lamarque 2005

Physicochemical Behavior of Homogeneous Series ofAcetylated Chitosans in Aqueous Solution: Role of Various

Structural Parameters

Guillaume Lamarque, Jean-Michel Lucas, Christophe Viton, and Alain Domard*Laboratoire des Materiaux Polymeres et des Biomateriaux - UMR CNRS 5627, Domaine scientifique de la

Doua, Batiment ISTIL, 15, Bd. A. Latarjet, 69622 Villeurbanne Cedex (France)

Received June 24, 2004; Revised Manuscript Received September 15, 2004

Physicochemical properties of four different homogeneous series of chitosans with degrees of acetylation(DA) and weight-average degrees of polymerization (DPw) ranging from 0 to 70% and 650 to 2600,respectively, were characterized in an ammonium acetate buffer (pH 4.5). Then, the intrinsic viscosity ([!]0),the root-mean-square z-average of the gyration radius (RG,z), and the second virial coefficient (A2) werestudied by viscometry and static light scattering. The conformation of chitosan, according to DA and DPw,was highlighted through the variations of R and " parameters, deduced from the scale laws [!]0 ) KwM w

R

and RG,z ) K′M w" , respectively, and the total persistence length (Lp,tot). In relation with the different

behaviors of chitosan in solution, the conformation varied according to two distinct domains versus DAwith a transition range in between. Then, (i) for DA < 25%, chitosan exhibited a flexible conformation; (ii)a transition domain for 25 < DA < 50%, where the chitosan conformation became slightly stiffer and, (iii)for DA > 50%, on increasing DPw and DA, the participation of the excluded volume effect becamepreponderant and counterbalanced the depletion of the chains by steric effects and long-distance interactions.It was also highlighted that below and beyond a critical DPw,c (ranging from 1 300 to 1 800 for DAs from70 to 0%, respectively) the flexibility of chitosan chains markedly increased then decreased (for DA >50%) or became more or less constant (DA < 50%). All the conformations of chitosan with regards to DAand DPw were described in terms of short-distance interactions and excluded volume effect.

Introduction

Chitin and chitosan are linear copolymers of (1f4)-linked-2-acetamido-2-deoxy-#-D-glucan (GlcNAc) and 2-amino-2-deoxy-#-D-glucan (GlcN). If chitin is fully insoluble in bothaqueous and usual organic solvents, as long as the copolymeris soluble in dilute acidic media, it is termed chitosan. Despitechitosan being much less widespread in biomass than chitin,it can be obtained from partial N-deacetylation of chitin undersevere alkaline conditions.1-4 Chitosan finds numerousapplications5,6 in agriculture,7 biomedicine8-10 (for exampleas drug delivery system11-14), paper-making, water treatment,or food-industry.15,16 Its properties strongly depend on theDegree of acetylation (DA),17 corresponding to the molarratio of GlcNAc units within the chain, and the molecularweight,10,16 which influence not only its physicochemicalbehaviors18-20 but also its biological activity.21-23 Theknowledge of the chitosan behavior and conformation inaqueous solution versus DA, the weight-average degree ofpolymerization (DPw), and the polydispersity index (Ip) isthen of major interest, for example, in the production ofnanoparticles.13,24In dilute acidic media, thanks to the protonation of the

amino functions, chitosan is fully soluble and behaves as acationic polyelectrolyte. The value of the intrinsic pK (pK0)

was already demonstrated to increase from 6.46 to 7.32 as afunction of DA.19,25 Thus, both electrostatic and steric effectsmust be taken into account when studying the influence ofDA and DPw on the solution behavior of chitosan. Indeed,in the case of flexible polymer chains, the accumulation ofcharges would lead to a considerable expansion due to elec-trostatic repulsions. Among the analytical techniques allow-ing the investigation of chitosan behaviors, viscometry26-32and static and dynamic light scattering18,30,33-36 are the mostcommonly used. Although the studies on the physicochemicalbehavior and conformation of chitosan in aqueous solutionat high ionic strength are numerous, most of them onlyreported analysis on a very restricted range of chitosan interms of DA28,31,32,35-37 and DPw.29 Commercial samples weregenerally used without further purification and degraded bymeans of sonication,31,38-41 nitrous deamination,42 or acidichydrolysis, leading to relatively high and variable polydis-persity indexes.33 Thus, preparation of the samples could havesome important incidences on Ip39,43 and aggregate forma-tion42 by self-association of chitosan chains leading to curvedshapes of Zimm plots44 and rendering the determination ofMw (weight average molecular weight), RG,z45 (root-mean-square of the z-average gyration radius), and A242,46 (secondvirial coefficient) very tricky. In this work, the conformationof the copolymer was studied thanks to the calculations ofthe total persistence length Lp,tot, R and " parameters (from

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

131Biomacromolecules 2005, 6, 131-142

10.1021/bm0496357 CCC: $30.25 © 2005 American Chemical SocietyPublished on Web 12/16/2004

the scale laws [!]0 ) KwM wR and RG,z ) K′M w

" ), whichwere preferred to the determination of the empirical B31,47parameter.Studying polyelectrolytes in aqueous solution required

addition of salt to minimize the electrostatic contributionsand screen ionic sites along the chains. If some authorstraditionally used NaCl48 or NaOAc49 in the case of chitosan,it has been shown that these salts were unable to avoidaggregation of chitosan chains,26,42 especially at high DA.By breaking interchain hydrogen bonds in solution, am-monium salt, first proposed by Domard and co-worker50 thenAnthonsen et al.,42 is now preferred by several authors aseluent for HP-SEC51 and for the characterization of chitosanin aqueous solution18 for its ability to prevent chitosan fromaggregation. The choice of the solvent and salt,49,52 the ionicstrength52 and pH34,52 of the solution, and dn/dc (refractiveindex increment) as a function of DA seemed to bepreponderant in the consistency and reliability of the resultspublished.53,54 Furthermore, some authors processed theirchitosans either by deacetylation or reacetylation27,51 (inheterogeneous or homogeneous media, respectively) leadingto a blockwise and statistical ordering of the GlcN andGlcNAc residues along the polymer chain.55 These composi-tions heterogeneities could also represent a major drawbackfor the study of the laws of behavior of chitosan in aqueoussolutions and the calculation of the persistent lengths.44Indeed, Brugnerroto et al.53,56 demonstrated by molecularmodeling that a block, an alternate and a statistic orderingof the GlcNAc and GlcN units along the polymer chain, ledto three different variations of Lp as a function of DA. Thiswas attributed to the enhanced flexibility brought about byblocks of several screened GlcN units and could explain whyseveral authors found different behaviors for chitosans havingthe same DA but prepared from different processes. Amongthe means to process chitosan samples, reacetylation in ahydro alcoholic medium at different DAs is a well-knownreaction that allows us to reach precise DA values19 withoutdegrading the copolymer.20 The relatively narrow Ip of theproceeded samples was also demonstrated to decrease as DAincreased.18,20 Improving the clarification of chitosan solu-tions is also important to suppress aggregates that couldinduce wrong measurements and lead to curved Zimm plotsor negative second virial coefficients.42 Even though themechanism of the aggregate formation is still an opendiscussion, most of them can be removed either by centrifu-gation or filtration on decreasing the membrane pore size.29

