Versatile and Efficient Formation of Colloids of Biopolymer-Based Polyelectrolyte Complexes

Versatile and Efficient Formation of Colloids ofBiopolymer-Based Polyelectrolyte Complexes

Christophe Schatz,†,‡ Alain Domard,‡ Christophe Viton,‡ Christian Pichot,† andThierry Delair*,†

Unite mixte CNRS-BioMerieux, UMR 2142, ENS Lyon, 46, Allee d’Italie, 69364 Lyon Cedex 07 France,and Laboratoire des Materiaux Polymeres et des Biomateriaux, UMR CNRS 5627, ISTIL,

Domaine Scientifique de la Doua, 15 Bd. Latarjet, 69622 Villeurbanne Cedex France

Received April 9, 2004; Revised Manuscript Received May 31, 2004

The formation of colloids based on polyelectrolyte complexes (PECs) of biopolymers was investigated throughthe complexation between two charged polysaccharides, chitosan as polycation, and dextran sulfate aspolyanion. The slow dropwise addition of components, generally used for the formation of PECs, allowedto elaborate both cationic or anionic particles with an excess of chitosan or dextran sulfate, respectively.The PEC particles featured a core/shell structure, the hydrophobic core resulting from the segregation ofcomplexed segments whereas excess component in the outer shell ensured the colloidal stabilization againstfurther coagulation. Considering the host/guest concept for the formation of PECs, the influence of themolecular weight of components on particles sizes could be well explained by the chain length ratios of thetwo polymers. As an irreversible flocculation occurred with a dropwise approach for both cationic andanionic PEC particles when the mixing ratio was close to unity, a more versatile, and simpler to setup,method was designed: the one-shot addition of one solution to the other. Because process of addition isfaster than the flocculation, cationic or anionic particles could be elaborated irrespective of the order ofaddition of the reactant. Characterization of these particles by quasielastic light scattering, electrophoresis,and scanning electron microscopy revealed very similar properties to those obtained by a slow dropwiseapproach. Critical coagulation concentrations of 0.12 and 0.09 M (with sodium chloride) for cationic andanionic particles evidenced a mostly electrostatic stabilization.

1. Introduction

Polyelectrolyte complexes (PECs) are formed in solutionby strong electrostatic interactions between charged micro-domains of at least two oppositely charged polyelectrolytes(PEL). At low ionic strength, the complexation is entropydriven by the release of small counterions, initially boundto the polyelectrolytes.1 According to the chemical composi-tion of PEL, the process of complexation may be cooperativein the sense of secondary interactions, such as hydrogenbonds, can take place and favor the ion-pairing betweenPEL.2 The conditions of formation, the properties, andapplications of PECs were extensively studied but, at thecolloidal level, less works detailed the elaboration of PECs,especially with biopolymers. Such structures were investi-gated with synthetic PEL in the framework of water-solublePECs by the groups of Kabanov3,4 and Tsuchida.5,6 Indeed,complexation between polyelectrolytes having significantlydifferent molecular weights, weak ionic groups and mixingin nonstoichiometric ratios leads to the formation of water-soluble aggregates on a molecular level. These particlesconsist in a long host molecule sequentially complexed withshorter guest polyions of opposite charge, according to a“zip” mechanism. Another way to produce quasisoluble PEC

particles was comprehensively reported by Dautzenberg.1,7-9

Complexation between polyelectrolytes with comparablelarge molecular weight and/or strong ionic groups leads tonearly spherical colloidal structures in conditions of highdilution (<10-4 g/mL) and nonstoichiometric mixing of PELsolutions.

In a recent work,10 we described the formation and thecharacterization of positively charged PEC particles, havingsubmicronic sizes, from the complexation of chitosan, a weakpolybase and dextran sulfate, a strong polyacid under theform of a sodium salt. Chitosan is a (1f4)-linked copolymerof 2-amino-2-deoxy-â-D-glucan (GlcN) and 2-acetamido-2-deoxy-â-D-glucan (GlcNAc; Figure 1) mainly issued fromthe partial N-deacetylation of chitin under alkaline condi-tions.11 The main structural parameter of chitosan is the de-gree of acetylation (DA), i.e., the molar fraction of acetylatedresidues inside the polymer which is responsible for boththe charge density of the polymer and the balance betweenhydrophilic and hydrophobic interactions.12,13In the field ofbiopolymers, chitosan is well-known for its outstandingbiological properties such as biodegradability14 or biocom-patibility15 allowing its use for the development of genedelivery systems.16 Dextran sulfate (DS) (Figure 1) is mainlyknown as a substitute of heparin because of its anti-thrombogenic properties, but it has been also applied in thetreatment of AIDS17 and ulcers. As the polyelectrolytecomplexation did not alter the integrity of the components,

* To whom correspondence should be addressed. E-mail: [email protected].

† CNRS-BioMerieux.‡ UMR CNRS.

1882 Biomacromolecules 2004,5, 1882-1892

10.1021/bm049786+ CCC: $27.50 © 2004 American Chemical SocietyPublished on Web 08/06/2004

chitosan/dextran sulfate PEC particles might find excitingapplications in the field of delivery devices of biologicalmacromolecules such as proteins, peptides, or nucleicacids.18,19

PEC particles were obtained from the dropwise additionof a default amount of dextran sulfate to chitosan in excess,both solutions being adjusted at pH) 4.0. We investigatedthe influence of molecular characteristics of PEL (chainlength, charge density) and environmental parameters (ionicstrength, concentration of PEL) on the sizes and polydis-persity of PEC particles. This study allowed us to demon-strate that chitosan chain length was the main parameterimpacting the sizes of colloidal PECs: higher molar masschitosan leading to the highest particle size. Resultingparticles whose shape, by scanning electron microscopy, wasshown to be broadly spherical were probably constituted bya neutral core surrounded by a chitosan shell ensuring thecolloidal stabilization. Such a structure was evidenced bymeasurements of electrophoretic mobilities revealing that thepositive charge of the particles was decreasing with increas-ing pH, in relation with the neutralization of excess glu-cosamine hydrochloride moieties.

