Typical Physicochemical Behaviors of Chitosan in Aqueous Solution

Biomacromolecules 2003, 4, 641-648

641

Typical Physicochemical Behaviors of Chitosan in Aqueous Solution Christophe Schatz,†,‡ Christophe Viton,‡ Thierry Delair,† Christian Pichot,† and Alain Domard*,‡ Unite´ mixte CNRS Biome´ rieux, UMR 2142, ENS Lyon, 46, Alle´ e d’Italie, 69 364 Lyon Cedex 07, France, and Laboratoire des Mate´ riaux Polyme` res et des Biomate´ riaux, UMR CNRS 5627, ISTIL, Domaine Scientifique de la Doua, 15 Bd. Latarjet, 69 622 Villeurbanne Cedex, France Received November 7, 2002; Revised Manuscript Received January 15, 2003

Physicochemical properties of a homogeneous series of chitosans with different degrees of acetylation and almost the same degree of polymerization were investigated in an ammonium acetate buffer. Techniques such as interferometry, static light scattering (in batch or coupled on line with a chromatographic system), and viscometry were processed. All of the results agree with a unique law of behavior only depending on the degree of acetylation of the polymer. Indeed, values of the refractive index increment, radius of gyration, second virial coefficient, and intrinsic viscosity are decreasing in the same way as DA is increasing. Three distinct domains of DA were defined and correlated to the different behaviors of chitosans: (i) a polyeletrolyte domain for DA below 20%; (ii) a transition domain between DA ) 20% and 50% where chitosan loses its hydrophilicity; (iii) a hydrophobic domain for DAs over 50% where polymer associations can arise. Conformations of chitosan chains were studied by the calculations of the persistence lengths (Lp). The average value was found to be close to 5 nm, in agreement with the wormlike chain model, but no significant variation of Lp with the degree of acetylation was noticed. Introduction In the field of water-soluble natural polymers, chitosan, a copolymer of (1f4)-2-amino-2-deoxy-β-D-glucan (GlcN) and (1f4)-2-acetamido-2-deoxy-β-D-glucan (GlcNAc), was the subject of numerous works concerning its molecular characteristics in aqueous solution. Commercial chitosans are mainly produced from partial N-deacetylation of chitin under alkaline conditions.1 The degree of acetylation (DA) corresponds to the molar fraction of acetylated units within the polymer chains. Chitosan is a very promising glycosaminoglycan for biological applications such as drug delivery systems2 where the knowledge of parameters such as molecular weight polydispersity, chain dimensions, and chain conformations is of major interest. The characterization of chitosan is usually carried out in dilute acidic solutions where it dissolves thanks to the protonation of the amino functions. Depending on DA, the value of the intrinsic pK (pK0) varies within 6.46 and 7.32.3,4 In these conditions chitosan behaves as a relatively weak polycationic polyelectrolyte. Nevertheless, addition of salts is necessary to screen electrostatic repulsions between ionic sites along the chains. Among the analytical techniques allowing the investigation of chitosan behaviors, viscometry5,6 and static and dynamic light scattering7-10 are the most commonly used. For all of them, a good solubility of the polymer is required. According to the origin of chitosan, this is not always the case, especially with most of the com* Corresponding author. E-mail: [email protected]. † Unite ´ mixte CNRS Biome´rieux. ‡ Laboratoire des Mate ´ riaux Polyme`res et des Biomate´riaux.

mercial products. Indeed, chitosans obtained by heterogeneous deacetylation from chitin have a tendency to form aggregates in solution.11 This phenomenon can be attributed to the presence of residual sequences of acetylated residues along the chains, thus favoring hydrophobic interactions and hydrogen bonding. Therefore, for the study of chitosan in solution, it seems better to use a homogeneous series of acetylated chitosans related to a random distribution of the two kinds of repeating units. Changes in the molecular distributions represented by the polydispersity index can also constitute a major drawback for the study of laws of behavior of chitosans. Then, techniques of molecular separation such as size exclusion chromatography (SEC) may be used to study chitosan solutions corresponding to narrow molecular weight distributions. The conformation and the size of chitosan chains in dilute solutions are strongly dependent on the degree of acetylation and the ionic strength. Then, at the lower values of DA, electrostatic repulsions between protonated amino groups are predominant, leading to an expansion of the chains. However, the observed stiffness for low DAs is also dependent on the ionic strength. Indeed, salt addition has a screening effect on electrostatic repulsions, which are largely suppressed at high salt concentrations. Nevertheless, one cannot consider that chitosans are in theta conditions at infinite ionic strength. Indeed, even at infinite ionic strength, we may assume that the electrostatic contribution to the excluded volume is not completely inhibited.12 This is supported by results of viscometry5 showing that whatever DA within 15-60%, the exponent “a” (see below) is largely over 0.5. As DA is increasing, different kinds of interactions must be considered.