Recently, the existence of a universal law of behavior forchitosan in aqueous solution according to DA was revealedby Sorlier et al.19 and then confirmed by Schatz et al.18 Bymeans of homogeneous series of acetylated chitosans withsimilar Ip, they demonstrated that properties of chitosandeduced from: potentiometry, interferometry, SLS andviscometry measurements underscored a triple behaviordepending on DA. When DA was increasing, properties ofchitosan were changing through three distinct domainsevolving from a polyelectrolyte state to that of isolatedcharges on polymer chains in a hydrophobic environmentand, between, a transition range. Since their purpose was tocheck the influence of DA on the physicochemical properties

of chitosan, their studies were restricted to a polymer seriesof DPw ) 2200 and 1000, respectively. The very narrowrange of DPw studied did not allow them to determine theconformational parameters R and ", or the persistence lengthsversus DPw. Thus, it was of major importance to estimatethe influence of DPw on the rigidity of chitosan to be thenable to foresee its conformation, whatever DA and DPw,according to a theorical approach taking into account theexcluded volume effect.To perform that, about thirty chitosans were homogen-

eously reacetylated (DAs ranging from 0 to 70%) from fourdifferent initial samples of low DA differing in their DPw(from 650 to 2600) and were either analyzed by capillaryviscometry and static light scattering (SLS), in “batch mode”,or coupled with steric exclusion chromatography (SEC,“micro-batch mode”) in an ammonium acetate buffer (pH4.5). The data resulting from these analyzes enabled us forthe very first time to better understand the combined role ofDPw and DA on the behavior of chitosan in aqueous solutionsat high ionic strength.

Experimental SectionPurification and Acetylation of Chitosan. Four different

batches of chitosan (batches 144, 154, 124, and 114) withDAs of 2.23, 1.2, 1.51, and 2.65%, respectively, werepurchased from France Chitine (France). The chitosanpowders were dissolved (5 g/L (w/v)) in an aqueous aceticacid solution, filtered successively through 5, 1.2, 0.8, and0.45 µm pore size membranes (Millipore), and precipitatedby means of a dilute aqueous solution of ammonia. Afterrepeating washings with deionized water and centrifugationscycles (until the conductivity of the supernatant reached thatof water), purified chitosans were then lyophilized.Acetylation was performed as follows. Purified chitosans

were then dissolved in a 0.1% acetic acid/1,2-propanediolmixture. Different amounts of a solution of pure and freshacetic anhydride in 1,2-propanediol were added, in stoechio-metric conditions, to reach the desired DA. At the end ofthe reaction, reacetylated chitosans were fully precipitatedby addition of dilute aqueous ammonia and washed severaltimes with deionized water at pH 7.5 in order to preservethe -NH2 form. Reacetylated chitosans with the lowest DPwand highest DA were fully soluble whatever the pH. In thelatter case, salts and propanediol were removed by ultrafil-tration (Amicon) against deionized water (pH 7.5) througha YC05 membrane (MwCO 500, Millipore). In all cases,reacetylated chitosans were finally lyophilized.Characterization of Chitosans. Reacetylated samples

were dissolved in dilute acidic D2O (at pD 3-4), and theirDAs were analyzed by 1H NMR spectroscopy, as proposedby Hirai et al.57 Spectra were recorded on a Bruker-Spectrospin AM 300 spectrometer (300 MHz). About 200-250 scans were acquired. The absence of propanediol withinthe reacetylated chitosans was confirmed for all samples. Thewater content was determined by thermo-gravimetric analysis(DuPont Instrument TGA 2000).Viscometry. Intrinsic viscosities were measured at 25 (

0.1 °C using an Ubbelohde automatic capillary viscometerwith a inner diameter of 0.53 mm (Viscologic TI 1

132 Biomacromolecules, Vol. 6, No. 1, 2005 Lamarque et al.

SEMATech). Reacetylated chitosans were dissolved (0.1-0.3% (w/w)) in a degassed 0.2 M acetic acid/0.15 Mammonium buffer (pH 4.5). The critical concentration ofchain entanglement C* was determined considering C*[!]0) 1. The following analyses were achieved with solutionsof concentration at least 0.8 times lower than C*.Static Light Scattering (SLS), Batch Mode. Each chi-

tosan was dissolved in the same degassed acetate buffer asthe one used for viscometry measurements. All samples werefiltered on 0.45 and then 0.22 µm pore size membranes, andanalyses were performed 24 h after the last filtration so thatthe solutions could degas. Samples with the highest DAsand DPws were further filtered on a 0.22 µm pore sizemembrane before analysis to remove aggregates, if necessary.Multi-angle laser light scattering (MALLS) detection wasused in batch (i.e., alone) and micro-batch mode (i.e., coupledwith a HP-SEC device). Batch measurements were per-formed with a Dawn DSP-EOS (Wyatt) equipped with a 25mW Ga/As laser operating at $ ) 690 nm. Solutions wereanalyzed in scattering glass cells at five different concentra-tions, below C*, and scattered light intensities were measuredat 18 angles ranging from 23 to 147°. Light intensitymeasurements were derived following the Rayleigh-Debyeequation, andMw, the weight-average molecular weight, RG,z,the root-mean-square z-average of the gyration radius, andA2, the second virial coefficient, were deduced by means ofZimm plots. The refractive index increment dn/dc was chosenas a function of each DA, according to Schatz et al.18

High-Performance Size Exclusion Chromatography,Micro-Batch Mode. The polymer separation was performedon two serially connected columns (TSK G3000-PW and

TSK G6000-PW, i.d. ) 7.8 mm, l) 300 mm). The detectionwas operated by a differential refractometer (Waters 410)coupled on line with a MALLS detector (Dawn DSP-F,Wyatt) equipped with a 5 mW He/Ne laser operating at $ )632.8 nm. Analyses were performed in micro-batch modeusing the K5 flow cell. A degassed AcOH (0.2 M)/AcONH4(0.15 M) buffer (pH 4.5) was used as eluent after two fil-trations on a 0.22 µm pore size membrane (Millipore). Theconcentrations used ranged from 0.04 to 0.1% (w/w) beforeinjection so that the results obtained for the different chito-sans were independent of the starting solution concentrations.All chitosan solutions were filtered on 0.45 and then 0.22µm pore size membranes. The flow rate was maintained at0.5 mL/min, and the amount of sample injected was 200µL.

Results and Discussion

Characterization of the Studied Samples. The mainpurpose of this work was to investigate laws of behaviorrelated to the physicochemical properties of chitosan inaqueous solutions in relation with both DA and DPw. Theionic strength of the solution was fixed to µ ) 0.15 M withan ammonium salt to sufficiently screen the protonated aminogroups, hinder electrostatic repulsions, and prevent chainsfrom aggregation.Four purified chitosans with starting DA < 5% and

differing in their initial degree of polymerization werereacetylated in mild conditions allowing us to produce fourseries of chitosans with DAs of 10, 30, 40, 50, 60, and 70%without degrading the polymer (Table 1). For more clarity,processed samples were named Cx,y, where x represents theaverage DPw of the batch, determined by HP-SEC (i.e., 650,

Table 1. Molecular Characteristics of Chitosans of Various DAs and DPw Determined by SLS in Batch and Micro-Batch Modea

HP-SEC analysis batch modebatch DA(%) Mw(105g/mol) DPw Mw (105g/mol) DPw Ip114 2.7 6.024( 0.181 3716( 112 4.188( 0.125 2583( 77 1.44( 0.04