With the aim of developing a general method toward theelaboration of biopolymer-based colloids, positively ornegatively charged, the dropwise addition of default amountsof chitosan to excess dextran sulfate was investigated in thisstudy. But, for both cationic and anionic particles, the samelimitation was encountered, namely the irreversible floc-culation observed close to neutrality. This flocculation couldnever be avoided even by adding a large excess of titrant,probably because of both the high difference of intrinsic pKbetween PEL (∆pK0 ∼ 4.5) and the ability of polysaccharidesto associate via numerous hydrogen bonds. This promptedus in this contribution to design an alternate approach,versatile enough to allow the synthesis of both negativelyor positively charged colloids, irrespective of the order ofaddition of reactants.

2. Experimental Section

2.1. Materials. Two weakly acetylated chitosans wereused in this study: one, with a medium molecular weight(Aber technologies, batch A32E03, DA∼1%,Mw ∼ 150 000g mol-1) and the other with a relatively high molecularweight (France Chitine, batch 114, DA∼5%,Mw ∼ 350 000g mol-1). To obtain a high purity material, chitosans weresuccessively dissolved at 1% (w/v) in a 0.1 M acetic acid

solution, filtered trough membranes (Millipore) having adecreasing porosity from 3 to 0.22µm, reprecipitated withdilute ammonia, extensively rinsed with deionized water untilneutral pH, and finally freeze-dried. Purified chitosans wereN-acetylated at DA) 15% in homogeneous media withacetic anhydride. The reaction was performed in a hydro-alcoholic mixture according to the procedure previouslydescribed by Vachoud et al.20 After N-acetylation, chitosanswere neutralized, rinsed with deionized water, and thenfreeze-dried. Finally, chitosan with a medium molecularweight was hydrolyzed with sodium nitrite.21-22 Then,chitosan was dissolved at 0.5% (w/v) in a 0.2 M acetic acid/0.1 M sodium acetate buffer. A 0.04 M sodium nitritesolution was added to chitosan solutions to obtain a nitrite/glucosamine units molar ratio of 0.05. The reaction wasperformed under moderate magnetic stirring during 15 h.Hydrolyzed chitosan was recovered by precipitation with adilute ammonia solution, purified by several washings withdeionized water until neutrality and lyophilized.

The degree of acetylation of chitosans was determined by1H NMR spectroscopy (Varian, 500 MHz), according to themethod developed by Hirai et al.23 Hydrolysis did not affectthe acetyl content of chitosan since the same DA was foundbefore and after reaction with the sodium nitrite solution.The weight-average molecular weight (Mw), the z-averageroot-mean-square of the gyration radius (RG,z), and thepolydispersity index (Ip) were measured by gel permeationchromatography (3000 and 6000 PW TSK gel columns, ID) 7.8 mm andL ) 300 mm) coupled on line with adifferential refractometer (Waters 410) and a multi-angle-laser-light-scattering (MALLS, Wyatt, Dawn DSP) equippedwith a 5mW He/Ne laser operating atλ ) 632.8 nm. Adegassed 0.2 M acetic acid/0.15 M ammonium acetate buffer(pH ) 4.5) was used as eluent and solvent. Light intensitymeasurements were derived from the classical Rayleigh-Debye equation24 considering a refractive index incrementof 0.189 mL/g.13 The water content was determined bythermo-gravimetric analysis (DuPont Instrument 2950).

Dextran sulfate having∼5 × 103 (DS 5k),∼104 (DS 10k),and∼5 × 105 (DS 500k) g mol-1 weight average molecularweights were provided by Sigma and used without furtherpurification. The molecular characteristics of the highmolecular weight dextran sulfate (DS 500k) were determinedby gel permeation chromatography using the same systemas described above and a PL aquagel-OH mixed column (ID) 7.8 mm andL ) 300 mm). A 0.1 M NaNO3 solutionadjusted at pH) 7.0 was used as solvent and eluent. In these

Figure 1. Chemical structures of the polysaccharides used in the study. DA (%) refers to the degree of acetylation of chitosan.

Colloids of Polyelectrolyte Complexes Biomacromolecules, Vol. 5, No. 5, 2004 1883

conditions, a molecular weight of 1.5× 106 g/mol was foundfor the DS 500k (see Table 1). The degree of sulfation, i.e.,the number of sulfate functions per glucoside units, wasdetermined for each sample by colloidal titration usingtoluidine blue as colored pigment10 and a UV/vis spectrom-eter (µQuant, Bio-Tek instrument). The following degreesof sulfation were obtained: 2.2( 0.2 (DS 5k), 2.4( 0.3(DS 10k), and 2.2( 0.2 (DS 500k). Water content wasprovided by the manufacturer.

2.2. Methods.2.2.1. Preparation of Polyelectrolyte Solu-tions.According to the water content, chitosan was dispersedin deionized water at various concentrations. The dissolutionwas achieved by adding a stoichiometric amount of 0.1 Mhydrochloric acid with respect to the free amino functions.Sodium chloride was added to vary the ionic strength (µ) ofthe solutions. Then, chitosan solutions were adjusted at pH) 4.0 with 0.1 M sodium hydroxide and the added amountwas considered to calculate the final concentration and ionicstrength. Dextran sulfate solutions, at different concentrations,were prepared directly in deionized water. Sodium chloridewas added to obtain the required ionic strength and pH wasadjusted to 4.0 with 0.1 M hydrochloric acid. Before use,all solutions were filtered through 0.22µm pore sizeMillipore membranes.

2.2.2. Polyelectrolyte Complexes Formation.Polyelectro-lyte complexations were performed at room temperatureusing either chitosan or dextran sulfate as the startingsolution. Two processes of mixing were investigated. First,polyelectrolyte solutions were added dropwise at a flow rateof 20 mL/hr to oppositely charged polyelectrolyte solutionsof same ionic strength. At defined molar mixing ratios (n+/n-) between cationic and anionic charged units, aliquots of2 mL were collected and left to stand 12 h before furthercharacterizations. Chitosan and dextran sulfate concentrationsin the mixture were recalculated to deduce the requiredamount of titrant solution to achieve the nextn+/n- ratio.Second, complexations were performed for variousn+/n-

ratios by a fast one-shot addition of a polyelectrolyte solutionto a solution of oppositely charged polyelectrolyte at identicalionic strength, the total volume of polyelectrolyte mixturevarying between 5 and 30 mL and each experiment being atleast repeated twice. A constant magnetic stirring of 500 rpmwas maintained during all titrations.

