templates have their limitations and there can be nouniversal template for the LbL technology. There is an ongoing

Alexander I. Petrov graduated to aMasters Degree in Chemistry in 1970 atLomonosov Moscow State University.Since 1972 he has been working in theInstitute of Biophysics (in 1989 theInstitute of Theoretical and ExperimentalBiophysics) in Puschino, Russia. Hereceived his Ph.D. degree in biology in1985. Since 1988 he has been SeniorResearcher at this institute (Laboratoryof Physical and Radiation Chemistry ofBiopolymers). His research interests arein biophysical chemistry and molecularbiophysics.

Dinesh B. Shenoy received his PhD in

Pharmaceutical Sciences from the Collegeof Pharmaceutical Sciences, Manipal,India. He worked as a post-doctoralfellow at MPI of Colloids and Interfaces,making contributions in employing LbLnano-engineering for biological andpharmaceutical applications. Currentlyhe is working as an Associate ResearchScientist at Department of Pharmaceuti-cal Sciences, Northeastern University,Boston, USA.

Helmuth Mohwald received his Diplomain Physics in 1971 at the University ofGottingen, Germany, and his Ph.D. degreein Physics in 1974 at the Max Planck

Institute of Biophysical Chemistry,Gottingen. In 1978, he finished hishabilitation at the University of Ulm,Germany. In 1981–1987, he was C3professor at the Technology Universityof Munchen, Germany. In 1987–1993, heheld the Chair of Physical Chemistry (C4professor) at the University of Mainz,Germany. Since 1993, he has been thedirector and scientific member of the MaxPlanck Institute of Colloids and Inter-faces, Potsdam, Germany. His mainresearch interests include biomimetic sys-tems, chemistry and physics in confinedspaces, dynamics at interfaces, and supra-molecular interactions.

Helmuth MohwaldAlexander Petrov Dinesh Shenoy

2 0 7 4 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

search for new templates for custom-based applications and thecarbonate cores have been a part of this effort in our group. Inaddition a ‘‘clean’’ core removal is also a mandatory prerequisitefor establishment of defined physico-chemical properties.

Applicability of the polyelectrolyte capsules for micro-encapsulation of substances of varying physicochemical pro-perties and biological functionalities and their subsequentrelease have been intensively studied in last years. Collectively,researchers have utilized covering of precipitated bioactivematerial,6,7 the sensitivity of the polyelectrolyte multilayers toconditions such as pH, salt concentration and solvent tocontrol their permeability and hence movement of substancesacross the shells. Incorporation of charged and non-chargedmacromolecules has been achieved via fabrication of double-walled capsules with subsequent decomposition of the innerwall.13 Moreover, polymer encapsulation via controlled poly-meric synthesis inside the capsules has also been demon-strated.14 These approaches have some disadvantages such asformation of stable cores with certain surface properties for thefirst method, low incorporation efficiency for the second oneand limited use of employed polymers for the other appro-aches. Encapsulation of reactive polymers or enzymes pro-motes the concept of microreactors of open type where smallsolutes can penetrate the shell wall while macromolecule stay inthe interior.15 The permeability of small molecules and ions canalso be tuned by variation of the layer number.16 In summary, atemplate that can impart the multiple functionalities listedabove is desirable and we have found that the CaCO3 templatefulfills many of the desirable properties.

A pre-requisite for an effective biological application of thesepolyelectrolyte micro- and nano-capsules synthesized viatemplate-directed assembly is the use of biocompatible shelland core components and a biofriendly processing technique.In the current article, we address the issue of the core substancethat is compliant not only with the requirement of biocom-patibility but also possesses multiple functionalities withrespect to biological applications.