9.8 4.260( 0.128 2580( 79 3.899( 0.115 2362( 70 1.37( 0.0329.4 4.521( 0.136 2608( 84 4.198( 0.119 2421( 69 1.27( 0.0240.0 4.494( 0.135 2528( 84 4.276( 0.123 2405( 69 1.28( 0.0249.5 4.578( 0.137 2518( 85 4.315( 0.124 2374( 68 1.29( 0.0259.9 4.391( 0.132 2359( 82 6.012( 0.187 3230( 100 1.51( 0.0766.0 4.619( 0.129 2447( 80 12.590( 0.210 6671( 113 2.13( 0.1059.9b n.d. n.d. 4.406( 0.137 2367( 70 1.20( 0.0466.0b n.d. n.d. 4.477( 0.213 2372( 69 1.18( 0.05

144 2.2 1.232( 0.049 761( 30 0.986( 0.023 609( 14 1.61( 0.138.6 1.182( 0.035 718( 22 0.900( 0.021 546( 13 1.43( 0.0822.3 1.211( 0.048 711( 28 0.889( 0.023 522( 14 1.34( 0.0940.1 1.166( 0.035 656( 20 0.873( 0.022 491( 12 1.31( 0.0745.9 1.154( 0.035 640( 19 0.867( 0.024 481( 13 1.24( 0.1362.1 1.138( 0.046 608( 24 0.857( 0.026 458( 14 1.17( 0.0769.4 1.301( 0.047 684( 25 0.861( 0.028 453( 15 1.11( 0.06

124 1.5 3.413( 0.089 2112( 55 3.048( 0.110 1886( 68 1.65( 0.069.8 3.548( 0.106 2149( 64 3.105( 0.098 1880( 59 1.37( 0.0630.9 3.940( 0.118 2265( 68 3.311( 0.112 1903( 64 1.29( 0.0639.4 4.106( 0.099 2313( 56 3.373( 0.121 1900( 68 1.28( 0.0548.2 4.052( 0.109 2236( 60 3.483( 0.119 1922( 66 1.29( 0.0559.7 4.129( 0.121 2219( 63 5.823( 0.175 3129( 94 1.53( 0.0571.0 4.170 ( 0.083 2185 ( 44 7.583 ( 0.196 3974 ( 103 2.14 ( 0.0559.7b n.d. n.d. 3.565 ( 0.115 1916 ( 62 1.20 ( 0.0571.0b n.d. n.d. 3.565 ( 0.126 1868 ( 55 1.12 ( 0.05

154 1.2 2.668 ( 0.069 1652 ( 43 2.134 ( 0.079 1322 ( 49 1.57 ( 0.0410.4 2.813 ( 0.084 1701 ( 51 2.250 ( 0.081 1361 ( 49 1.31 ( 0.0331.6 2.844( 0.100 1632( 57 2.512( 0.082 1441( 47 1.20( 0.0339.8 2.977( 0.060 1675( 34 2.371( 0.080 1334( 45 1.26( 0.0249.9 2.904( 0.087 1596( 48 2.427( 0.085 1334( 47 1.27( 0.0260.1 2.974( 0.089 1597( 48 2.455( 0.090 1318( 48 1.20( 0.0270.5 3.153( 0.095 1654( 48 2.600( 0.086 1364( 45 1.16( 0.02

a n.d., not determined. b Additional filtration on 0.22 µm.

Acetylated Chitosans in Aqueous Solution Biomacromolecules, Vol. 6, No. 1, 2005 133

1 600, 2 200, and 2 600 for the batches 144, 154, 124, and114, respectively), and y the DA.All of the samples proceeded exhibited a random distribu-

tion of the GlcNAc/GlcN residues along the polymer chainby 1H NMR spectroscopy, according to the method of Vårumet al.58HP-SEC measurements revealed a specially narrow range

of DPw for each series. Thanks to the elimination of solubleoligomers during the repeated washing/centrifugation cyclesof the reacetylation process,18,19 Ip monotonically decreasedon increasing DA. Ip found by SLS in batch mode for thetwo highest DAs of the series 114 and 124 strongly suggested

that these four chitosan solutions underwent self-association.These aggregates were concentration dependent since Zimmplots revealed a positive curvature of the concentrationfunction (Figure 1). Most of these aggregates were removedby means of an additional filtration through 0.22 µm poresize membrane (Figure 1, parts b and c) with a loss of matter<7%, and a Ip decreasing as a consequence. The loss ofmatter could affect the measurements achieved by static lightscattering (see below) but was therefore reproductive andthe filtrated out fraction was composed of chitosan chainswith similar amino group distribution. Thus, it seemed thataggregation was only dependent on the DA and molecular

Figure 1. Zimm plots (in negative scale representation) for various chitosans in solution in an AcOH (0.2M)/AcONH4 (0.15 M) buffer: (a)C2600,2.7; (b) C2600,66 before and (c) after further filtration on 0.22 µm pore size membranes.


http://dontstartme.literatumonline.com/action/showImage?doi=10.1021/bm0496357&iName=master.img-000.png&w=327&h=533

weight of chitosan and not on the mode of distribution ofits repetitive units.Despite the fact that the mechanism of aggregate formation

is not well-understood, some authors stated that acetamidogroups of GlcNAc could be involved51 in the formation ofhydrophobic interactions between polymer chains, but thiswas recently contested by Philippova et al.59 To ourknowledge, the mode of aggregation (end to end or side byside aggregation) was, in particular, still unsolved. Thus, ifnecessary, the characteristics of the four chitosans C2600,59.9,C2600,66, C2200,59.7, and C2200,71 will be presented before andafter the additional filtration to evaluate the influence of theaggregates on the physicochemical parameters studied. Oursample preparation and conditions of analysis being closeto those of Schatz et al.,18 our results were compared to theirs,as mentioned in the different figure legends.Viscometry Measurements. The critical concentration of

chain entanglement C* of chitosan was calculated throughthe determination of the intrinsic viscosity [!]0 (Figure 2).The polydispersities of the samples were taken into account

according to a logarithmic distribution of the molecularweight60

where Ip ) (Mw/Mn) ) (h + 1/h) and ! is the gammafunction.In the following results, K was preferred to Kw, for its

independence with Ip. The variation of [!]0 versus DA wasquite similar to that illustrated by the general law of behaviorwe may now observe for almost all of the physicochemicalproperties of chitosan, which depend on the structural chargedensity of the polymer.19,20 Three domains were clearlyidentified: (i) for low DA values, chitosan exhibited a typicalpolyelectrolyte behavior revealed by high [!]0 values, (ii)for higher DAs (>50%), [!]0 fell down to much lower values,and (iii) in between, hydrophilic and hydrophobic effects onchain dimensions were counterbalanced and a transition rangewas observed. Thus the variation of C*, defined as thereverse of [!]0 in the case of Gaussian coils, underscoredthe same behavior (data not shown) and varied from 0.63 to

4 mg/mL. Considering these findings, concentrations wechoose for the parent and fractionated solutions (0.4-1 mg/mL and 0.05-0.4 mg/mL for SLS in micro-batch and batchmode, respectively) guaranteed analyses in dilute regime.A first insight in the study of chitosan conformation re-

quired the determination of the K and R set of constants ofthe Mark-Houwink-Kuhn-Sakurada (MHKS) equation [!]0) KM w