2.2.3. Particles Purification.Two molar mixing ratios,n+/n- ) 2 and 0.2, were considered for subsequent size andmobility measurements as a function of pH or ionic strength,

microscopic observation, and colloidal stability. To avoidphysi-sorption of free polymer onto the surface of theparticles, these ones were separated from the polyelectrolytemixture by centrifugation for 30 min at 12000 rpm. Thesupernatant was discarded, the sediment re-suspended indeionized water, and the centrifugation repeated. Finally theproduct was re-suspended in a minimum volume of deionizedwater.

2.2.4. Colloidal Characterization. Dynamic Light Scat-tering. Dynamic light scattering measurements of polyelec-trolyte complexe dispersions were carried out using a CoulterN4MD equipped with a 4 mWHe/Ne laser operating atλ )632.8 nm. All measurements were performed at a fixed angleof 90° at 25.0( 0.2 °C. Hydrodynamic diameter (Dh) andpolydispersity index (PI) were derived from a cumulantanalysis.25 Three to five measurements of 120s were per-formed for each sample. For the clarity of the presentation,bars were not plotted on each graph but from the reproduc-ibility of the measurements, one could consider relative errorsof ∼5% and∼10% for size and polydispersity index.

Turbidimetry. The turbidity of the complex dispersions wascharacterized by the optical density (OD) measured atλ )500 nm with a µQuant UV/VIS spectrometer (Bio-Tekinstrument). At this wavelength, the polyelectrolytes do notabsorb. Deionized water was used to establish the baseline.Optical density was expressed as the average of at least twomeasurements.

Electrophoretic Mobility.The electrophoretic mobility (µE)of the particles was determined at 25°C with a MalvernZetasizer HS3000 equipped with a 10mW He/Ne laser beamoperating atλ ) 633 nm.µE was expressed as the averageof 15 measurements with a relative error of∼5%.

Scanning Electron Microscopy. Surface morphology andshape of chitosan-dextran sulfate particles were examinedusing scanning electron microscopy (Hitachi S800) operatingat 15 eV. A droplet of 0.1% (w/v) particles dispersion wasdeposited on a sample holder, air-dried at room temperature(24 h), and coated with gold-palladium in a cathodeevaporator (Technics Hummer II) under an argon atmo-sphere.

2.2.5. Colloidal Stability.The colloidal stability of PECparticles was investigated as a function of the salt concentra-tion in the medium. The aggregation process was monitoredat 25°C as a function of time by means of a turbidimetricmethod (Jasco V-550, UV-visible spectrophotometer).Particles dispersions were highly diluted in 10-2 M acetic

Table 1. Molecular Characteristics of Chitosans and Dextran Sulfates Determined by Gel Permeation Chromatography Coupled with StaticLight Scattering Either in a 0.2 M Acetic Acid/0.15 M Ammonium Acetate Buffer (pH ) 4.5) for Chitosan or in a 0.1 M NaNO3 Solution (pH) 7.0) for Dextran Sulfate

Mw 105 (g/mol) DPwa Ipb RG,z (nm) Lp (nm)c b (nm)d Lc (nm)e

Chitosan (Degree of Acetylation ) 15%)0.160 ( 0.008 96 ( 5 1.21 ( 0.06 f f 0.61 48 ( 33.65 ( 0.09 2182 ( 54 1.41 ( 0.05 91.2 ( 2.4 8.6 ( 0.8 0.61 1100 ( 28

Dextran Sulfate (Degree of Sulfation ) 2.2-2.4)∼0.05 ∼13 ND ND ND 0.22 ∼6∼0.1 ∼26 ND ND ND 0.20 ∼1215.1 ( 2.0 3910 ( 520 1.52 ( 0.09 72.5 ( 3.0 1.6 ( 0.5 0.22 1920 ( 250

a Weight-average degree of polymerization. b Polydispersity index. c Persistence length (see refs 10,13). d Average distance between charges centers.e Contour length. f Molecular weights are too low for determination of gyration radii and persistence lengths by static light scattering.

1884 Biomacromolecules, Vol. 5, No. 5, 2004 Schatz et al.

acid neutralized at pH) 4.0. Aggregation rate constants(dOD/dt) were deduced by measuring the optical densityvariation atλ ) 500 nm as a function of time, after a rapidaddition of NaCl solution in the optical cell. The stabilityfactor (W) at a given NaCl concentration was determined asfollows:

where (dOD/dt)f and (dOD/dt)s are the initial slopes of theoptical density variation versus time for fast and slowaggregation processes, respectively. The critical coagulationconcentration (CCC) of particles dispersion was deducedfrom the stability factor variation as a function of the salinityplotted in log-log scale.

3. Results and Discussion

Before investigating the formation of polyelectrolytecomplexes, molecular characteristics and conformation ofinitial components were determined by gel permeationchromatography coupled on line with a multiangle laser-light-scattering (Table 1). The conformation in solution ofchitosan (DA) 15%) assessed by the persistence lengthcorresponds to a wormlike model (Lp ) 8.6 nm for highmolecular weight chitosan) mainly due to the intrinsic rigidityof the â, (1f4) glycosidic bonds and the possibility ofintramolecular hydrogen bonds.26 In contrast, dextran sulfate(DS) exhibited a rather low persistence length (Lp ) 1.6 nm)with regard to its high charge density. This relatively highflexibility could be due (i) to a high counterion condensationevidenced by a high value of the Manning parameter (êM )3.3) which reduces the apparent charge density, (ii) low levelof intramolecular hydrogen bonds, since only an average of0.8 hydroxyl group is available per sugar residue, and (iii)R, (1f6) linked sugar residues with a branched structurefor high molecular weight dextran sulfates.27

3.1. Formation of Particles by Dropwise Addition ofPolyelectrolytes.In a previous work, we evidenced that thechain length of initial components is a key parameter on thePEC particle size.10 To gain deep insight into the role ofthis parameter on the formation and properties of PEC,complexation between low and high molecular weightchitosans with three different molecular weight dextransulfates was investigated. The results are reported in termsof size evaluated by the hydrodynamic diameter, polydis-persity index and optical density of colloidal dispersions(Figure 2).