10.1021/bm025724c CCC: $25.00 © 2003 American Chemical Society Published on Web 02/27/2003

642

Biomacromolecules, Vol. 4, No. 3, 2003

Then, acetylated units can be involved in intramolecular hydrogen linkages, thus limiting the rotation around the β(1f4) glycosidic bonds. Furthermore, acetamido groups are more bulky than amino functions and hence are responsible for a steric hindrance also limiting the chain rotations.13 As a consequence, the stiffness of the chains should increase with DA. As already mentioned above, a large amount of acetyl groups along the chain can also favor the formation of hydrophobic interactions. The role of the degree of acetylation on the chain stiffness has essentially been assessed by viscometry considering the Mark-HouwinkKuhn-Sakurada relationship (MHKS), [η] ) KMa, and by static light scattering through the relation RG ) K′Mν, where [η] and RG correspond to the intrinsic viscosity and the radius of gyration, respectively. Experiments have been performed either with parent samples having different degrees of polymerization6 or with samples fractionated by size exclusion chromatography.14 Many authors found an increase of the a coefficient with DA and concluded that acetyl groups induce some stiffness to the chains5,6,15 whereas the dependence of ν on DA is less significant.14-16 Calculations of the persistence length (Lp) were also used by some authors to deduce the effect of an increase of the acetyl content on the chain conformation. But, the values of Lp were rarely calculated on a large range of DA, and then some uncertainties remain about its variation with DA. With several commercial chitosans, Terbojevich et al.17 found Lp ) 22 nm for DA ) 15% and 42%, Berth et al.7 found Lp ) 6 nm for a DA close to 25%, whereas with homogeneously acetylated chitosans Brugnerotto et al.16 found a slight increase of Lp from 11 to 15 nm for DAs of 2% and 60%, respectively. These results suggest that the model of the “flexible wormlike chains” can describe the conformation of chitosan in solution. Nevertheless, its dependence with the degree of acetylation is still a current topic of research. The origin of chitosan seems to explain some divergent results in the literature. The aim of this present work was to study physicochemical characteristics of chitosan solutions in the largest range of DA as possible, allowing us to determine laws of behavior. Contrary to most of the cases reported in the literature, we worked with a homogeneous series of polymers. Thus, six water-soluble chitosans with DAs between 1% and 71% obtained by homogeneous acetylation of the same initial sample of low DA were either analyzed by static light scattering (in batch and coupled on line with steric exclusion chromatography) and by capillary viscometry. The dimension and conformation parameters resulting from these measurements should enable us to understand the behavior of chitosan chains as a function of DA. Experimental Section Purification and Acetylation of Chitosan. Chitosan (batch A32E03) with a DA of 1% was purchased from Aber Technologies (France). To obtain a high-purity material, chitosan was successively dissolved at 1% (w/v) in a 0.1 M acetic acid solution, filtered trough membranes (Millipore) having a decreasing porosity from 3 to 0.22 µm, reprecipi-

Schatz et al.

tated with dilute ammonia, extensively rinsed with deionized water until neutral pH, and finally freeze-dried. Acetylation was performed according to the procedure described by Vachoud et al.18 Briefly, purified chitosan was dissolved in a 0.1 M acetic acid/1,2-propanediol mixture. A solution of acetic anhydride in 1,2-propanediol was added at different ratios according to the required DA. Then, the reaction medium was neutralized and the chitosan thus precipitated rinsed with deionized water and lyophilized. Five chitosans were acetylated in such a way, and their degrees of acetylation were found successively as 11.5%, 24%, 38%, 51%, and 71%. Characterization of Chitosan. After lyophilization, the degrees of acetylation were calculated from 1H NMR spectroscopy (Varian, 500 MHz) according to the method developed by Hirai et al.19 1H NMR also allowed us to check the absence of propanediol within acetylated chitosans. The water content was determined by thermogravimetric analysis (DuPont Instrument 2950). Static Light Scattering (SLS). Multiangle laser light scattering (MALLS) detection was used in batch and coupled on line with a chromatographic system. Batch measurements were performed with a Dawn DSP-EOS (Wyatt) equipped with a 25 mW Ga/As laser operating at λ ) 690 nm. Solutions were analyzed in scattering glass cells, and scattered light intensities were measured at 18 angles between 23° and 147°. Light intensity measurements were derived following the classical Rayleigh-Debye equation, allowing us to deduce Mw, the weight-average molecular weight, RG,z, the root-mean-square z-average of the gyration radius, and A2, the second virial coefficient by means of Zimm plots. Refractive Index Increment (dn/dc). dn/dc was evaluated independently for each DA with a differential interferometer operating at λ ) 632.8 nm (NFT Scan Ref). Five solutions (a parent solution and four dilutions) were analyzed to determine each value of dn/dc. The interferometer was also used to control the possible loss of matter during the filtration stage and to evaluate effects of dialysis and filtration on dn/ dc values. High-Performance Size Exclusion Chromatography (HP-SEC). The polymer separation was performed on two serially connected columns (TSK G3000-PW and TSK G6000-PW, i.d. ) 7.8 mm, l ) 300 mm). The detection was 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 λ ) 632.8 nm. Analyses were performed in microbatch mode using the K5 flow cell. A degassed 0.2 M acetic acid/0.15 M ammonium acetate buffer (pH 4.5) was used as eluent after filtration on a 0.22 µm pore size membrane (Millipore). The flow rate was maintained at 0.5 mL/min, and the amount of sample injected was 100 µL at a concentration of 0.1% (w/w). Viscometry. Measurements were performed at 25 °C with an Ubbelohde capillary viscometer (Viscologic TI.1, SEMATech) with an inner diameter of 0.53 mm. The intrinsic viscosity [η] was calculated by extrapolating to zero concentration either the Huggins or the Kraemer equations. As differences between the [η] values obtained from these