R, calculated by plotting log([!]0) versus log(Mw).Mw was determined either by SLS in batch or micro-batchmode (Figure 3). Variations of R and log(K) versus DA wererepresented in Figure 4 and followed the same trend as thoseof Figure 2. Unfortunately, R and log(K) values derived fromMw, determined either by HP-SEC or SLS, revealed somedisparities. The highest Mw could be overestimated by HP-SEC because of a nonoptimum column separation and/orpolymer chain associations (Table 1). Values of the Hugginsconstant K, decreasing versus DA, were consistent withothers,27,28 indicating that the solubility of chitosan in thebuffer decreased as a function of DA, and were confirmedby the variation of A2 (see below). For batch measurements,chain conformation of chitosan varied from a random coilin a good solvent (R ) 0.65) to a wormlike chain in perturbedconditions (R ) 1.05).61 Thus, the increase of rigidity versusDA (i.e., the decreasing charge density) suggested that theammonium salt sufficiently screened the amino groups tocounterbalance the electrostatic repulsions responsible forthe chain expansion. Chitosan was subsequently allowed tocoil up and its apparent flexibility enhanced.Nevertheless, R values were largely over 0.5 (i.e., in %

conditions) whatever the DA within 0-25%, indicating that,

Figure 2. Variation of [!]0 as a function of DA for the four series ofreacetylated chitosans of average DPw ) 650 ()), 1600 (4), 2200(O) and 2600 (0). The results were compared to those of Schatz etal.18 (DPw ) 1000, 3)

[!]0 ) KwMwR and Kw ) K

2!(R + h + 1)hR!(h + 1)

Figure 3. Determination of R and log(K) versus DA. Mws weredetermined by SLS in (a) micro-batch and (b) batch measurements.Experimental points in dotted circles were those of Schatz et al.18




despite the high ionic strength, we may assume that theelectrostatic contributions to the excluded volume effect wasnot completely inhibited for very low DAs. As DA increased,Wang et al.28 proposed that intramolecular hydrogen bond-ings could prevent the acetamido group and the #-(1f4)linkage of the glucopyranose ring from free rotation and beresponsible for an increase of excluded volume caused bythe bulky acetamido groups. Even though R depends on thenature of the solvent, ionic strength and Ip, Wang et al.28 (0< DA < 31%) and then Anthonsen et al.27 (0 < DA < 60%)found R linearly increasing with DA, but no one observed atransition range within DA) [20;50], as we did. The plateauwe observed could be the main consequence of the differentsolvents and dn/dc values we used to perform our studies.The ammonium salt used in this report could be responsiblefor particular interactions between the polymer and thesolvent and explain the different results of Anthonsen et al.27However, the same behavior was already observed by Sorlieret al.62 for the variation of dn/dc in an aqueous solventcontaining KClO4 as added salt. The different solvent theyused and the similar behaviors they obtained suggested thatonly the purity and the structure of the sample, and, on theother side, the pH and ionic strength of the solution areresponsible for the results observed. Furthermore, comparedto Wang et al.,28 the nonlinear decrease of the refractive indexincrement along with DA could also induce importantvariations in the determination of Mw and subsequently lead

to a quite different behavior of R. These two phenomenacould explain the absence of a transition range in their results.Nevertheless, we may remind the reader that dn/dc is aparameter traducing the electric polarizability and a parameterlargely influenced by the apparent charge density, then byDA and the screening of the protonated amino group.Static Light Scattering measurements. (A) Variations

of RG,z and Lp,tot. Variations of RG,z were followed as afunction of DA by SLS in batch and micro-batch mode. RG,zvalues determined by HP-SEC were somewhat dependenton the separation conditions of the samples, especially forlow DA, and could explain the differences observed withthose determined by SLS in batch mode (Figure 5). Depend-ing on DPw of the chitosan series analyzed, two kinds ofbehavior were observed on increasing DA. For DPw > 1600,the variation of RG,z seemed to follow a different behavioras those observed until now and increased versus DA. Onthe contrary, for DPw < 1600, the decrease of RG,z versusDA, which followed the same trend as [!]0, was surprisingwhen compared to the variation of R. The polyelectrolyticbehavior was characterized by the relatively high values ofRG,z, then, as the proportion of GlcNAc units increased, thehydrophobicity of the chains was responsible for theirdepletion and the fall of RG,z. On the opposite, the conforma-tion change of chitosan from a flexible to a stiff molecule

Figure 4. Variations of (a) log(K) and (b) R from the MHKS equationversus DA. The curves were derived from Mw determined either byHP-SEC (s) or SLS (- - -) in batch mode.

Figure 5. Variation of RG,z versus DA determined by SLS for thefour series of reacetylated chitosans: DPw ) 650 ()), 1600 (4), 2200(O), and 2600 (0). (a) With the HP-SEC device and (b) in batchmode. The results were compared to those of Schatz et al.18 (DPw )1000, 3). Solutions of chitosan were filtered on 0.45 and 0.22 µmpore size membranes before analysis. A further filtration on 0.22 µmwas necessary for the four chitosans of highest DPw and DA.




should subsequently imply an increase of RG,z.34 The presenceof a critical DPw,c for which we observed a transition ofconformation suggested short- and long-distance interactionswere involved, this latter being represented by the globalexcluded volume effect #tot(DA, DPw).As a first attempt to explain the different behaviors of

chitosan in solution, we propose to share the global excludedvolume effect into an electrostatic, #el, and a steric term, #st

If the electrostatic contribution #el would mainly depend onthe charge density, and subsequently to DA, on the opposite,both DPw and DA should have a strong influence on #st.Indeed, whatever the increase of the number of repetitiveunits or the bulky GlcNAc residue proportion, they bothimply a loss of chain flexibility by steric hindrance orentropic effects. We are totally conscious of the approxima-tions introduced by the split of #tot(DA, DPw) into twocontributions. According to the Odijk and Houwaart63relation, #tot(DA, DPw) ) 8&Lp,tot

2 $D, where $D is the Debyescreening length, the Odijk-Skolnik-Fixman (OSF) theorypredicted that the persistence length Lp,tot can be shared intoLp,i, the intrinsic persistence length characterizing the localstiffness of the chain, and Lp,e, the electrostatic persistencelength related to the electrostatic repulsions between adjacentionic sites: Lp,tot ) Lp,i + Lp,e. Introducing Lp,tot in the formerequation results in a “coupled term” (∝ Lp,i ! Lp,e) in theexpression of #tot(DA, DPw) we must take into account. Thiscondition was satisfied when studying the variation of #totversus DA, since both #el(DA) and #st(DA,DPw) varied withDA. On the other hand, we will see below that theelectrostatic contribution to #tot can be reasonably neglectedwhen modeling the conformation of chitosan versus DPw.Considering this point of view, for DAs ranging from 50

to 70%, the continuous loss of charge density of chitosanchains leading to the decrease of #el could be counterbalanced(DPw ) 1600) or even overcome (DPw > 1600) by thecontinuous increase of #st, ascribable to the steric hindrancebrought about by GlcNAc units. As a result, the increase of#tot reflected the extension of the chains prevented fromdepletion. This hypothesis will be comforted by the variationsof Lp,tot and A2 (see below). However, the study on thevariations of RG,z became trickier for very high values ofDA because of self-association of chitosan, which led to theformation of aggregates (Figure 5b) and curved Zimm plots(Figure 1). An additional filtration of these solutions on 0.22µm pore size membranes enabled us to decrease RG,z to morereliable values, but caused experimental uncertainties.A second empirical method to follow the chitosan con-

formation was to determine the K′ and " parameters derivedfrom the scale law RG,z ) K′M w

" . These parameters werecalculated as described previously for viscometric measure-ments, but a chromatographic method was also used todetermine " average values, as follows. Each set of the fourchitosans with equivalent DAs were injected in HP-SEC.The average values of " calculated from each eluted peaksby means of the software ASTRA 4.7 (Wyatt) were thencompared to those derived from Figure 5. Except for thevalues determined by this last method, both variations of

log(K′) and " agreed with the previous observations and alsorevealed an increasing chain stiffness (Figure 6).