1. Synthesis and characterization of calciumcarbonate (CaCO3) microparticles

Crystallization of calcium carbonate from supersaturated solu-tions has been the subject of many studies due to its importancein geo-/bio- and material sciences, as well as due to its wideapplications in industry, technology, medicine and many otherfields. Most of the research has been dedicated to elucidatingthe formation of inorganic nanoparticles in saturated solu-tions,17–19 understanding kinetics and mechanisms of crystal-lization and mutual transformations between CaCO3

polymorphs—calcite, aragonite, vaterite.17–23 As a rule, thedirect mixing of soluble salts of Ca21 and CO3

22 results in anamorphous precipitate initially, which eventually transformsinto aggregated CaCO3 microcrystals with a particular mor-phology. The CaCO3 microparticles obtained by this simpleroute are uniform and homogeneously sized, non-aggregated,highly porous spheres. The quality of the resultant micro-particles was found to be strongly dependent on the experi-mental conditions such as the type of the salts used, theirconcentration, pH, temperature, rate of mixing the solutionsand the intensity of agitation of the reaction mixture—parameters affecting the rate of the nucleation process.17–23

Inclusion of different additives such as divalent cations,organic solvents and macromolecules (synthetic or natural)added to the reaction mixture were shown to exert a profoundeffect on the morphology of the CaCO3 microparticlesformed.24–26

We have standardized a simple and reproducible procedurefor preparation of spherical CaCO3 microparticles with anarrow size distribution ranging from 4 to 6 mm.27 The

amorphous nanoprecipitates instantly formed upon mixingthe CaCl2 and Na2CO3 solutions were found to transforminto microparticles with spherical morphology with porous,channel-like internal structure (Fig. 1a, b).

Porosity is an important feature of CaCO3 microparticles.We applied the Brunauer–Emmett–Teller (BET) method ofnitrogen adsorption/desorption in order to determine thesurface area of CaCO3 microparticles (mean diameter 5 mm)and an effective pore size distribution. Nitrogen adsorption–desorption measurements revealed a surface area of 8.8 ¡

0.3 m2 g21 and an average pore size of 35 nm (Fig. 1c).27

Theoretical calculations of the surface area of non-porous

microparticles (MF) with an identical diameter yield a surfacearea value of 0.8 m2 g21. This demonstrates that due to itshighly porous architecture the effective surface area of theCaCO3 microparticles is about 11 times greater than that of acompact particle.

Fig. 1 SEM images of CaCO3 microparticles. Single particle (right topcorner—overview) (a) and broken particle (b). Pore size distribution ofCaCO3 microparticles (c). Scale bar 1 mm.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1 2 0 7 5

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

2. Adsorption on CaCO3 particles

2.1. Permeability studies: effect of pH on macromoleculeadsorption

High surface area and the presence of nanometer-sized poresand channels in CaCO3 particles offer a unique opportunity tocapture biomacromolecules such as proteins via physicaladsorption/pore diffusion, thus enabling very high substrateloading. When we consider macromolecule loading in such atype of particles, molecular interactions between the substrateand the surface of CaCO3 microparticles play an importantrole. The pH of the medium would be a decisive factor whendealing with a charged substrate or a macromolecule bearingionizable groups (e.g. protein). However, when the macro-molecule is uncharged (e.g. dextran), it was observed thatdiffusion-related limitations resulted in different permeabilitiesof CaCO3 microparticles depending on the molecular weight(Fig. 2a, b).28 We investigated the phenomenon mentionedabove with a protein (lactalbumin, Lact) and a polysaccharide(dextran, MW y4 kDa). These molecules have an average sizeof about 3–4 nm and hence can easily permeate into the internalvolume of CaCO3 microparticles through the pores of varyingsize (20–60 nm as determined by BET studies) or remain localizedon the surface of the particle by virtue of adsorption.28

As experimental evidence shows in Fig. 2c, the amount ofadsorbed dextran was independent of the pH conditions atwhich the microparticle was exposed to the macromoleculesimply due to the absence of a charge on the polysaccharide.

The isoelectric point (pI) for CaCO3 microparticles as mea-sured by microelectrophoresis was 8.5.28 Hence, the particlesare positively charged when the pH of the suspending mediumis less than 8.5 (7.0 and 8.0 in our experiments). The proteincandidate (Lact) is negatively charged at this pH. As a result,we observe a doubled extent of adsorption for Lact at pH 7.0and 8.0 compared to their alkaline counterparts (pH 9.0 and10.0) at which the surface charge of the core particles isreversed to negative—resulting in electro-repulsion with theproteins having identical charge.28