Physicochemical Behaviors of Chitosan

equations were smaller than measurement errors, intrinsic viscosities were expressed as the average of the two methods. Preparation of Solutions. A 0.2 M acetic acid/0.15 M ammonium acetate buffer (pH ) 4.5) was used as solvent for both the preparation of polymer solutions and all dilutions. Prior to use, it was filtered twice on 0.22 µm pore size membranes. Taking into account their water contents, chitosans of varying DAs were dissolved in this solvent at a concentration of 1 mg/mL. Before light scattering and viscometry experiments solutions of chitosans were successively filtered on 0.45 and 0.22 µm pore size membranes (Millipore). Results and Discussion The purpose of this work was to investigate the laws of behavior related to the physicochemical properties of chitosan in aqueous solutions in relation with the degree of acetylation (DA). In this kind of study, the use of homogeneous samples having the same origin and a high purity is of great importance. This is why all of the samples used in this work were acetylated in homogeneous conditions from the same purified chitosan having initially a DA of 1%. Acetylation was performed with acetic anhydride in a hydro-alcoholic media. Such conditions lead to a random distribution of acetyl groups along the polymer chains and preserve their structures against degradation. Characterization of chitosan is often performed in acetic acid/sodium acetate solutions, and small amounts of sodium chloride are sometimes added in order to adjust the ionic strength.7 In our case, an acetic acid/ammonium acetate buffer (pH ) 4.5) was preferred for its ability to break hydrogen bonding. This solvent first suggested by Domard et al.20 as an eluent for HP-SEC analyses was also chosen by Ottoy et al.21 As, whatever the DA, the pH of the buffer is largely below the intrinsic pK,4 we may consider that, in all cases, we are very close to the full ionization. Solubility of Chitosan Samples. As for most of the natural polymers, chitosan is well-known for its tendency to self-aggregate in aqueous solution. Evidence of these macromolecular associations was underscored by Anthonsen et al.,11 who, by static light scattering, studied chitosans having different DAs and molecular weights. Especially at high DAs, their Zimm plots revealed a positive curvature of the curves representing the concentration dependence, thus leading to negative values of the second virial coefficient. Such a phenomenon is known to be characteristic of a process of concentration-dependent aggregation. The mechanism of formation of these associations is not yet well understood, but some authors stated that hydrophobic moieties in chitosan, i.e., acetyl groups and glucosidic rings, play a significant role on aggregation21,22 in the formation of hydrophobic interactions. Thanks to the use of a fluorescent probe interacting with hydrophobic regions, Philippova et al.23 revealed the presence of aggregates from an initial concentration of chitosan close to 1 mg/mL. By different ways, such as addition of urea or ethanol, increase of temperature, and variation of ionic strength, the authors tried to elucidate the nature of these aggregates. They concluded