The formation of colloidal polyelectrolyte complexesbetween chitosan and dextran sulfate was studied as afunction of the mixing ration+/n-, i.e., the molar ratiobetween cationic and anionic units. Complexation wasperformed both by the dropwise addition of a default dextransulfate solution to an excess chitosan leading to the formationof cationic particles (n+/n- > 1) or by the reverse addition,a default chitosan to an excess dextran sulfate leading to theformation of anionic particles (n+/n- < 1). To ensure a highreproducibility of QELS measurements, a single solution of

excess component was used in each experiment and succes-sive dropwise additions of default polymer were performedto get the required mixing ratios. The common feature ofthe two complexation processes mentioned above was theincrease of turbidity assessed by the optical density atλ )500 nm as the concentration of titrant increased (Figure 2).This evidenced the formation of increasing amounts ofinsoluble polyelectrolyte complexes since neither chitosannor DS absorbed light atλ ) 500 nm. Until the flocculationof the system characterized by a drop of the optical density,the colloidal dispersions were stable, indicating that elec-trostatic or electrosteric stabilization prevented further co-agulation. This stabilization might be ensured by an excessof binding of the major component, likely to form astabilizing shell around the particles (see below).

In this work, the degree of acetylation of chitosan wasrestricted to 15% as we demonstrated that a low acetylcontent favored the formation of less swollen particles.10

Charge density of dextran sulfate was also kept at a constantvalue corresponding to 2.2-2.4 sulfate functions per glu-cosidic unit. In all experiments, the initial solutions ofpolyelectrolytes were adjusted at pH) 4.0 to ensure the fullprotonation of chitosan,12 DS being also fully dissociated atthis pH. In addition, using solutions having the same pHalso allowed us to avoid any neutralization reactions. Finally,initial concentrations of chitosan and dextran sulfate wereset at 10-3 and 5× 10-4 g/mL as we demonstrated that theseconditions allowed the formation of small PEC particles.10

3.1.1. Formation of Cationic Particles.Cationic particleswere formed by the addition of increasing amounts of dextransulfate to an excess chitosan (decreasingn+/n- ratios from20 to ∼1). Figure 2 evidenced an antagonist effect of themolecular weight of chitosan and DS on the PEC particleproperties. Indeed, high molecular weight chitosan led to thehighest particle sizes and optical densities whatever DSmolecular weights and mixing ratios. In contrast, the higherthe molecular weight of DS, the smaller the particle sizesand the optical densities. Thus, smallest particles are obtainedby the complexation of chitosan havingMw ) 1.6 × 104

g/mol with DS havingMw ) 1.5 × 106 g/mol. Anothercharacteristic of the complexation process is the decrease ofhydrodynamic diameters during the titration until a secondaryaggregation was obtained preceding the irreversible floc-culation. It means that the addition of dextran sulfate allowedboth the formation of new particles, evidenced by theincrease of the optical density and their rearrangementstoward more compact structures. The range of the particlesize reduction decreased also with increasing dextran sulfatemolecular weight, meaning that colloids formed with highmolecular weight polyanions were denser than those obtainedwith the lowest molar mass. The polydispersity index (PI)of the colloidal dispersions displayed about the same behavioras the hydrodynamic diameter toward the effect of themolecular weight of components. However, its variation asn+/n- decreased was more limited in range. Finally, it isworth noting that very high values of PI, obtained with highmolecular weight chitosans, cannot be interpreted since above0.3 PI loses its physical meaning.

W )(dOD/dt)f

(dOD/dt)s


On the basis of these results, one can consider that theparticle forming process relies on the formation of hydro-phobic segments on polymer chains by charge neutralizationand segregation of these segments into particles stabilizedby free unpaired charges. Thus, for a set chitosan molar mass,the size of the titrant component impacts on the dimensionof the objects in such a way that ion-pairing is optimum forcharge neutralization. Hence, the longer the segment withion pairs, the stronger the interactions between these seg-ments and the smaller the particles. On this assumption, twodifferent situations may be considered. First, for low mo-lecular weight polyanions (Mw ∼ 5 × 103 and∼104 g/mol),chitosan chains in excess amount can act as host polycations(see the respective contour lengths in Table 1). Then, thecomplexes formed bear many unpaired positive charges stillavailable for complexation. Being more hydrophilic, uncom-plexed segments of chitosan constitute the corona of theparticles (Figure 3a). Thus, the higher the chitosan chainlength, the thicker the corona and the higher the hydrody-namic volume of particles. However, as the chitosan chains

exhibit a relative stiff conformation, one cannot rule out thepossibility that the initial larger coil volume for the highermolecular weight chitosan chains contributes to the formationof more swollen core and hence to high particle sizes.Addition of more polyanion led to the observed size reductionby consumption of the residual charges with concomitantcollapse of the newly formed hydrophobic segments ontothe hydrophobic core of the objects. Second, when using highmolecular weight DS (Mw ) 1.5× 106 g/mol), the polyanioncan function as a host molecule capable of accommodatingseveral polycation chains. Then, formed hydrophobic seg-ments are longer and can strongly interact with one anotherto yield more compact objects (Figure 3b). This requires thatnot all charges of chitosan be neutralized, otherwise floc-culation can occur as we will see later for anionic particles.

Finally, it is worth noting that the values of mixing ratioat the flocculation point are close to the unity whatever themolecular weight of initial components evidencing thatalmost all the charges of the polyions were complexed.However, the high mismatching between charge distance of

Figure 2. Sizes, polydispersity indexes, and optical densities of PEC particles obtained by slow addition of a DS solution in a chitosan solution(n+/n- > 1) and by slow addition of a chitosan solution in a DS solution (n+/n- < 1) for (a) low molecular weight chitosan (Mw ) 1.6 × 104

g/mol), (b) high molecular weight chitosan (Mw ) 3.65 × 105 g/mol) and three molecular weight dextran sulfates: (0) Mw ) 5 × 103 g mol-1,(2) Mw ) 104 g mol-1, (4) Mw ) 1.5 × 106 g mol-1. Conditions: pH ) 4.0, µ ) 5 × 10-2 M, flow rate ) 20 mL/h, [chitosan]0 ) 10-3 g/mL, [DS]0) 5 × 10-4 g/mL.


chitosan and dextran sulfate (Table 1) hinders a regular 1:1arrangement of the polyions required for a strict ladder-likePEC structure. It means that each DS unit, bearing 2.2-2.4sulfate functions, should be able to complex more than oneglucosamine moiety and thus act as an ionic cross-linker.