A change in the adsorbed amount of less than 20% for Lactadsorption on charged poly(styrene sulfonate) latices wasdemonstrated when the adsorption reached the maximumestimated in ref. 29. This indicates that a contribution of lateralrepulsion of Lact in a compact monolayer structure influencesthe pH-dependent adsorption but does not induce a largechange in pH 7.0–10.0 far from the pI, which is 4.5. Surfacesaturation for Lact is reached at a protein capacity higher thanthe obtained value. The calculated value of the adsorbedamount for a compact monolayer is 1.5 pg per microparticle.28

It can point at a not close-packed structure of adsorbedmolecules resulting in lower lateral repulsion. At the same timeone can assume that not all the internal surfaces of CaCO3

microparticles determined by BET are available for proteinmolecules. Molecules of N2 (BET) can easy reach all internalsurfaces in even small pores and the surfaces of irregular andrough internal channels of the porous CaCO3 structure.Because of steric difficulties, only part of the internal surfacearea is used for Lact adsorption which can lead to a smalleramount of adsorbed protein than needed to form a compactmonolayer.

We observe a significantly large adsorption of protein on thesurface of CaCO3 particles even when both species possess thesame sign of charge.28 The same is true when an unchargedsubstrate like dextran is exposed to CaCO3. Although globalelectrostatic forces undoubtedly affect adsorption, they are notthe only prevailing forces and hence do not dominate thisprocess. There are various sub-processes that determine macro-molecule adsorption onto solid–liquid interfaces30,31 like: a)electrostatic interaction; b) steric interactions due to thepolymeric components at the sorbent surface that extend

into the surrounding aqueous phase; c) changes in the state ofhydration and; d) rearrangements in the macromolecularstructure. The pH of the medium alters the degree of influenceof the electrostatic part of the particle–substrate interaction(especially when the substrate is charged or chargeable)providing a tool to manipulate adsorption–desorption kinetics,hence an opportunity to control the amount of chargedmacromolecules incorporated.

3. Formation of matrix-type polyelectrolytemicrocapsules

The porous nature of the CaCO3 microparticles that could beused as decomposable templates to obtain polyelectrolytecapsules offers yet another interesting feature—the possibilityto obtain a matrix consisting of a polyelectrolyte complex

Fig. 2 Confocal images of CaCO3 microcores after incubation withdextrans followed by washing: dextran-FITC 4 kDa (a) and 2000 kDa(b). Adsorption isotherms for dextran and Lact in CaCO3 micro-particles as a function of pH (c).

2 0 7 6 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

within the capsule interior having similar chemistry as the shell.When the polyelectrolyte multilayer shell was built around theCaCO3 core by the LbL protocol and the core was extractedwith a chelating agent, the resultant assembly was not just thehollow polyelectrolyte capsule, but a capsule having a definedmatrix-type interior.27 Fig. 3 shows the schematic representa-tion of the concept. We used two oppositely charged poly-electrolytes as shell components, namely, polystyrene sulfonate(PSS, negatively charged polymer) and poly(allylamine)hydrochloride (PAH, positively charged polymer). We usedscanning confocal Raman microscopy as the tool to identifythe chemical composition of the capsule wall as well as theinterior.27 We could detect both polyelectrolyte species at bothlocations indicating the formation of the polyelectrolyte com-plex (interpenetrating network) within the interior as wellgiving rise to a matrix type structure. This was confirmed byconfocal laser scanning microscopy (CLSM) imaging by usingfluorescently labeled PAH.27

We used microgravimetry (quartz crystal microbalance,QCM) as a tool to register the polyelectrolyte complex that wasformed not only as a wall component but also as a matrixcomponent.27 The average weight of a single CaCO3 microcorewas approximately 88 pg. After the elimination of the core, theweights of the polyelectrolyte capsules with 6 and 16 layerswere 11 and 24 pg respectively. As reported earlier, PSS-PAHmultilayers were assembled on MF templates having a smooth,non-porous morphology (limiting the formation of the poly-electrolyte complex only to the capsule wall and not to theinterior), the weight of the hollow capsule was about 1.4 pg for8 layers.32 The increased mass of the capsule is attributed to thepresence of an increased amount of polyelectrolyte complex inthe capsules obtained with CaCO3 microsupports and whenanalyzed in the light of microscopy data, it indicates the con-tribution coming from the matrix-type interior of the capsule.