" was respective of DA and varied from 0.5 (Gaussiancoils in % conditions) to 0.65 (wormlike chains with excludedvolume effect). The value of " ) 0.46 obtained for DA <5% in batch mode seemed to be therefore underestimated.This contradicts previous observations by coupled SLSmeasurement in batch mode33,45 and HP-SEC53 reporting "as being constant on increasing DA. Only Wu et al.37 found" > 0.6 in a 0.2 M AcOH/0.1 M AcONa buffer by meansof dynamic light scattering, characteristic for stiffer polymers,but the DA of 9% they used was far from confirming ourresults. Determination of RG,z being more sensitive toexperimental errors than [!]0, variations of " and log(K′) onincreasing DA were proportionally much lower than thoseof R and log(K) and definitely more difficult to underline.Despite all of the difficulties we met with in the determi-nation of RG,z, the variation of the persistence length Lp,totalong with DA and DPw could be helpful to confirm or notthe previous variations of conformation.Persistence lengths were calculated for each chitosan from

the Benoit-Doty relation, valid for monodisperse wormlikechains in % conditions64

#tot(DA, DPw) ) #el(DA) + #st(DA,DPw)

Figure 6. Variations of log(K′) and " as a function of DA, derivedfrom Figure 5 (0, micro-batch and O, batch modes) and average "values calculated from SEC-peaks of each set of the four chitosanswith equivalent DA (4).

RG,z2 )

LLp,tot3 - Lp,tot

2 +2Lp,tot

3

L -2Lp,tot

4

L2(1 - e-L/Lp,tot)



L corresponds to the contour length of the polymer chain,and the length per monomer unit was kept constant to 4.9Å. Determining accurate Lp,tot implied the consideration ofthe polydispersity of the samples. It was evaluated by anormalized and logarithmic theorical function44 that fittedcorrectly our experimental weight distribution (by HP-SEC).#tot was estimated through the evaluation of the Floryexpansion coefficient for the gyration radii RG and Lp,totcalculated thanks to the iterative method proposed by Berthet al.33 from RG,z and Mw determined by SLS in batch mode.Lp,tot was also determined by fitting a theorical line dependingon Lp,tot to every plot (y ) log(Mw), x ) log(RG,z)) obtainedfrom the HP-SEC peaks of each chitosan samples, thanksto the software Astra. Unfortunately, as pointed by Berth etal.,29 one can hardly highlight any systematic effect of bothDA and DPw on the RG,z-Mw relationship, when derivedfrom chromatographic peaks. This was also illustrated bythe irrespective variation of " versus DA by using this lastmethod (Figure 6b). The difficulty to obtain well-Gaussianchromatographic peaks led us to estimate Lp,tot on a narrowdistribution of Mw, where the radius of gyration dependenceversus Mw was strictly linear, which was only partiallyrepresentative of each chitosan samples. This could explainthe inconsistent results observed when calculating Lp bymeans of the linear fitting (Table 2) and underlined that bothRG,z and " values derived from HP-SEC/MALLS peaks mustbe considered very cautiously when trying to study theirdependence against DA, because they both depend on thequality of sample separation. In contrast, determination ofLp,tot by SLS in batch mode led to more reliable results(Figure 7). It is worth noting that the similarity between allthe plots of Figures 4-7 confirm the consistency of our data.Thus, the universal law of behavior describing the variation

of a given property of chitosan with DA seemed to be alsovaluable for the description of its conformation in aqueoussolution for dilute systems.Even though this increase was theorically proposed by

molecular modeling,56,65 this is the first experimental reportof a significant rise of Lp,tot along with DA by SLS mea-surements in batch mode. Only Brugnerotto et al.53 observedthe same trend by SEC/MALLS measurements in micro-batch mode but their results relied on the same techniquewe used to determine Lp,tot by chromatographic fit, whichled, in our case, to inconsistent results. Moreover, they useda constant dn/dc ratio when performing their analysis, despiteits increase along with DA is now established. Nevertheless,in batch mode measurements, one can conclude that, for DPw) 650, the increase of Lp,tot versus DA was inconsistent withthe progressive fall of RG,z (Figures 5b and 7a). Thisunexpected result should rather be ascribable to the slightfall in the polydispersity index along with DA of this chitosanseries (Table 1). Specifically, the logarithmic function weused to describe each mass distribution was very sensitiveto the polydispersity index of the sample.44 The fact that Ipobviously decreased on increasing DA could lead to under-and over-estimated Lp,tot values, for DA < 25% and DA >50%, respectively, and may contribute to the describedvariations. The artifact in the calculation of Lp,tot resultingfrom the decrease of Ip on increasing DA could then takepart of the increase of chain rigidity observed in Figure 7a,

Table 2. Persistence Lengths Determined from Each ElutedHP-SEC Peaka

batch DA(%) Lp(nm) DPw Ip114 2.7 12.0( 0.7 3716( 112 1.28( 0.02

9.8 12.6( 0.8 2580( 79 1.28( 0.0229.4 13.4( 1.0 2608( 84 1.27( 0.0240.0 13.0( 1.0 2528( 84 1.25( 0.0249.5 13.0( 1.0 2518( 85 1.24( 0.0259.9 14.4( 1.2 2359( 82 1.13( 0.0766.0 10.6( 0.9 2447( 80 1.13( 0.10

144 2.2 7.8( 0.5 761( 30 1.27( 0.138.6 9.0( 0.6 718( 22 1.23( 0.0822.3 6.8( 0.4 711( 28 1.24( 0.0940.1 7.6( 0.5 656( 20 1.23( 0.0745.9 6.0( 0.4 640( 19 1.19( 0.1362.1 7.2( 0.5 608( 24 1.15( 0.0769.4 8.6( 0.6 684( 25 1.10( 0.06

124 1.5 14.4( 1.2 2112( 55 1.43( 0.069.8 12.0( 0.7 2149( 64 1.37( 0.0630.9 12.6( 0.8 2265( 68 1.22( 0.0639.4 13.2( 1.0 2313( 56 1.21( 0.0548.2 12.9( 1.0 2236( 60 1.14( 0.0559.7 11.2( 0.7 2219( 63 1.15( 0.0571.0 12.3( 0.8 2185( 44 1.18( 0.05

154 1.2 10.3( 0.9 1652( 43 1.24( 0.0410.4 10.6( 0.9 1701( 51 1.26( 0.0331.6 11.4( 0.7 1632( 57 1.15( 0.0339.8 12.8( 1.0 1675( 34 1.17( 0.0249.9 12.5( 0.8 1596( 48 1.18( 0.0260.1 10.9( 0.9 1597( 48 1.15( 0.0270.5 12.1( 0.7 1654( 48 1.09( 0.02

a The Ip were those derived from the linear dependence of the (y )log(Mw), x ) log(RG,z)) plots.