Biomacromolecules, Vol. 4, No. 3, 2003 643

that neither H-bonds nor hydrophobic interactions were responsible for aggregation of chitosan, and the debate remains still open. To remove these aggregates, clarification of the solutions is an important step prior to characterization;24,25 filtration and/or ultracentrifugation are currently used. In this work, all chitosan solutions were successively filtered on 0.45 and 0.22 µm pore size membranes. The high solubility of chitosan samples in AcOH/AcNH4 buffer and the absence of aggregates were verified by different methods. (I) The loss of matter during the filtration was evaluated by interferometric measurements on filtered solutions after dialysis in the same solvent. Whatever the degree of acetylation (except for DA ) 71%; see below), less than 2% of polymer was retained after the two successive filtrations, which means that chitosan is highly soluble in the buffer. (II) HP-SEC chromatograms display a Gaussian shape for all acetylated chitosans, and no bimodal distribution, characteristic for high molecular weight aggregates in solution, appeared. In addition, the recovery rates were over 95%, except for DA ) 71% (see below), indicating a weak adsorption of polymer on the stationary phase. (III) All solutions of chitosan seemed to be free of concentration-dependent aggregates since all Zimm plots revealed no typical curvature of the concentration dependence curves (Figure 1). (IV) Finally, all values of the second virial coefficients were over 10-3 mol‚mL/g2 characteristic for a polyelectrolyte in a good solvent.15 Refractive Index Increment dn/dc. Accurate determination of the refractive index increments (dn/dc) is required for the evaluation of the weight-average molecular weights and the second virial coefficients, especially because dn/dc arises at a power of 2 in the Rayleigh-Debye equation. Measurements of the refractive index increments of polyelectrolytes are usually performed after dialysis of the polymer solutions in the solvent used for their preparation. We noticed that the same result could be obtained with the same nonfiltered solution of chitosan before or after dialysis in the ammonium acetate buffer used as solvent. On the contrary, the phase shift between laser beams of reference and sample cells was decreasing 10% and 30% after filtration on 0.45 and 0.22 µm pore size membranes, respectively. Conductivity was also decreasing after filtration. Inversely, only a small reduction of intensity was detected by both the refractometric and MALLS detections after a chromatographic separation of these samples. Then, the losses of interferometric responses after each filtration were rather due to a modification of the chemical potential of chitosan solutions by dissolution of CO2 into the solution during the filtration process than to a loss of polymer matter. Indeed, the dialysis of filtered solutions in ammonium acetate buffer allowed us to fully recover the same interferometric response as with initial solutions before filtration. As a consequence, in this work, all of the dn/dc of chitosan solutions were measured on solutions without any filtration. In the literature, the variations of dn/dc with DA were currently not clearly demonstrated. On one hand, some authors state that this parameter is independent of DA15,26, but the range of investigated DA values is often sharp, and moreover, their chitosans are commercial samples of various

644


Schatz et al.

Figure 1. Zimm plots for various chitosans in solution in an acetic acid/ammonium acetate buffer (µ ) 0.15 M): (a) DA ) 1%; (b) DA ) 51%; (c) DA ) 71%.

origins and molecular distributions, corresponding to polymers obtained by a heterogeneous deacetylation. On the other hand, some authors such as Terbojevich27 or Wang6 found a significant decrease of dn/dc with the degree of acetylation for commercial and deacetylated chitosans, respectively. In our case, working with a homogeneous series of chitosans covering all of the range where chitosan can be considered as soluble, we had a random distribution of the acetyl groups and a same molecular distribution of the polymer chains. Therefore, DA was actually the only variable. The results reported in Figure 2 exhibit a strong dependence between dn/dc and DA. This variation can be simply explained if we consider that this parameter is related to the electric polarizability of chitosan chains, highly sensitive to the charge density. The latter depends on both the distance between two ionizable groups and their ionization state as related to their intrinsic pK (pK0). The law of behavior

illustrated on Figure 2 is quite similar to the general law of behavior we may now observe in all of the studies where we investigate the variation of a physicochemical property of chitosan, which depends on both the structural charge density and the ionization state of the polymer.4,28 Indeed, the curve can be divided into three domains. (1) At DAs below 20%, dn/dc exhibits high values which correspond to a polyelectrolyte behavior where electrostatic interactions are predominant but decrease with DA in relation with the decrease of the apparent charge density. (2) For DAs over 50%, the decrease of dn/dc must be related to both a decrease of the structural charge density and an increase of the hydrophobic domains.4,28 It is worth noting that a chitosan of DA ) 71% has a dn/dc close to 0.15 mg/mL, characteristic of the values obtained for neutral polysaccharides such as dextrans or pullulans in aqueous solvents.29 (3) The interval of DA values between 20% and 50% corresponds to the



Figure 2. Variation of dn/dc, the refractive index increment of chitosans in acetic acid/ammonium acetate buffer (µ ) 0.15 M) at λ ) 632.8 nm.

Figure 3. Dependence of the second virial coefficient on the degree of acetylation of chitosan determined by Zimm plot in solution in an acetic acid/ammonium acetate buffer (µ ) 0.15 M).

Table 1. Molecular Characteristics of Chitosans of Various DAs Determined in Batch Mode and from Chromatographic Experimentsa unfractionated samples DA (%)

Mw (105 g/mol)

1 11.5 24 38 51 71

1.62 ( 0.04 1.71 ( 0.02 1.68 ( 0.02 1.73 ( 0.02 1.87 ( 0.04 2.72 ( 0.07

a

fractionated samples

DPw

Mw (105 g/mol)

DPw

Ip

1000 ( 30 1020 ( 10 980 ( 10 980 ( 10 1030 ( 20 1430 ( 40

1.28 ( 0.03 1.38 ( 0.02 1.42 ( 0.02 1.45 ( 0.03 1.58 ( 0.03 1.88 ( 0.04

790 ( 20 830 ( 10 830 ( 10 820 ( 20 870 ( 20 990 ( 20

1.64 ( 0.06 1.46 ( 0.05 1.36 ( 0.04 1.37 ( 0.04 1.35 ( 0.05 1.26 ( 0.03

Ip corresponds to the polydispersity index.