3.1.2. Formation of Anionic Particles.Anionic particleswere obtained by addition of increasing amounts of chitosanto an excess dextran sulfate (increasingn+/n- ratios from0.1 to ∼1). As for the cationic particles, the formation ofcomplexes mainly depended on the size of the polymer inexcess. Here, the most surprising phenomena was obtainedwith a DS of 5× 103 g/mol. Indeed, when a solution ofchitosan having low or high molecular weight was addedinto an excess solution of DS 5k to getn+/n- ) 0.1, themixture was turbid during less than 5 min and thenflocculated. With a DS of 104 g/mol, colloids were formedin a rather small range of molar mixing ratios, flocculationoccurring atn+/n- ) 0.5 and 0.3 for low and high molecularweight chitosan, respectively (Figure 2). The poor colloidalstability was confirmed by the irreversible aggregationencountered during the centrifugation of these particles.Using a DS of 1.5× 106 g/mol, the domain of formation ofcolloids increased atn+/n- ) 0.6-0.7 according to thechitosan molecular weight (Figure 2). These results evidencedthat the key parameter controlling the formation and thestability of anionic particles was the chain size ratio betweenchitosan and DS. Indeed, two different situations can be

identified depending on whether chitosan may act as hostmolecules or not.

First, low colloidal stability occurred as long as dextransulfate remained smaller than the titrant, in other words whilethe titrant (chitosan,Mw ) 1.6 × 104 or 3.65× 105 g/mol)could host several molecules of polyanion (dextran sulfate,Mw ∼ 5 × 103 or 104 g/mol). Reasons for the prematureflocculation observed with a DS 5k and the poor colloidalstability with DS 10k may be the followings. Dextran sulfatebeing highly flexible, more charged than chitosan and usedin excess for the formation of anionic colloids, smallmolecules of DS 5k would favor a complete ion-pairing withchitosan leading to quasi-neutral complexes without enoughunpaired sulfate functions to ensure the stabilization ofparticles (Figure 3c). Hence, formed complexes quicklyaggregated. On increasing the molecular weight of dextransulfate to 104 g/mol, conformational adaptation of DS for acomplete-ion pairing was less favorable and let enoughunpaired sulfate groups participate in the colloidal stabiliza-tion. The very short decrease of the mean diameter at theonset of the dextran sulfate titration by low molecular weightchitosan (Figure 2a), illustrating a limited compaction uponaddition of chitosan, confirms the presence of some freesegments of DS in PEC particles, probably located at theirsurface. Such a decrease of size was not observed with highmolecular weight chitosan (Figure 2b), complexation beingearlier limited by the flocculation. Hence 104 g/mol is a

Figure 3. Formation of colloidal polyelectrolytes complexes based on chitosan (light gray) and dextran sulfate (black) through the hydrophobicsegregation of complexed segments. Four limit cases are considered according to the charge of the colloid and the molecular weight of components(schemes are not to scale). (a), (c) high Mw chitosan versus low Mw dextran sulfate; (b), (d) low Mw chitosan versus high Mw dextran sulfate.


threshold molar mass for dextran sulfate to elaborate anionicPEC particles with the two chitosans used in this study.Conversely, in the reverse process, the complexation of anexcess of low molecular weight chitosan by a high molecularweight dextran sulfate, highly stabilized PEC particles wereobtained. This can be explained by considering that chitosanis weakly charged and relatively stiff. Hence, chargeneutralization cannot be totally achieved and some glu-cosamine residues remained unpaired ensuring electrostaticstabilization (Figure 3b).

Second, when DS could act as a host molecule, i.e., forthe highest molecular weight (Mw ) 1.5 × 106 g/mol),chitosan counterparts formed “hydrophobic patches” oncomplexation with the large DS molecules. Theses patchessegregated to form particles consisting in a hydrophobic coreand a large hydrophilic shell of uncomplexed dextran sulfatesegments (Figure 3.d), leading to larger diameters of particlesthan those observed with DS 10k (Figure 2). Then, a furtheraddition of chitosan led to a reduction of both size andpolydispersity index, by complexing with the negativelycharged shell, as previously observed for cationic particlesobtained from high molar mass chitosan. Figure 2 alsodisplays that anionic colloids of similar sizes were obtainedby complexing high molar mass DS with the two chitosans,evidencing that the hydrodynamic diameter of these PECparticles was indeed determined by the thickness of the largedextran sulfate outer shell.

Finally, whether chitosan or dextran sulfate were used asstarting solutions, flocculation at an+/n- ratio close to 1was irreversible and the addition of excess titrant never ledto the recovery of colloids with a reversed surface charge.

3.2. Formation of Particles by the One-Shot Additionof Polyelectrolytes. Since irreversible flocculation wasobserved when getting slowly close to electrical neutrality,the idea was to go beyond this critical value as quickly aspossible in order to achieve a fast charge inversion, whichwould allow maintaining the colloidal stability. Hence, thetitrant solution was added as quickly as possible to thestarting solution. This new way to proceed was tested toelaborate both anionic and cationic colloids by adding excessdextran sulfate solution to default chitosan or excess chitosanto default dextran sulfate. The influence of the order ofaddition of reactants on the formation of colloids was alsoinvestigated by adding the polyelectrolyte in default to excessof oppositely charged PEL as for the dropwise addition.

The mean diameters of particles obtained by these twoprocesses were determined for the complexation of a lowmolecular weight chitosan with three DS (Figure 4). What-ever the nature of the polyelectrolyte used as starting solutionand the order of addition (excess to default or the opposite)colloidal PECs were obtained at variousn+/n- ratios, usingthe same polyelectrolyte concentration range as before. Moresurprisingly, the shape of the curves of the variations of thehydrodynamic diameter with varyingn+/n- ratios for the one-shot addition looked quite similar to the ones obtained bythe slow addition process. This suggests high kinetics forthe particle formation process for both, one-shot and slowadditions. However, the one-shot addition of the titrant seemsto be faster than the reaction, otherwise flocculation would

have impeded the formation of positive colloids when addingexcess chitosan to dextran sulfate (or inversely for anioniccolloids) since oppositely charged components are in sto-ichiometric amount for a brief time during the mixing.