Figure 7. Variations of Lp,tot versus (a) DA and (b) DPw. Only samplesC2600,59.9, C2600,66, C2200,59.7, and C2200,71 after their additional filtrationon 0.22 µm pore size membrane were represented. The results ofSchatz et al.18 were only represented on Figure 7(a).



for DPw) 650. Thus, for DPw< 1 600, we strongly assumedthat the variation of Lp,tot must decrease or, at least, be moreor less constant on increasing DA, as that observed by Schatzet al.18 Compared to ours, the relatively high Ips of thesamples analyzed by Schatz et al.18 also seemed to beresponsible for the relatively low Lp,tot they calculated. Thus,their results could have fitted correctly our laws of behaviorin the single case where the polydispersity indexes were moreor less equal to those presented here. As a conclusion, it isworth noting that, for DPw < 1600, the excluded volumeeffect is negligible and only short-distance interactions wereresponsible for the conformation of chitosan.On the other hand, in the case of DPw > 1 600, variations

of Lp,tot and RG,z versus DA were consistent with each otherand could be explained by the theory of the excluded volume,in the case of long-distance interactions became preponderantto the detriment of any others. As DA increased, #st(DA,-DP) was successively negligible (DA < 25%), equal (25%< DA < 50%), and higher (DA > 50%) than #el(DA).Besides, #el(DA) could keep on falling as the charge densityalong the polymer chains dropped (Figure 8). This simplerepresentation could be used to explain the conformation ofchitosan chains observed in solution as a function of DA. Itis of major importance to realize this representation impliesthat #st(DA,DP) and #el(DA) are coupled each other (throughDA), which is respective to the OSF theory.In contrast, the variation of Lp,tot versus DPw underlined

quite different behaviors (Figure 7b). Even though Lp,tot couldhave been overestimated for the chitosan series C650, itseemed therefore that, whatever the DA, the increase of thechain length from DPw ) 650 to a critical chain length(illustrated by DPw,c) led to enhance the polymer flexibility.This strongly suggested that, in the range of DPw ) [650;DPw,c], the chain stiffness was mainly limited to short-distance interactions, and neither #st(DA,DP) nor #el(DA)varied in a large extent, whatever the DA. On the other hand,for DPw > DPw,c, long-distance interactions became prepon-derant and two kinds of behavior were highlighted. (i) ForDA below 50%, Lp,tot seemed to be nearly constant as thechains became longer, underlining that #st(DA,DP) = #st-(DA) was almost constant versus DPw. Thus, for a nonneg-ligible charge density, where the electrostatic contributionto the excluded volume effect was superimposed to that of

the steric hindrance, the influence of the chain length didnot seem to be preponderant. (ii) In contrast, on increasingDA from 50% to 70%, the fall of the electrostatic contribu-tion to the global excluded volume effect, due to the dropof charge density, was counterbalanced by the increase of#st(DA,DP) because of the addition of new GlcNAc residuesalong the chain, which emphasized the participation of sterichindrance to the excluded volume effect. As a consequence,for DA > 50%, chitosan chains tended to become stiffer asDPw increased (Figure 9).It is worth noting that previous studies from Tsaih et al.31

on a chitosan of DA ) 17% confirm our results. Thanks toviscometric measurements, they pointed out that chitosanchains were rather stiff and flexible for DPws below andbeyond DPw,c) 1300, respectively. Unfortunately, the highlypolydispersed samples (Ip > 2.5) they used and the verynarrow range of DA and DPw studied led to imprecise andtoo specific results. The present work decisively extends thestudy to a wide range of chitosans in terms of DA and DPw.It also underlines that a slight shift of the conformationaltransition from DPw ) 1800 to 1300 occurs for DAs rangingfrom 0 to 70%, respectively. This was ascribable to anincreasing participation of #st(DA,DP) to the global excludedvolume effect as DA increased (Figure 8). From thesefindings, it is now established that the whole conformationsand behaviors of chitosan in a dilute aqueous system versusDPw can be simply explained by considering two participa-tions, #st(DA,DPw) and #el(DA), to the global excludedvolume effect. However, if splitting #tot in two terms can beconsidered as correct for describing a system versus DA,such an approximation could be applied versus DPw only if

Figure 8. Schematic representation of one possible variation of #st-(DA,DP) and #el(DA) versus DA, leading to the global behaviorobserved for #tot and Lp,tot in the case of DPw > 1600.

Figure 9. Schematic representation of one possible variation of #st-(DA,DP) and #el(DA) versus DPw leading to the global behaviorobserved for #tot and Lp,tot: (a) For DA < 50% and (b) for DA > 50%.




Lp,e= 0. In this case, the coupled term of the Odijk-Houwaartrelation can be neglected. For µ ) 0.15 M, Lp,e was de-termined as being lower than 0.2 nm for each degree of acet-ylation and then could be neglected in the calculation of Lp,tot.According to the Manning theory,66 ionic sites of a highlycharged polyelectrolyte are importantly screened by the coun-terions of a solution containing a high concentration of salt.Thus, at high ionic strength, the theory predicted that somecounterions of the solution could be condensed to the pro-tonated sites of the polymer, leading to the decrease of itsapparent charge density. This consequently implies that thesolution behavior of chitosan versus DPw, at high ionicstrength (µ ) 0.15 M), can be described more or less as aneutral copolymer.Nevertheless, for the highest DA values, self-association

of chitosan chains could play a major role in the behaviorobserved and aggregation artificially emphasize the apparentstiffness of chitosan. This led us to be very cautious whenconsidering the law of behavior on the variation of Lp,totversus DA or DPw, but both variations of R and " confirmedthe results observed for Lp,tot.As a matter of fact, to our knowledge, the behavior of

Lp,tot with DPw was never highlighted experimentally. Lp,totcovered the range of [4; 10] nm. These values were lowerthan those reported by Terbojevich et al.45 for two series ofcommercial chitosans of DA ) 15 (DPw ) [1200; 3750],Lp,tot ) [19; 21.5] nm) and 40% (DPw ) [950; 6700], Lp,tot) [20; 23] nm), and Brugnerotto et al.53 for one single seriesof chitosan with DPw " 680 of various DAs ()[0; 56] %,Lp,tot ) [11; 15] nm). In contrast, our values were moreconsistent with those reported by Berth et al.33 with Lp,tot )6 nm for a chitosan of DA ) 7% and a DPw ) 770. Thevarious preparations and clarification methods of the chitosansolutions we used before analysis were closer to those ofBerth et al.33 than the others, which could take part in thedifferences observed.Static Light Scattering Measurements. (B) Variation

of A2. The variations of A2 versus DA and DPw depend onthe intrinsic solubility of the copolymer and are linked to itsaffinity toward the solvent (Figure 10). The loss of solubilitywith DA was already attributed to the decrease of hydophi-licity of the polymer chains in favor of hydrophobic inter-actions brought about by GlcNAc residues18 and agreed withthat of log(K) (Figure 4). Despite the use of ammonium ace-tate as added salt, these measurements confirmed the forma-tion of aggregates for C2600,59.9, C2600,66, C2200,59.7, and C2200,71,revealed by a drop of A2 to very low or negative values.42On the other hand, the solubility of chitosan markedly

decreased as DPw increased from 650 to 1600 and thenleveled off for higher values. Both variations of A2 versusDA and DPw must be related to the variation of #el and #st.(i) According to Figure 8, on increasing DA, the fall of #elwas counterbalanced by that of #st and consequently resultedin the decrease of A2. This led to the increase of segment/segment interactions between chains to the detriment of theinteractions between chains and the solvent and wouldexplain the chain depletion observed for the smallest DPw(Figure 5). (ii) Second, whatever the DA, A2 drastically felldown to lower values on increasing Mw, proving again the

preponderant influence of the steric contribution to theexcluded volume effect on the chitosan solution behavior.More specifically, for DA > 50% (i.e., copolymer almostneutral) and DPw < 2600 (i.e., before aggregation), A2 wasproportional to M2, as predicted by the Flory-Krigbaumexpression A2 ) (NaNk2#tot/2M2)h(ztot), where Na is Avogadro’snumber, Nk is the number of Kuhn segment, and ztot is theexcluded volume parameter (h(ztot) ) (1/5.73ztot)ln(1 +5.73ztot)). However, the experimental and predicted valueswere nonconsistent between each other, which could be theconsequence of (a) the nonnegligible effect of DA on #st,(b) the slight contribution of #el on the global volume effect,even at high DA and DPw, and (c) a relation morecomplicated between #st and #el that the one we proposed.