transition between hydrophilic and hydrophobic behaviors. In this domain, dn/dc is close to 0.185 mL/mg. Some authors working on commercial chitosans often located in this range of DA quoted such a value.26,30 Static Light Scattering and Viscometry. (A) Experiments Performed in Batch. As shown in Figure 1, whatever the degree of acetylation, static light scattering provides us with well-defined Zimm plots. Thus, we may consider that two successive filtrations of the polymer solutions on 0.45 and 0.22 µm pore size membranes are sufficient to clarify the solutions of chitosan. Only very slight curvatures, at the low scattering angles, were detected for the highest concentrations (C ) 0.8 and 1 mg/mL), especially for chitosans of high DA (Figure 1). This tendency can be attributed to small amounts of aggregates but has no significant influence on the results since similar values of Mw, RG,z, and A2 were obtained by considering all five concentrations or by omitting the two highest ones. Nevertheless, in the following, all Zimm plots were derived with the five concentrations. Normally, the weight-average molecular weight, Mw, slowly increases with the degree of acetylation although the weight-average degree of polymerization, DPw, remains more or less constant. This result agrees with conditions of acetylation sufficiently soft to preserve the integrity of polymer chains against degradation (Table 1). The nonnegligible drop of DPw observed with chitosan of DA ) 71% reveals the presence of aggregates. High contents of acetylated moieties are probably at the origin of these polymer associations by hydrophobic interactions preferentially to

Figure 4. Root-mean-square z-average of the gyration radius determined by static light scattering in batch mode (9) and coupled with the chromatographic system (4). Solutions of chitosan in an acetic acid/ammonium acetate buffer (µ ) 0.15 M) were filtered on 0.45 and 0.22 µm pore size membranes before analysis.

hydrogen bonding. Indeed, H-bonds should be less involved in this kind of self-association since ammonium acetate is generally considered as preventing their formation. The variations of A2 and RG,z with the degree of acetylation are plotted in Figures 3 and 4, respectively. The decrease of the second virial coefficient with the degree of acetylation is to be related with the increase of the segment/segment interactions to the detriment of the interactions between the chain segments and the solvent. This loss of solubility can be explained by a decrease of the hydrophilicity of the polymer chains in favor of hydrophobic interactions brought about by the increase of the number of N-acetyl residues along the chains. The variation of A2 on increasing DA must also be related with the decrease of the electrostatic excluded volume representing the major contribution of the total excluded volume for polyelectrolytes. To our knowledge, such a dependence of A2 with DA was never pointed out. With homogeneous acetylated chitosans Berth et al.14 obtained an irregular decrease of A2 with DA, but only

646


Figure 5. Variations of the intrinsic viscosity (9) and the Huggins constant (4) with DA in solution in an acetic acid/ammonium acetate buffer (µ ) 0.15 M).

soluble fractions of the polymer were considered since their samples exhibited a poor solubility in the acetate buffer used as solvent. The variations of the values of the gyration radius follow the same trend. For DAs ) 1% and 11.5%, the relatively high values of RG,z correspond to the polyelectrolyte behavior where intramolecular repulsions between adjacent ionic sites lead to an expansion of the chains. Between DA ) 24% and 51%, the dimensions do not really increase, whereas for DA ) 71%, the self-association of chitosan chains mentioned above leads to a significant increase of the gyration radius. Nevertheless, this increase must be regarded as a sort of artifact. Indeed, after an additional filtration of the chitosan solution of DA ) 71% on a 0.22 µm pore size membrane, RG,z decreased from 51.6 to 42 nm with a loss of matter equal to 11%. This value of RG,z becomes below that observed at DA ) 51% and agrees with a continuous depletion of the free chains on increasing DA. This possibility of removing aggregates by filtration signifies that these structures are relatively stable and do not depend on the polymer concentration. Moreover, the value of the second virial coefficient for this DA still remains far from theta conditions and confirms their relative stability.11 For chitosans having a DA between 1% and 51% the same additional filtration on a 0.22 µm pore size membrane only led to a weak loss of matter and decrease of RG,z of no more than 2% and 2 nm, respectively, thus confirming that a chitosan of DA ) 71% is partially aggregated. Viscometric measurements were performed with the same solutions of chitosan as those used for static light scattering. Since the chitosans had almost the same DPw, the differences of viscometric behaviors could actually be explained by changes of chain dimensions related to the balance between hydrophilic and hydrophobic interactions. As mentioned above, the steady decrease of A2 is linked to an enhancement of the segment/segment contacts and therefore also to a decrease of the chain expansion. This was especially revealed by the values of RG,z in the range of DA within 1-50%. The expected decrease of the intrinsic viscosity with DA was also observed experimentally (Figure 5). It is interesting to notice the similarity between all of the plots of Figures 2-5, thus confirming a general law of behavior when we study