The effect of dextran sulfate molecular weight on theparticle size was very similar for both slow and fastprocesses. It was impossible to obtain stable anionic colloidswith the lowest molecular weight dextran sulfate (DS 5k).Using DS 10k, the size variations of anionic particles withincreasingn+/n- ratios, display a regular increase in diam-eters, suggesting that the aggregation mode was slightly morepredominant in this process than in the dropwise one.Moreover, experimentally, these particles have a very poorcolloidal stability since we observed their aggregation duringthe centrifugation as for those obtained by dropwise addition.Again, it was with DS of the highest molar mass that anionicPECs could be obtained over the broadest range ofn+/n-

ratios. Similarly to the dropwise method, a “U” shaped curvewas obtained for the variations of the particle size withincreasingn+/n- ratios from 0.1 to 0.7 (Figures 4c and 2a),

Figure 4. Sizes of PEC particles obtained by a one-shot addition ofa DS solution in chitosan solutions (2) and by a one-shot addition ofa chitosan solution in DS solutions (4) for various weight-averagemolecular weights of DS: (a) Mw ) 5 × 103 g mol-1 (b) Mw ) 104 gmol-1 (c) Mw ) 1.5 × 106 g mol- 1. Conditions: pH ) 4.0, µ ) 5 ×10-2 M, [chitosan]0 ) 10-3 g/mL, [DS]0 ) 5 × 10-4 g/mL.


confirming that anionic particles were surrounded by a largecorona of uncomplexed DS whose thickness decreased whenthe concentration of chitosan increased.

As a general feature, the average particle sizes obtainedby the one-shot method were slightly smaller than by thedropwise process (Figure 2) and so were the polydispersityindexes (inferior to 0.2 in identical experimental conditions).Interestingly, the mean diameter was more elevated whenthe excess was added to the default (which can also beobserved in Figures 5 and 6) for both positively andnegatively charged particles. Indeed, these particles areformed by passing through the flocculation point, causing aslight increase of their degree of aggregation and thus oftheir size.

3.2.1. Influence of External Parameters on the ParticleFormation Process.To investigate the influence of the poly-mer concentration and of the mixing order, a dextran sulfateof 1.5 × 106 g/mol was selected as this polymer led to themost stable colloids forn+/n- ratios of 2, for cationic par-ticles, and 0.2, for anionic particles. Figures 5 and 6, as inFigure 4, confirm that the addition of the excess polymer tothe default led to colloids with larger diameters and a slightincrease in polydispersity indexes for the cationic particlesonly.

Dextran sulfate initial concentration could be increasedat least up to 5× 10-3 g/mL without affecting thestabilization of the colloids. The increase in particle size,polydispersity indexes and optical density was sharp onreaching this critical concentration, whatever the mixingorder and then+/n- ratios (Figure 5), meaning that a

relatively high concentration of polyelectrolyte favored thedegree of aggregation of PEC particles as demonstrated forsynthetic polyelectrolytes.28,29The size increase with dextransulfate concentration had been observed previously for thedropwise approach.10

If an increase of the chitosan concentration led to higheroptical densities as for the dextran sulfate (Figure 6), its effecton the particle size was quite limited in range as alreadynoticed for the slow addition process, for which the oppositetendency was observed, namely a reduction in particle size.10

This evidences that a higher content of chitosan favored theformation of more particles without affecting significantlytheir sizes. The lower reactivity of chitosan in comparisonto the highly charged dextran sulfate may be responsible ofthis tendency but further investigations are needed to obtaina better understanding of this effect.

3.2.2. Colloidal Characterization. Scanning ElectronMicroscopy. Two types of particles based on low molecularweight chitosan and high molecular weight DS and obtainedby the one-shot addition of default component to excess onewere analyzed by scanning electron microscopy. Micropho-tographs show that particles have almost a spherical shape(Figure 7). It is worth noting that these pictures were obtainedafter elimination of the free excess polymer in the mixtureby two successive centrifugations followed by re-dispersionin water. On crude particles no clear data could be obtainedas excess polymer was bound onto the colloid surface,leading to an overestimation of the particle size.10 At n+/n-

ratios of 2 or 0.2, parts a and b of Figure 7 respectively,particles look fairly identical and confirm the values of

Figure 5. Influence of dextran sulfate (Mw ) 1.5 × 106 g mol-1) concentration and mixing order on the properties of PEC particles obtained bya one-shot addition with chitosan (DA ) 15%, Mw ) 1.6 × 104 g mol-1) for n+/n- ) 2 (a) and n+/n- ) 0.2 (b). Conditions: pH ) 4.0, µ ) 5 ×10-2 M, [chitosan]0 ) 10-3 g/mL, total volume of solutions (after mixing) ) 5 mL. (light gray) [DS]0 ) 5 × 10-4 g/mL, (dark gray) [DS]0 ) 10-3

g/mL, (black) [DS]0 ) 5 × 10-3 g/mL.


hydrodynamic diameters obtained by QELS, though thenegatively charged one appeared slightly smaller than by lightscattering. This might be due to the drying step of the samplepreparation, inducing a collapse of the large and mobile outershell of dextran sulfate. Whether the particles have beenobtained by the dropwise or the one-shot method, identicalmorphologies have been observed (data reported in ref 10).

Colloidal Stability Versus pH. The variations of thehydrodynamic diameters and electrophoretic mobilities withpH have been reported in Figure 8 for the same PEC particlesas those studied by SEM. The experiments for which theincrease of pH was obtained by acetic acid neutralizationwith sodium hydroxide do not have a constant ionic strengthall along the investigated pH range. Hence, another set ofdata was recorded with a controlled ionic strength set at 10-2

M by the addition of sodium chloride.The decrease in size observed for the cationic particles

(Figure 8a) can be interpreted as resulting from the neutral-ization of the protonated glucosamine units leading to acollapse of chitosan chains in the particle corona. Theprogressive neutralization of chitosan was confirmed by theconcomitant decrease of electrophoretic mobilities (Figure8b). For pH values higher than 5.5, the effective over-chargeof the particles was reduced to such an extent that secondaryaggregation occurred, impeding further measurements.