3.1. Macromolecule encapsulation into pre-formedmicrocapsules

Our previous investigations have shown that many low/highmolecular weight water-soluble substrates such as dyes, poly-electrolytes, polymers and proteins can be retained within thepolyelectrolyte multilayer capsules obtained by using MF-particles as decomposable templates.33 Later, we showed thatthese were caused by the gel-like internal structure of theresidual MF oligomers adsorbed to the polyelectrolytes usedfor capsule preparation.34

When CaCO3 particles are used as decomposable templatesfor assembly of polyelectrolyte multilayers by the LbL tech-nique, the subsequent elimination of the support by exposure toan acid or chelating agent results in complete removal of thetemplate without any residues.11,27 As described in the

preceding section, the capsule interior has only a matrix-typestructure, essentially composed of the complex of polyelec-trolyte pairs used for the LbL assembly. When thesepolyelectrolyte capsules are exposed to a macromolecule(such as a protein—bovine serum albumin (BSA) or apolysaccharide—dextran), one observed capture and sponta-neous accumulation of the macromolecules within the capsuleinterior.27 The concept was depicted in the scheme (Fig. 3) andis demonstrated in Fig. 4. Fluorescence spectroscopy and QCManalysis were used to determine the amount of labeled BSA anddextran captured within the microcapsules.27 The results arefound to be analogous for two independent analyses and themicrocapsules show a high active loading capacity for BSA anddextran. With QCM analysis, the weight of the BSA capturedwas found to be about 15 pg for a single capsule having 16polyelectrolyte layers. This gives an average concentration ofabout 250 mg mL21 of protein per capsule. Comparable figureswere obtained when the amount captured was estimated byfluorescence measurements. We do not know the microstruc-ture of the capsule interior—except that it consists grossly of amatrix made from a polyelectrolyte complex. The qualitativeanalysis of the loaded capsules indicates that the capturedmacromolecule is essentially in a complexed or aggregatedform rather than in a free state.

4. Biocompatible capsules

Calcium carbonate is a component of the biological systemand is considered as safe material for administration into a

Fig. 3 Scheme of microcapsule fabrication and encapsulation of macromolecules into capsules.

Fig. 4 CLSM image of microcapsules (16 steps of polyelectrolyteadsorption, first—PAH) after incubation with BSA-FITC followed bywater washing.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1 2 0 7 7

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

biological system. When it comes to its usage as a decom-posable template, we have proven that it is completely elimi-nated with no residual element left behind within thepolyelectrolyte capsules.11,27 This leaves us to deal solelywith the polyelectrolytes that constitute the capsule shell. Inprevious paragraphs, we described microcapsules made of themost widely investigated polyelectrolyte pairs, namely PSS andPAH. Both are regarded as bio-incompatible and hence the useof capsules made with them for biological applications is notpossible. One needs to refabricate the capsules with biocom-patible polyelectrolytes to make them bio-friendly. In order toimpart biocompatibility to the polyelectrolyte capsules, wehave exploited several biopolymers as multilayer constituents.They include polyanionic polymers such as alginates, poly-glutamic acid, dextran sulfate etc. and polycationic polymerssuch as polyornithine, protamine and chitosan. We haveinvestigated possible combinations of these biopolymer pairswith CaCO3 microparticles as the core substance. The moststable capsules were obtained with the polyornithine (150 kDaM.W.)-alginate/polyglutamic acid pair. The exploitation ofother combinations of polyelectrolyte pairs leads to strongaggregation of particles during the LbL assembly procedure.This was more pronounced when chitosan was used as thepolycation. Apparently, it can be suggested that the adsorbedchitosan does not introduce enough electrostatic repulsionand hence screens the inter-particulate adhesion. The use ofprotamine as polycation failed mainly at the stage of coredissolution, as polyelectrolyte capsules obtained with prota-mine with any of the polyanions listed above could not retainstability due to the limited number of positively charged groups