Schatz et al.

the variation of a given property of chitosan with DA. This important result has already been suggested above and in our other studies4,28 made on this kind of homogeneous series of chitosans. At low values of DA, chitosan exhibits a typical polyelectrolyte behavior revealed by high values of the intrinsic viscosity. Expansion of the chains by intramolecular electrostatic repulsions (tertiary electroviscous effect31,32) is certainly the most obvious reason to explain the increase of the viscosity, but we must also consider the importance of the primary electroviscous effect that affects directly the intrinsic viscosity. This effect, independent of the polymer concentration, corresponds to the deformation by the flow of the surrounding electric double layer of each polymer chain assimilated to rigid particles.31,32 The energy dissipated by this deformation leads to a significant increase of viscosity in comparison with neutral polymers. In the range located within DA 20-50%, no significant variation of the viscosity was observed, showing thus that, in this range, hydrophilic and hydrophobic effects on chain dimensions are equilibrated. Since both A2 and RG,z are also relatively constant in this region, we may consider that the decrease of hydrophilicity is opposed to a partial increase of the hydrophobicity brought about by the increase of the intrinsic pK of the amino groups of the glucosamine residues.4 For DA ) 71%, SLS has allowed us to point out the probable formation of polymer chain associations in relation with the drop of RG,z and DPw values. Here, the fall of the intrinsic viscosity at the same DA confirms the formation of highly compact structures that should really correspond to the formation of aggregates. Values of the Huggins constant, k′, are lower than those reported by different authors for equivalent DAs,14,33 which indicates a very high solubility of chitosan in ammonium acetate buffer even for high acetyl contents (Figure 5). The law of variation of k′ also reveals the solubility dependence of chitosan on the hydrophilic/hydrophobic balance. The values of intrinsic viscosities also allow us to determine C*, the critical concentration of chain entanglement. Thus, the values of C* calculated from the reverse of the intrinsic viscosities vary from 1.7 mg/mL for DA ) 1% to 2.0 mg/mL for DA ) 71%. This result demonstrates that the concentration we chose for the parent solutions (1 mg/ mL) is lower than C*, whatever the DA. This assumption is reinforced by the fact that, for all DAs, the experimental values illustrating both the Huggins and Kraemer equations agree quite well with linear variations, illustrating the absence of change in the regime of the solutions. (B) Experiments Performed by Chromatography. A chromatographic system coupled with a refractometer and a MALLS spectrometer on line provided us with the molecular mass and the gyration radius for each eluted fraction of polymer. Analyses were performed on the same samples as those used above for static light scattering and viscometric measurements. Except for DA ) 71%, the variations of Mw (Table 1) and RG,z (Figure 4) were very close to those obtained previously in batch mode. Nevertheless, values of Mw were lower probably because of retention of some large polymer associations on the columns. For the chitosan sample of DA ) 71%, this retention was the most pronounced since the rate of polymer recovery was only 83% compared to at



Table 2. Chain Stiffness Parameters of Chitosans Varying in DA Determined from Batch and Chromatographic Data persistence length (nm) DA (%)

ν ((0.01)

from batch data

from chromatographic fit

1 11.5 24 38 51 71

0.55 0.55 0.57 0.57 0.57 0.56

4.7 ( 0.6 4.3 ( 0.5 4.7 ( 0.5 5.0 ( 0.6 4.5 ( 0.5 5.0 ( 0.6

5.2 ( 0.3 4.6 ( 0.3 5.9 ( 0.4 5.8 ( 0.3 5.3 ( 0.4 4.5 ( 0.4

relation (eq 1) valid for monodisperse wormlike chains in theta conditions. L corresponds to the contour length of the polymer chain.

Figure 6. Molecular weight distributions of chitosans eluted on TSK gel columns in solution in an acetic acid/ammonium acetate buffer (µ ) 0.15 M).