With anionic particles, a decrease in particle size wasobserved until pH 5 (Figure 8a) whereas electrophoreticmobilities remained almost constant in all the investigatedpH range (Figure 8b), resulting from the presence of sulfategroups at the surface, always protonated in the pH rangeinvestigated. Relatively high values of hydrodynamic diam-eters in comparison to those obtained during the complex-ation (Figure 4) have to be related to the low ionic strength

of the media (see below). The decrease in particle sizebetween pH 3.5 and 5 could result from a loss of hydrationof the negatively charged shell due to a decrease in the H3O+

ion concentration. Beyond pH 8, colloidal characterizationcould not be performed as a drastic decrease in opticaldensity due to the dissolution of the complex was observed.Indeed, most of the protonated glucosamine units wereneutralized at this pH,30 inducing a decomposition of thePECs into the two hydrophilic macromolecules. As particleswere highly diluted, precipitation of chitosan was notobserved.

Colloidal StabilityVersus Ionic Strength.The investigationof the effects of ionic strength on particle diameters andelectrophoretic mobilities was undertaken with PEC basedon low molar mass chitosan and high molar mass DS. InFigure 9a, a sharp decrease in particle size was observedfor anionic particles with increasing salt concentrations,conversely to cationic ones. This size reduction was probablydue to a collapse of the large DS-rich corona as a result ofthe charge screening by increasing the electrolyte concentra-tion. Above 0.075 M in salt, the colloidal stability was lost,and irreversible aggregation was observed. Cationic particlesdisplayed a very limited size reduction with increasing saltconcentration and aggregation occurred at 0.1 M NaCl. Asalready observed in Figure 2a, the size of the chitosan-richouter shell was smaller than for dextran sulfate coronas.Anyway, for both types of particles, the stabilization mech-anism seemed mainly to be of electrostatic type. The var-iations of electrophoretic mobilities with ionic strength werereported in Figure 9b. For cationic particles, the mobilitieslogically decreased with increasing salt content. For anionicparticles, the mobilities remained almost constant as charge

Figure 6. Influence of chitosan (DA ) 15%, Mw ) 1.6 × 104 g mol-1) concentration and mixing order on the properties of PEC particlesobtained by a one-shot addition with dextran sulfate (Mw ) 1.5 × 106 g mol-1) for n+/n- ) 2 (a) and n+/n- ) 0.2 (b). Conditions: pH ) 4.0, µ) 5 × 10-2 M, [DS]0 ) 5 × 10-4 g/mL, total volume of solutions (after mixing) ) 5 mL. (white) [CS]0 ) 5 × 10-4 g/mL, (light gray) [CS]0 ) 10-3

g/mL, (dark gray) [CS]0 ) 5 × 10-3 g/mL, (black) [CS]0 ) 10-2 g/mL.


shielding could be compensated by the increase in chargedensity due to the reduction in particle size.

Colloidal stability of PEC particles was also assessed fromthe critical coagulation concentration (CCC), i.e., the con-centration of salt at which the particles are no longer stableand begin to aggregate. CCC was determined by plottingthe stability ratio (W) as a function of ionic strength using alog-log scale (Figure 10). Physically, 1/W, corresponds tothe fraction of particle-particle collision resulting in asticking event.31 Intersection of the negative slope line withthe horizontal one, corresponding toW ) 1 (for high ionicstrength), provided the values of CCC which were found tobe close to 0.12 and 0.09 M for cationic and anionic particles,respectively. These relatively low values, in good agreement

with previous size variations (Figure 9a), are in accordancewith a mostly electrostatic stabilization for both types ofparticles.

Finally, on a process standpoint, particles were easily andreproducibly produced on a larger scale, up to 40 mL ofsolutions for an+/n- ratio of 2 (no higher volume wasactually tested) and 30 mL for the 0.2n+/n- ratio, after whichparticle size variations could reach 20-30%. In fact, sizesof anionic PEC particles are less accurate with regard to theirlarge corona of dextran sulfate. Particles have been concen-trated to 3% and 0.7% solids for respectively the cationicand anionic ones, lower solid content for anionic particlesbeing due to their more swollen structure. After four monthsof storage at room temperature, with no or gentle stirring,no alteration of the particle mean size was observed.

4. Conclusion

The influence of the molecular weight of components onthe formation of chitosan/dextran sulfate PEC particlesobtained by a dropwise addition was investigated for bothcationic and anionic colloids. First, when the component inexcess could act as host and accommodate several moleculesof the titrant, the segregation of hydrophobic segments ledto particles electrostatically stabilized by a large hydrophiliccorona of excess component. Second, when the componentin excess was smaller than the titrant, the outcome of theprocess depended on the capacity of the excess polymer at

Figure 7. Scanning electron microphotographs of chitosan (Mw )1.6 × 104 g/mol)/dextran sulfate (Mw ) 1.5 × 106 g/mol) PEC particles,obtained by a one-shot addition of components, for two mixing ratios,(a) n+/n- ) 2 (b) n+/n- ) 0.2. Conditions: [chitosan]0 ) 5 × 10-3

g/mL, [DS]0 ) 5 × 10-4 g/mL, pH ) 4.0, µ ) 5 × 10-2 M. Particleswere twice washed before analysis.

Figure 8. Hydrodynamic diameters (a) and electrophoretic mobilities(b) of chitosan/dextran sulfate PEC particles (see experimentalconditions in Figure 7) as a function of pH in various media: (0) n+/n- ) 2 in neutralized 10-2 M acetic acid solutions. (9) n+/n- ) 2 inneutralized 10-2 M acetic acid solutions adjusted at µ ) 10-2 M withNaCl. (2) n+/n- ) 0.2 in neutralized 10-2 M acetic acid solutionsadjusted at µ ) 10-2 M with NaCl. (4) n+/n- ) 0.2 in 10-2 Mphosphate buffer.


performing charge neutralization. When charge neutralizationcould be optimum, flocculation occurred for low amountsof added titrant (case of chitosan into DS 5k or DS 10k inexcess). When charge neutralization could not be optimum,the formation of dense particles, stabilized by unpairedgroups was observed (case of DS 500k added to lowmolecular weight chitosan).