available in protamine for electrostatic complexation. Themost remarkable feature of successfully fabricated biocom-patible capsules having polyornithine and polyanions was theabsence of a matrix structure within the capsule interior.Fig. 5a shows capsules made of 3 layer-pairs of polyornithine/alginate with the first layer of polyornithine being labeled witha rhodamine derivative. We clearly see no fluorescence from theinterior of the capsules (as was the case when the interiorshowed a matrix-type structure with a PSS-PAH pair). Thiscould be because of slow penetration of the bulkier biomacro-molecules into pores of the template. When the first layer ofmacromolecules was still on the exterior of the CaCO3 particle(not yet diffused into the pores), it poses the possibility of beingcomplexed there itself by the incoming second layer of oppo-sitely charged polyelectrolyte—thus effectively blocking anyfurther pore-mediated diffusion. In Fig. 5a one can see theconfocal microscopy image of polyornithine/alginate acid cap-sules in the presence of FITC-dextran. The capsules are notpermeable for high molecular weight dextran in a wide range ofpH and salt concentration indicating the closure of pores foractive loading of the substrates. In any case, the CaCO3

microparticles offer the feasibility to fabricate completelybiocompatible capsules via template-directed LbL assembly.This opens up an opportunity to explore these capsules in theirinteraction with cell tissue in order to develop delivery systems.

5. Fabrication of microgel particles/capsules bytemplate-assisted-shell reinforcement

Unlike classical decomposable templates that function assupports for build-up of multilayers on them, CaCO3 cores canalso serve as a source for cross-linking that hardens thedeposited shell instantly during the core dissolution process.Depending upon the shell wall material and the method usedfor its deposition, the CaCO3 assisted shell reinforcement mayresult in formation of micro-gel-like structures. A schematicrepresentation of the concept is depicted in Fig. 6.

Consider a shell being formed by deposition of sodiumalginate using the surface controlled precipitation (SCP)method taking advantage of its insolubility in hydro-alcoholicsolutions (approx. 30% v/v or more of ethanol). It is establishedthat the SCP technique results in formation of thicker shellwalls [for details about SCP, see refs. 8, 13, 35, 36]. In addition,in the case of CaCO3, a considerable amount of alginatematerial that is being precipitated as a fine colloidal substancegets trapped within the fenestrations and channels of thehighly porous template. When CaCO3 is decomposed with

Fig. 5 Confocal fluorescence images. A – Polyornithine labeled withrhodamine/alginate capsules made of 3 double layers on porouscalcium carbonate microparticles. B – Capsules made of 4 double layersof polyornithine/alginate being impermeable for dextran labeled withfluorescein (molecular weight 70 kDa). The images were taken in water.

Fig. 6 Schematic representation of fabrication of microcapsules by SCP followed by template-reinforcement of the shell and fabrication of protein-filled microcapsules. (A) CaCO3 template, (B) deposition of SA by SCP, (C) core decomposition with HCl and in-situ crosslinking of SA by Ca21,(D) exposure to protein, (E) deposition of SA by SCP to form inner wall, (F) LbL adsorption of oppositely charged PEs to form the outer wall,(G) PE capsule filled with alginate-immobilized protein after treatment with EDTA, decomposition of the core and cross-linking of inner wall.

2 0 7 8 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

hydrochloric acid, the calcium ions generated in-situ instantlycrosslink the sodium alginate present in the vicinity (by virtueof physisorption and pore diffusion) leading to formation of amicro-gel-like structure.

A small quantity of external cationic dye (like rhodamine6G) that stains the anionic polymeric coating helps in visualiza-tion of the capsular microgel structure using confocal micro-scopy (Fig. 7A). It should be noted that if the microstructuresobtained were exposed to chelating agents (such as EDTA) or ifEDTA was employed for dissolution of the template, nomicrostructures could be seen indicating that the capsular-gelwas essentially made of a calcium-cross-linked alginate matrix.In addition, we confirm the micro-architecture of the systemusing high resolution SFM imaging (Fig. 7B). Investigation ofthe surface texture and thickness of the shell layer of thecollapsed capsules at the nanometric scale reveals a consider-ably thick, pan-cake-like structure for the microgel formed.The height of the dried gel-particles varied from 77 to 116 nmdepending upon the concentration of the sodium alginatesolutions used for SCP coating onto CaCO3 particles (2 and5 mg ml21 sodium alginate respectively). Higher amounts ofalginate contribute to the increased dry volume of the gelparticles due to a considerable amount of precipitating alginategetting trapped in the voids that subsequently undergo gelationwhen divalent calcium ions are released during the core dis-solution process. Exploration of such capsular gels themselveswas difficult due to irregular cross-linking and significant inter-particular adhesion unless additional layers were deposited.Nevertheless, the method forms a basis for a simple core-assisted gel-capsule fabrication.