least 95% for the other degrees of acetylation. It confirms that a part of the polymer was certainly aggregated in a low amount and then not really separated on the columns. This leads to lower values of RG,z (41 nm) and Mw in comparison with light scattering measurements in the batch mode. The decrease of polydispersity as DA increases (Table 1 and Figure 6) could appear as surprising since the acetylation reaction does not degrade chitosan chains. Two reasons may explain this phenomenon. First, the range of pH where chitosan is soluble is known to enlarge as DA increases. So, the numerous rinsing with deionized water after each acetylation can contribute to eliminate an increasing amount of low molecular weight polymer fractions as DA increases and thus induce a decrease of Ip. This behavior was not observed for another homogeneous series of chitosans of much higher molecular weights.4 Second, the increasing tendency to self-association with DA induces a growing retention of a few aggregates inside the chromatographic columns and a lesser efficiency of the chromatographic separation mentioned above. The exponent “ν” in the relation RG ) KMν allows us to deduce the type of chain conformation present in solution. Indeed, ν is increasing from 0.5, for Gaussian coils, in an unperturbed state, i.e., in theta conditions, to 1 for rigid rods. This coefficient is determined from chromatographic analyses by plotting log RG as a function of log M for each eluted fraction. The relation is linear between 5 × 104 and 106 g/mol for all degrees of acetylation, and in this range of molecular weights, values of ν are located between 0.55 and 0.57 without significant variation with DA (Table 2). Such values are close to those found by Berth et al.14 with homogeneous acetylated chitosans and are consistent with the wormlike chain model. Nevertheless, considering the weak variation of ν with DA and the nonnegligible experimental errors, we cannot conclude on a chain conformation dependence with DA. Persistence Length. The dependence of the chain stiffness on the degree of acetylation was also investigated by calculations of the persistence lengths (Lp) for each chitosan. Persistence lengths were deduced from the Benoit-Doty

2 RG,Θ

LLp 2L3p 2L4p 2 - Lp + - 2 (1 - e-L/Lp) ) 3 L L

(1)

For polyelectrolytes, the persistence length is the sum of two contributions: Lp,i, the intrinsic part, related to the stiffness of the backbone, and Lp,e, the electrostatic persistence length, related to the electrostatic charge effects on the chain rigidity.25 The latter was determined for each degree of acetylation, and it appears that its contribution to the global persistence length, in our experimental conditions, is negligible (Lp,e< 0.2 nm) even for a low DA, where the ionization degree of macromolecules is high. Berth et al.7 proposed a method to calculate Lp, taking into account the polydispersity of the system and the effects of excluded volume. The polydispersity was evaluated by a normalized weight distribution34 which, in our case, fits correctly the experimental weight distribution, and the influence of the excluded volume is estimated by the calculation of the Flory expansion coefficient R. An iterative procedure allows us to get Lp thanks to SLS results in batch mode, i.e., from Mw and RG,z and the polydispersity index deduced from chromatographic analyses. Thanks to this method, the values of the persistence lengths for all chitosans were found as located between 4.3 and 5 nm without significant variation upon changing DA, as previously observed for the coefficient ν (Table 2). These values are close to Lp ) 6 nm, the value found by Berth et al.,7 for commercial chitosans of DA close to 25%. Since the polydispersity index is a very sensitive variable for the determination of Lp, one can suppose that a weak difference between the experimental and normalized weight distribution may induce uncertainties. This is why we decided to also directly estimate the values from chromatographic data by fitting eq 1 to experimental RG(M) curves for each sample of chitosan.16 In this case, the polydispersity of the system is really considered, and only the contribution of the excluded volume must be added for each iteration. Using this second method, values of Lp were found between 4.6 and 5.9 nm, and there, also, the influence of DA on Lp remained limited (Table 2). If Lp and RG,z present a similar relative variation, experimental errors are much higher for the values of Lp than for those of RG,z. Thus, it is difficult to conclude on an influence of the acetyl content on the chain stiffness. The average persistence length deduced from the whole results leads to approximately Lp ) 5 nm. This value confirms that,

648


in our experimental conditions, whatever the degree of acetylation, chitosan chains adopt the conformation of a more or less flexible wormlike chain constituted by approximately 50 Kuhn segments of 10 nm. This behavior must also be related to the fact that chitosan chains never correspond to a strong polyelectrolyte since the apparent charge density is related to the presence of no more than one ionizable group every 0.51 nm. Conclusion By means of a homogeneous series of acetylated chitosans with almost the same molecular distribution, we studied the physicochemical properties of chitosan in solution as a function of only one structural parameter, the degree of acetylation. Depending on DA, interferometry, static light scattering, and viscometry underscored a triple behavior: (i) At DAs below 20%, electrostatic interactions are predominant, and chitosan behaves as a cationic polyelectrolyte characterized by a high expansion of the chains. (ii) For DA values between 20% and 50%, physicochemical parameters remain more or less constant in relation with the fact that hydrophilic and hydrophobic interactions are counterbalanced. (iii) For DA over 50%, associations of polymer chains lead to the formation of stable aggregates. Since the buffer contains ammonium acetate, hydrogen bonding cannot be actually largely involved in this aggregation phenomenon. Thus, hydrophobic interactions due to high acetyl content are probably responsible for the molecular associations. All of these results confirm previous conclusions of Sorlier et al.,4,28 who found similar characteristic behaviors of chitosan by studying the variation of pK0 with DA in 0.1 M KClO4. Thus, one can conclude that physicochemical properties of chitosan in solution may be simply represented by a single and general law of behavior which can be described as follows: when DA is increasing, properties of chitosan are changing through three distinct domains evolving from a polyelectrolyte state to that of isolated charges on polymer chains in a hydrophobic environment and, in between, a transition range. Acknowledgment. This work is financially supported by a grant from the Ministry of Research and Technologies. References and Notes (1) Roberts, G. A. F. Chitin chemistry; Macmillan Press Ltd.: London, 1992.