The one-shot addition of polyelectrolyte was developedto overcome the flocculation occurring in the dropwiseapproach when the mixing ratios were close to unity. Thismethod proved very versatile for the elaboration of colloidalPECs. Indeed, the process was far less sensitive to the orderof mixing than the previous one, as the excess polymer couldbe used either as titrant or as starting solution. Variations ofthe particle mean size on varying n+/n- ratios were similar

for both modes of addition (slow and fast). This one-shotprocess, simple to setup, easy to perform, insensitive to themixing order of reactants, led to highly stable chargedbiopolymer-based colloids, which could be obtained in fairamount and at high solid contents. Further work is underwayto exemplify this approach toward colloids based on biopoly-mer polyelectrolyte complexes.

Acknowledgment. This work was financially supportedby a grant from the French Ministry of Research andTechnologies. C.S thanks P. Sorlier for the preparation ofthe high molar mass chitosan.

Note Added after ASAP Posting.Minor changes weremade to sentences in the first and second paragraphs of theIntroduction for the sake of clarity. A change was also madeto the caption of Table 1. The manuscript was originallyposted on 8/6/2004. The corrected manuscript was repostedon 8/23/2004.

References and Notes

(1) Dautzenberg, H. InPhysical Chemistry of polyelectrolytes; Radeva,T., Eds.; Surfactants Science; Dekker: New York, 2001; Vol. 99, p743.

(2) Bekturov, E. A.; Bimendina, L. A.AdV. Polym. Sci.1980, 43, 100.(3) Kabanov, V. A.; Zezin, A. B.Makromol. Chem., Suppl. 1984, 6,

259.(4) Kabanov, V. A.Polym. Sci.1994, 36, 143.(5) Tsuchida, E.; Abe, K. InDeVelopments in ionic polymers 2; Wilson,

A. D., Prosser, H. J., Eds.; Elsevier Applied Science: New York,1986; Chapter 5, p 191.

(6) Tsuchida, E.; Abe, K.AdV. Polym. Sci. 1982, 45, 1.(7) Dautzenberg, H.; Karibyants, N.Macromol. Chem. Phys. 1999, 200,

118.(8) Dautzenberg, H.; Hartmann, J.; Grunewald, S.; Brand, F.Ber. Bunsen-

Ges. Phys. Chem.1996, 100, 1024.(9) Dautzenberg, H.; Jaeger, W.Macromol. Chem. Phys. 2002, 203, 2095.

(10) Schatz C.; Lucas J-M.; Viton C.; Domard, A.; Pichot, C.; Delair,T. Langmuir2004, in press.

(11) Roberts, G. A. F.Chitin chemistry; Macmillan Press Ltd: London,1992.

(12) Sorlier, P.; Viton, C.; Domard, A.Biomacromolecules2002, 3, 1336.(13) Schatz, C.; Viton, C.; Delair, T.; Pichot, C.; Domard, A.Biomac-

romolecules2003, 4, 641.(14) Lee, K. Y.; Ha, W. S.; Park, W. H.Biomaterials1995, 16, 1211.(15) Hirano, S.; Seino, H.; Akiyama, Y.; Nonaka, I.Polym. Mater. Sci.

Eng. 1988, 59, 897.(16) Liu, W. G.; De Yao, K.J. Controlled Release2002, 83, 1.(17) Mitsuya, H.; Looney, D. J.; Kuno, S.; Ueno, R.; Wong-Staal, F.;

Broder, S.Science1988, 240, 646.(18) Cui, Z.; Mumper, R. J.J. Controlled Release, 2001, 75, 409.(19) Janes, K. A.; Calvo, P.; Alonso, M. J.AdV. Drug DeliVery ReV. 2001,

47, 83.(20) Vachoud, L.; Zydowicz, N.; Domard, A.Carbohydr. Res.1997, 302,

169.(21) Allan, G. H.; Peyron, M.Carbohydr. Res.1995, 277, 257.(22) Allan, G. H.; Peyron, M.Carbohydr. Res.1995, 277, 273.(23) Hirai, A.; Odani, H.; Nakajima, A.Polym. Bull.1991, 26, 87.(24) Evans, J. M. InLight Scattering from Polymer Solutions; Huglin,

M. B., Ed.; Academic Press: New York, 1972; Chapter 5.(25) Koppel, D. E.J. Chem. Phys. 1972, 57, 4814.(26) Wang, W.; Bo, S.; Li, S.; Qin, W.Int. J. Biol. Macromol. 1991, 13,

281.(27) Beyer, P.; Nordmeier, E.Eur. Polym. J. 1995, 31, 1031.(28) Buchhammer, H. M.; Petzold, G.; Lunkwitz, K.Colloid Polym. Sci.

2000, 278, 841.(29) Dautzenberg, H.Macromolecules1997, 30, 7810.(30) Schatz, C.; Delair, T.; Pichot, C.; Viton, C.; Domard, A.Langmuir

2003, 19, 9896.(31) Walker, H. W., Granti, S. B.J. Colloid Interface Sci.1996, 179,

552.

BM049786+

Figure 9. Hydrodynamic diameters (a) and electrophoretic mobilities(b) of chitosan/dextran sulfate PEC particles (see experimentalconditions in Figure 7) in 10-2 M acetic acid solution at pH ) 4.0 (µ) 1.5 × 10-3 M) as a function of added sodium chloride concentration.(4) n+/n- ) 2, (2) n+/n- ) 0.2.

Figure 10. Stability factor (W) as a function of the electrolyteconcentration (NaCl) at pH 4.0 for two compositions of chitosan/dextran sulfate particles (see experimental conditions in Figure 7):(2) n+/n- ) 2, (4) n+/n- ) 0.2.


Versatile and Efficient Formation of Colloids of Biopolymer-Based Polyelectrolyte Complexes

Documents

Transcript of Versatile and Efficient Formation of Colloids of Biopolymer-Based Polyelectrolyte Complexes