Conclusion

Thanks to the systematic study by viscometric and SLSmeasurements of four homogeneously reacetylated chitosanseries, differing in their DA and DPw, we extended all ofthe previous studies on this topic and highlighted differentlaws of behavior for the conformation of this polyelectrolyteversus DPw and DA. Conformations of chitosan werefollowed by the variation of the empirical conformationalparameters R and ", derived from the scale laws [!]0 ) KwM w

R and RG,z ) K′Mw" , respectively. We particularly high-

lighted that they could be represented by the universal lawof behavior of chitosan, valuable to describe the variation

Figure 10. Dependence of the second virial coefficient on: (a) DAand (b) Mw determined by SLS measurements in batch mode in AcOH(0.2 M)/AcONH4 (0.15 M) buffer.



of one of its physicochemical parameters versus DA. Theresults were then confirmed by SLS measurements and thevariation of Lp,tot. If experiments achieved by coupling onlinea SEC device led to unreliable results, those obtained in batchmode allowed us to calculate accurate and precise persistencelengths. Their variation versus DA agreed with that obtainedby viscometry measurement. Both conformation and solutionbehavior of chitosan at high ionic strength were explainedby splitting the excluded volume effect into an electrostaticand a steric contribution. In our model, splitting the excludedvolume effect was allowed because: (i) on varying DA, thecontributions were both coupled and (ii) on varying DPw,Lp,e can be easily neglected with regards to Lp,i. Theconformation of chitosan according to DA and DPw was thusexplained as follows. (i) Below a critical DPw (ranging from1300 to 1800, depending on DA) and for low DAs, therigidity of the chain was essentially due to short interactionsbetween adjacent protonated amino groups. On increasingDA, hydrophobic and steric interactions represented by #st-(DA,DPw) first counterbalanced (25% < DA < 50%) andthen overcame (DA > 50%) the polymer affinity with thesolvent, leading to the depletion of the chain. (ii) On theopposite, beyond DPw,c, the excluded volume effect wasassumed to lead to the different conformations of chitosanobserved in all of the ranges of DA. On increasing DA, theinfluence of #el(DA) was rapidly overcome by the increaseof #st(DA,DPw), because of the addition of new bulkyGlcNAc units along the chain. Concomitantly, as DAincreased, #st(DA,DPw) was also responsible for the slightshift of DPw,c to lower values, characterizing the change offlexibility of the chains versus their length. Thus, onincreasing DPw from 650 to 2600, the enhanced flexibilitywhich would have been the result of a chain length increasein the case of a Gaussian coil (DA < 25%) is thencounterbalanced (25 < DA < 50%) or even overcome (DA> 50%) by the steric contribution of the global excludedvolume effect. This subsequently led to the increase of thechain stiffness observed.To conclude, the continuous increases of #tot on increasing

both DA and DPw were responsible for the self-associationof chitosan chains revealed by curved shapes of Zimm plots,a drastic increase of RG,z and a drop of A2. Their presencecould lead to a colloidal dispersion, far from the dilute systemtheory and models we used. The radii of gyration as well asthe persistence lengths we determined could be then not well-representative of isolated chains in solution and lead to anoverestimated apparent rigidity. Even though the concentra-tion of these aggregates is rather low (<7%) to provoke afake conformation, on the other hand, it is sufficiently highto locally modify the critical concentration of chain entangle-ment C*, which was calculated in the case of a flexible andisolated chain. As a consequence, the theories and thecharacteristic dimension lengths describing the rigidity ofthe chains could be not well-adapted anymore: Lp,tot wouldbecome locally meaningless regarding the correlation length', generally used for a semidilute system for a blobrepresentation. Thus, the universal law of behavior firstpresented by Sorlier et al.,19 then confirmed by Schatz etal.,18 should rather be presented as underscoring two

domains: (i) For DA < 25%, the chitosan chains are quiteisolated, rather flexible and the theories related to the neutralpolymers in dilute solution can be used only in the case ofDPw > 1300. For DPw < 1300, the electrostatic contributionto the excluded volume must be taken in consideration. (ii)For DA > 50%, the steric contribution to the global volumeeffect contributed to the chitosan chains rigidity. The chainsbecome also mainly hydrophobic and a few of them undergoaggregation, which could change the theories involved todescribe a local and semidilute system. Between, we observea transition domain corresponding to a meta-stable systemin which the whole interactions and the thermal agitationcounterbalanced each other more or less. It is then prepon-derant to consider that the DA range generally studied ()[20;50]) in previous published works corresponded approxi-mately to the transition domain of our universal law ofbehavior. Added to the fact that the important Ips of thechitosan samples analyzed in the literature tended to leveloff the variations of the conformational parameters, this couldexplain why most of the previous studies reported R, ", andLp,tot as being constant according to DA. Thanks to this work,it is now worth considering the DPw when analyzing asolution property of chitosan.

Acknowledgment. These studies are taking part of theCARAPAX project from the 5th European frameworkProgram “Quality of Life and Management of LivingResources”.

References and Notes(1) Sannan, T.; Kurita, K.; Iwakura, Y. Makromol. Chem. 1975, 176,

1191.(2) Kurita, K.; Tomita, K.; Tada, T.; Ishii, S. J. Polym. Sci. Part A 1993,

31, 485.(3) Tolaimate, A.; Desbrieres, J.; Rhazi, M.; Alagui, A.; Vincendon, M.;

Vottero, P. Polymer 2000, 41, 2463.(4) Chang, K. L. B.; Tsai, G.; Lee, J.; Fu, W.-R. Carbohydr. Res. 1997,

303, 327.(5) Hirano, S. In Chitin and Chitosan; Skjåk-Braek, G., Anthonsen, T.,

Sandford, P., Eds.; Elsevier Applied Science: New York, 1989; p37.

(6) Sandford, P. In Chitin and Chitosan; Skjåk-Braek, G., Anthonsen,T., Sandford, P., Eds.; Elsevier Applied Science: New York, 1989;p 51.

(7) Vander, P.; Vårum, K. M.; Domard, A.; El Gueddari, N. E.;Moerschbacher, B. Plant Physiol. 1998, 118, 1353.

(8) Peluso, G.; Petillo, O.; Ranieri, M.; Santin, M.; Ambrosio, L.; Calabro,D.; Avallone, B.; Balsamo, G. Biomaterials 1994, 15, 1215.

(9) Gerentes, P.; Vachoud, L.; Doury, J.; Domard, A. Biomaterials 2002,23, 1295.

(10) Horner, V.; Pittermann, W.; Wachter, R. In AdVances in ChitinScience; Domard, A., Roberts, G. A. F., Vårum, K. M., Eds.; JacquesAndre Publisher: Lyon, France, 1997; p 671.

(11) Janes, K. A.; Calvo, P.; Alonso, M. J. AdV. Drug DeliVery ReV. 2001,47, 83.