We assembled additional polyelectrolyte multilayers on topof the alginate coating with multiple objectives: surfacestabilization of the core–shell structures obtained by SCPalone, construction of a defined capsular shell wall and reten-tion of the alginate matrix formed in-situ during subsequenttreatments to derive polymer-filled capsules (refer to section 6for details).

6. Fabrication of polymer- or protein-filledmicrocapsules

The polyelectrolyte microcapsules can be loaded with bio-materials of varying interest. We have successfully filled thecapsule interior with both organic (polymers, proteins, drugsetc.) and inorganic (metal clusters, semiconductor nanocrystalsetc.) materials. Generally, we have followed two principles:active loading (entrapment is achieved during fabrication ofthe microcapsules) and passive loading (encapsulation isachieved after the microcapsules are obtained and the materialof interest is driven into or generated in-situ within the capsules

using physicochemical forces like pH/concentration/ionic/polarity gradient, complexation, chemical reaction etc.). Inthe present manuscript, we shall limit the discussion specificallyto CaCO3-related results.

The profound porosity of the CaCO3 microparticle providesa strong parameter for manoeuvring both active and passiveloading phenomena. The scheme depicted in Fig. 6D–Gillustrates the mechanism of encapsulation. In the SEMimage of Fig. 1 we can observe a rough and porous textureof CaCO3 microparticles—both on the external surface and inthe interior (as radial channels). For protein encapsulationexperiments, we used 5 mm particles which provided increasedsurface area for adsorption/porous diffusion of larger amountsof protein. The driving force for active encapsulation wasessentially physical adsorption, pore diffusion and electrostaticinteraction. We deposited polyelectrolyte multilayers on theadsorbed protein layer with the intention of capturing it. Bovinserum albumin (BSA) was chosen as the model protein. Theevents following core dissolution (in-situ generation of a gel-like matrix) helped in immobilization and retention of theprotein within the microenvironment of the capsule interior.The label on the protein (in our case rhodamine tagged to theprotein) helped in tracing the location of the protein at variousstages of manufacturing (with CLSM imaging). CLSMevidence further showed that protein was partitioned partlyonto the wall (adsorbed fraction) and partly into the voids(diffused fraction) before EDTA treatment. Once the assemblywas exposed to EDTA, the protein was dislodged from the shellwall along with alginate and filled the interior of the capsule.The filling pattern was uniform and almost all capsules resultedin formation of protein-filled capsules eventually after exposureto EDTA (Fig. 8 top). We also observed similar results in theabsence of sodium alginate treatment (without formation of aninner layer by SCP) which could be attributed to penetration ofthe protein during the adsorption phase into the space offeredby the surface of the core and getting trapped within poly-electrolyte multilayers that are assembled later on. However,great care had to be taken not to subject such a system to physicalstresses like vortexing or centrifugation that dislodge theentrapped protein due to the absence of any gel-like structure.

SEM images of the core–shell structures and polymer/protein-filled capsules obtained (before EDTA treatment) aredepicted in Fig. 8 (bottom). There was a ‘‘filling-in’’ of therough and porous surface of the CaCO3 due to diffusion andsequential adsorption of protein, sodium alginate and poly-electrolyte pairs. However, when the templated support waseliminated, the shells retained the memory of the ‘‘filling-inpattern’’. Hence, one could observe considerably thick capsuleshells due to the presence of an inner, cross-linked alginate layerin addition to the contribution from penetrated protein/polymer/polyelectrolyte that occurred invariably during assembly of the

Fig. 7 A – CLSM images of alginate gel microcapsules templated on CaCO3 microparticles. The insert illustrates the staining of these capsules withrhodamine 6G. B – Pancake-like structure of dried microcapsules as revealed by SFM imaging.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1 2 0 7 9

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

shell. Filling of the capsule interior with polymer or proteincould not be captured by SEM studies (as swollen structures),principally because the gel-like structure collapses during thedrying of the sample for SEM measurements.