Schatz et al. (2) Janes, K. A.; Calvo, P.; Alonso, M. J. AdV. Drug DeliVery ReV. 2001, 47, 83. (3) Domard, A. Int. J. Biol. Macromol. 1987, 9, 98. (4) Sorlier, P.; Denuzie`re, A.; Viton, C.; Domard, A. Biomacromolecules 2001, 2, 765. (5) Anthonsen, M. W.; Varum, K. M.; Smidsrod, O. Carbohydr. Polym. 1993, 22, 00. (6) Wang, W.; Bo, S.; Li, S.; Qin, W. Int. J. Biol. Macromol. 1991, 13, 281. (7) Berth, G.; Dautzenberg, H.; Peter, M. G. Carbohydr. Polym. 1998, 36, 205. (8) Pa, J. H.; Yu, T. L. Macromol. Chem. 2001, 202, 985. (9) Kjoniksen, A. L.; Iversen, C.; Nystrom, B.; Nakken, T.; Palmgren, O. Macromolecules 1998, 31, 8142. (10) Buhler, E.; Rinaudo, M. Macromolecules 2000, 33, 2098. (11) Anthonsen, M. W.; Varum, K. M.; Hermansson, A. M.; Smidsrod, O.; Brant, D. A. Carbohydr. Polym. 1994, 25, 13. (12) Skolnick, J.; Fixman, M. Macromolecules 1977, 10, 944. (13) Errington, N.; Harding, S. E.; Varum, K. M.; Illum, L. Int. J. Biol. Macromol. 1993, 15, 113. (14) Berth, G.; Dautzenberg, H. Carbohydr. Polym. 2002, 47, 39. (15) Brugnerotto, J.; Desbrie`res, J.; Heux, L.; Mazeau, K.; Rinaudo, M. Macromol. Symp. 2001, 168, 1. (16) Brugnerotto, J.; Desbrie`res, J.; Roberts, G.; Rinaudo, M. Polymer 2001, 42, 9921. (17) Terbojevich, M.; Cosani, A.; Conio, G.; Marsano, E.; Bianchi, E. Carbohydr. Res. 1991, 209, 251. (18) Vachoud, L.; Zydowicz, N.; Domard, A. Carbohydr. Res. 1997, 302, 169. (19) Hirai, A.; Odani, H.; Nakajima, A. Polym. Bull. 1991, 26, 87. (20) Domard, A.; Rinaudo, M. Polym. Commun. 1984, 25, 55. (21) Ottoy, M. H.; Varum, K. M.; Christensen, B. E.; Anthonsen, M. W.; Smidsrod, O. Carbohydr. Polym. 1996, 31, 253. (22) Amiji, M. M. Carbohydr. Polym. 1995, 26, 211. (23) Philippova, O. E.; Volkov, E. V.; Sitnikova, N. L.; Khokhlov, A. R.; Desbrie`res, J.; Rinaudo, M. Biomacromolecules 2001, 2, 483. (24) Tabor, B. E. In Light Scattering from Polymer Solutions; Huglin, M. B., Ed.; Academic Press: New York, 1972, Chapter 1. (25) Berth, G.; Dautzenberg, H. Recent Res. DeV. Macromol. Res. 1998, 3, 225. (26) Beri, R. G.; Walker, J.; Reese, E. T.; Rollings, J. E. Carbohydr. Res. 1993, 238, 11. (27) Terbojevich, M.; Cosani, A.; Focher, B.; Naggi, A.; Torri, G. Carbohydr. Polym. 1992, 18, 35. (28) Sorlier, P.; Viton, C.; Domard, A. Biomacromolecules 2002, 3, 1336. (29) Theisen, A.; Johann, C.; Deacon, M. P.; Harding, S. E. RefractiVe increment data book for polymer and biomolecular scientists; Nottingham University Press: Nottingham, U.K., 2000. (30) Tshaih, M. L.; Chen, R. H. J. Appl. Polym. Sci. 1999, 71, 1905. (31) Domard, A. Thesis, University of Grenoble, France, 1976. (32) Jiang, L.; Yang, D.; Chen, S. B. Macromolecules 2001, 34, 3730. (33) Signini, R.; Desbrie`res, J.; Campana Filho, S. P. Carbohydr. Polym. 2000, 43, 351. (34) Dautzenberg, H.; Rother, G. Makromol. Chem., Macromol. Symp. 1992, 61, 94.

BM025724C

Typical Physicochemical Behaviors of Chitosan in Aqueous Solution

Recommend Documents