(12) Janes, K. A.; Alonso, M. J. J. Appl. Polym. Sci. 2003, 88, 2769.(13) Vila, A.; Sanchez, A.; Tobio, M.; Calvo, P.; Alonso, M. J. J.

Controlled Release 2002, 78, 15.(14) Cui, Z.; Mumper, R. J. J. Controlled Release 2001, 75, 409.(15) Muzzarelli, R. A. A. Carbohydr. Polym. 1996, 29, 309.(16) Sugano, M.; Watanabe, S.; Kishi, A.; Izume, M.; Ohtakara, A. Lipids

1988, 23, 187.(17) Chatelet, C.; Damour, O.; Domard, A. Biomaterials 2001, 22, 261.(18) Schatz, C.; Viton, C.; Delair, T.; Pichot, C.; Domard, A. Biomac-

romolecules 2003, 4, 641.(19) Sorlier, P.; Denuziere, A.; Viton, C.; Domard, A. Biomacromolecules

2001, 2, 765.(20) Sorlier, P.; Viton, C.; Domard, A. Biomacromolecules 2002, 3, 1336.(21) Sorlier, P., Ph.D. Thesis, Lyon, France, 2002.


(22) Hirano, S.; Nagao, N. Agric. Biol. Chem. 1989, 53, 3065.(23) Kauss, H.; Jeblick, W.; Domard, A. Planta 1989, 178, 385.(24) Schatz, C.; Pichot, C.; Delair, T.; Viton, C.; Domard, A. Langmuir

2003, 19, 9896.(25) Domard, A. Int. J. Biol. Macromol. 1987, 9, 98.(26) Roberts, G. A. F.; Wang, W. In AdVances in Chitin Science; Domard,

A., Jeuniaux, C., Muzzarelli, R. A. A., Roberts, G. A. F., Eds.;Jacques Andre Publisher: Lyon, France, 1996; p 279.

(27) Anthonsen, M. W.; Vårum, K. M.; Smidsrød, O. Carbohydr. Polym.1993, 22, 193.

(28) Wang, W.; Bo, S.; Li, S.; Qin, W. Int. J. Biol. Macromol. 1991, 13,281.

(29) Berth, G.; Dautzenberg, H. Carbohydr. Polym. 2002, 47, 39.(30) Errington, N.; Harding, S. E.; Vårum, K. M.; Illum, L. Int. J. Biol.

Macromol. 1993, 15, 113.(31) Tsaih, M. L.; Chen, R. H. Int. J. Biol. Macromol. 1997, 20, 233.(32) Kasaai, M. R.; Arul, J.; Charlet, G. J. Polym. Sci. Part B 2000, 38,

2591.(33) Berth, G.; Dautzenberg, H.; Peter, M. G. Carbohydr. Polym. 1998,

36, 205.(34) Pa, J.-H.; Yu, T. L. Macromol. Chem. Phys. 2001, 202, 985.(35) Buhler, E.; Rinaudo, M. Macromolecules 2000, 33, 2098.(36) Colfen, H.; Berth, G.; Dautzenberg, H. Carbohydr. Polym. 2001, 45,

373.(37) Wu, C.; Zhou, S.; Wang, W. Biopolymers 1995, 35, 385.(38) Signini, R.; Desbrieres, J.; Campana Filho, S. P. Carbohydr. Polym.

2000, 43, 351.(39) Chen, R. H.; Chang, J. R.; Shyur, J. S. Carbohydr. Res. 1997, 299,

287.(40) Chen, R. H.; Shyur, J. S.; Chang, J. R. In AdVances in Chitin Science;

Domard, A., Roberts, G. A. F., Vårum, K. M., Eds.; Jacques AndrePublisher: Lyon, France, 1997; p 437.

(41) Chen, R. H.; Chen, J. S. In AdVances in Chitin Science; Peter, M.G., Domard, A., Muzzarelli, R. A. A., Eds.; University of Potsdam:Potsdam, 1999; p 361.

(42) Anthonsen, M. W.; Vårum, K. M.; Hermansson, A. M.; Smidsrød,O.; Brant, D. A. Carbohydr. Polym. 1994, 25, 13.

(43) Tsaih, M. L.; Chen, R. H. J. Appl. Polym. Sci. 1999, 71, 1905.(44) Dautzenberg, H.; Rother, G. Makromol. Chem., Macromol. Symp.

1992, 61, 94.(45) Terbojevich, M.; Cosani, A.; Conio, G.; Marsano, E.; Bianchi, E.

Carbohydr. Res. 1991, 209, 251.

(46) Terbojevich, M.; Cosani, A.; Focher, B.; Naggi, A.; Torri, G.Carbohydr. Polym. 1992, 18, 35.

(47) Smidsrød, O.; Haug, A. Biopolym. 1971, 10, 1213.(48) Roberts, G. A. F.; Domszy, J. G. Int. J. Biol. Macromol. 1982, 4,

374.(49) Rinaudo, M.; Milas, M.; Dung, P. L. Int. J. Biol. Macromol. 1993,

15, 281.(50) Domard, A.; Rinaudo, M. Polym. Commun. 1984, 25, 55.(51) Ottøy, M. H.; Vårum, K. M.; Christensen, B. E.; Anthonsen, M. W.;

Smidsrød, O. Carbohydr. Polym. 1996, 31, 253.(52) Tsaih, M. L.; Chen, R. H. J. Appl. Polym. Sci. 1999, 73, 2041.(53) Brugnerotto, J.; Desbrieres, J.; Roberts, G. A. F.; Rinaudo, M.

Polymer 2001, 42, 9921.(54) Beri, R. G.; Walker, J.; Reese, E. T.; Rollings, J. E. Carbohydr. Res.

1993, 238, 11.(55) Lamarque, G.; Viton, C.; Domard, A. Biomacromolecules 2004, 5,

992.(56) Brugnerotto, J.; Desbrieres, J.; Heux, L.; Mazeau, K.; Rinaudo, M.

Makromol. Symp. 2001, 168, 1.(57) Hirai, A.; Odani, H.; Nakajima, A. Polym. Bull. 1991, 26, 87.(58) Vårum, K. M.; Anthonsen, M. W.; Grasdalen, H.; Smidsrød, O.

Carbohydr. Res. 1991, 211, 17.(59) Philippova, O. E.; Volkov, E. V.; Sitnikova, N. L.; Khokhlov, A.

R.; Desbrieres, J.; Rinaudo, M. Biomacromolecules 2001, 2, 483.(60) Brandrup, J.; Immergut, E. H. Polymer Handbook; John Wiley &

Sons: Eds; New York 1989.(61) Dautzenberg, H.; Jaeger, W.; Kotz, J.; Philipp, B.; Seidel, C.;

Stscherbina, D. Polyelectrolytes: Formation, Characterization andApplication; Carl Hanser Verlag: Munchen, Germany, 1994.

(62) Sorlier, P.; Rochas, C.; Morfin, I.; Viton, C.; Domard, A. Biomac-romolecules 2003, 4, 1034.

(63) Odijk, T.; Houwaart, A. C. J. Polym. Sci., Polym. Phys. Ed. 1978,16, 627.

(64) Benoit, H.; Doty, P. J. Phys. Chem. 1953, 57, 963.(65) Mazeau, K.; Perez, S.; Rinaudo, M. J. Carbohydr. Chem. 2000, 19,

1269.(66) Manning, G. S. J. Chem. Phys. 1969, 51, 924.

BM0496357


lamarque 2005

Documents