The measurement of the height of the capsular microgels bySFM provided further evidence for material loaded inside. Pre-formation of an inner polymer layer via SCP resulted in aseveral fold increase in the height of the capsule compared topolyelectrolyte multilayers formed by the LbL technique (e.g.when 2 mg ml21 sodium alginate was used for SCP, the con-tribution to the capsule height was about 75 nm). The biocom-patible polyelectrolyte pairs used (sodium alginate–chitosanand dextran sulfate–chitosan) provided an average height (orthickness) of 10 nm per adsorbed LbL layer. Data derived fromSFM measurements collectively suggest that the system has atotal height of about 160 nm, pointing to the existence of adiffusion barrier of about 80 nm around the gel-immobilizedprotein matrix.

There is yet another possibility of loading polymer-filledcapsules that we have successfully achieved—the passiveloading concept utilizing a pH/polarity/ionic/concentrationgradient. This concept, however, has been proven with adifferent core–shell assembly and polymer. The alginate matrixcan contribute more than one driving element for the low-molecular weight compound to move-in across the poly-electrolyte multilayers to be captured inside. It remains as oneof the experimental objectives for us to achieve with the currentcore–polymer assembly.

Conclusion

The present investigation proves multiple functionalities of aporous calcium carbonate microparticle besides being merelya decomposable template for microcapsule fabrication. The

principle of using the template for cross-linking and reinforcingthe capsule wall assembled by SCP could be improved forvarious applications demanding a soft microstructure formanipulations such as drug loading with a possibility offortifying the system by further coating by the LbL methodwith suitable polymers. The thickness of the PE multilayers canbe tuned with nanometric precision by the LbL technique andprovides a nanoengineered diffusion barrier for encapsulatedsubstances. The double-shell walled microcapsule provides aunique opportunity to encapsulate the substance of interest as ashell constituent or as an immobilized matrix via a dissociableinner shell. The diffusion of the protein captured within the gel-like interior of the microcapsule will be governed by theproperties of the microgel reservoir as well as the physico-chemistry of the capsule wall constituting the diffusion barrier.We are investigating this property and hence a novel method toachieve controlled release by a drug delivery system assembledusing components and the process that ensures the stability ofthe proteins.

Acknowledgements

This research project is supported by the Sofja Kovalevskajaprogram of the Alexander von Humboldt Foundation and theGerman Ministry of Education and Research, Germany.D. V. Volodkin thanks DAAD for support (referat 325,number A/03/01495). Annelise Heilig, Michelle Prevot andDr. Dmitry Shchukin are thanked for SFM and SEM mea-surements respectively.

References

1 C. Thies, A short history of microencapsulation technology, inMicrospheres, microcapsules and liposomes. vol. I: Preparation andchemical application, ed. R. Arshady, Citus books, London, 1999,p. 47.

2 G. Decher and J. D. Hong, Macromol. Chem., Macromol. Symp.,1991, 46, 321.

3 G. B. Sukhorukov, E. Donath, S. Davis, H. Lichtenfeld, F. Caruso,V. I. Popov and H. Mohwald, Polym. Adv. Technol., 1998, 9, 759.

4 A. A. Antipov, G. B. Sukhorukov, E. Donath and H. Mohwald,J. Phys. Chem. B, 2001, 105(12), 2281.

5 H. Ai, S. Jones, M. M. Villiers and Y. M. Lvov, J. ControlledRelease, 2003, 86, 59.

6 V. S. Trubetskoy, A. Loomis, J. E. Hagstrom, V. G. Budker andJ. A. Wolff, Nucleic Acids Res., 1999, 27, 3090.

7 D. V. Volodkin, N. G. Balabushevitch, G. B. Sukhorukov andN. I. Larionova, STP Pharm. Sci., 2003, 13(3), 163.

8 G. B. Sukhorukov, Studies in Interface Science, ed. D. Mobius andR. Miller, Elsevier, Amsterdam, 2001, p. 383.

9 C. Y. Gao, S. Moya, H. Lichtenfeld, A. Casoli, H. Fiedler,E. Donath and H. Mohwald, Macromol. Mater. Eng., 2001,286(6), 355.

10 D. B. Shenoy, A. A. Antipov, G. B. Sukhorukov and H. Mohwald,Biomacromolecules, 2003, 4(2), 265–272.

11 A. A. Antipov, D. Shchukin, Y. Fedutik, A. I. Petrov,G. B. Sukhorukov and H. Mohwald, Colloids Surf. A—Physicochem. Eng. Aspects, 2003, 224, 175.

12 S. Moya, L. Dahne, A. Voigt, S. Leporatti, E. Donath andH. Mohwald, Colloids Surf. A—Physicochem. Eng. Aspects, 2001,183, 27.

13 I. L. Radchenko, G. B. Sukhorukov and H. Mohwald, ColloidsSurf. A, 2000, 202, 127.

14 L. Dahne, S. Leporatti, E. Donath and H. Mohwald, J. Am. Chem.Soc., 2001, 123, 5431.

15 Y. Lvov, A. A. Antipov, A. Mamedov, H. Mohwald andG. B. Sukhorukov, Nanolett., 2001, 1(3), 125.

16 A. Antipov, G. B. Sukhorukov, E. Donath and H. Mohwald,J. Phys. Chem. B, 2001, 105(12), 2281.

17 L. S. Tracy, C. J. P. Francois and H. M. Jennings, J. Cryst.Growth, 1998, 193, 374.

18 N. Koga, Y. Nakagoe and H. Tanaka, Thermochim. Acta, 1998,318, 239.

19 D. Horn and J. Rieger, Angew. Chem., Int. Ed., 2001, 40, 4330.

Fig. 8 Protein-loaded microcapsules (capsule size is about 5 mkm).Top – BSA-labeled with rhodamine is embedded in a gel structure madeby alginate and reinforced by 4 layers of polyelectrolyte. Bottom – SEMimage of the capsules.

2 0 8 0 J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online

20 T. Ogino, T. Suzuki and K. Sawad, Geochim. Cosmochim. Acta,1987, 51, 2757.

21 N. Spanos and P. G. Koutsoukos, J. Cryst. Growth, 1998, 191,783.

22 M. Kitamura, J. Colloid Interface Sci., 2001, 236, 318.23 M. Kitamura, H. Konno, A. Yasui and H. Masuoka, J. Cryst.

Growth, 2002, 236, 323.24 T. Kato, T. Susuki, T. Amamiya, T. Irie, M. Komiyama and

H. Yui, Supramol. Sci., 1998, 5, 411.25 F. Manoli and E. Dalas, J. Cryst. Growth, 2000, 218, 359.26 H. Colfen and L. Qi, Chem. Eur. J., 2001, 7, 106.27 D. V. Volodkin, A. I. Petrov, M. Prevot and G. B. Sukhorukov,

Langmuir, 2004, 20(8), 3398.28 D. V. Volodkin and G. B. Sukhorukov, Biomacromolecules, 2004,

accepted.29 F. Galisteo and W. Norde, Colloids Surf. B: Biointerfaces, 1995, 4,

375.

30 C. A. Haynes and W. Norde, Colloids Surf. B: Biointerfaces, 1994,2, 517.

31 W. Norde, F. Macritchie, G. Nowicka and J. Lyklema, J. ColloidInterface Sci, 1986, 112, 447.

32 E. Donath, G. B. Sukhorukov, F. Caruso, S. A. Davis andH. Mohwald, Angew. Chem., Int. Ed., 1998, 37(16), 2202.

33 C. Y. Gao, E. Donath, H. Mohwald and J. Shen, Angew. Chem.,Int. Ed., 2002, 41(20), 3789.

34 N. G. Balabushevich, O. P. Tiourina, D. V. Volodkin,N. I. Larionova and G. B. Sukhorukov, Biomacromolecules,2003, 4(5), 1191.

35 V. Dudnik, G. B. Sukhorukov, I. L. Radtchenko and H. Mohwald,Macromolecules, 2001, 34(7), 2329.

36 D. V. Volodkin, A. I. Petrov, N. I. Larionova andG. B. Sukhorukov, COST 840 & XI International BRG Workshopon Bioencapsulation. State of Art of Bioencapsulation Science andTechnology, Strasbourg (Illkirch), France, 25–27 May 2003.

J . M a t e r . C h e m . , 2 0 0 4 , 1 4 , 2 0 7 3 – 2 0 8 1 2 0 8 1

Dow

nloa

ded

by S

tanf

ord

Uni

vers

ity o

n 23

May

201

2Pu

blis

hed

on 1

5 Ju

ne 2

004

on h

ttp://

pubs

.rsc

.org

| do

i:10.

1039

/B40

2617

A

View Online