Structure of Natural Polyelectrolyte Solutions: Role of the Hydrophilic


Structure of Natural Polyelectrolyte Solutions: Role of the Hydrophilic...

0 downloads 72 Views 2MB Size

pubs.acs.org/Langmuir © 2009 American Chemical Society

Structure of Natural Polyelectrolyte Solutions: Role of the Hydrophilic/ Hydrophobic Interaction Balance Simina Popa-Nita,† Cyrille Rochas,‡ Laurent David,*,† and Alain Domard† †

Laboratoire des Mat eriaux Polym eres et Biomat eriaux, Universit e de Lyon, Universit e Lyon 1, UMR CNRS 5223 IMP, B^ at. ISTIL, 15, bd Latarjet, F-69622 Villeurbanne Cedex, France, and ‡ Laboratoire de Spectrom etrie Physique, Universit e Joseph Fourier de Grenoble, UMR CNRS 5588, ^ Postale 87, F-38402 St. Martin d’H Boite eres, France Received January 8, 2009. Revised Manuscript Received March 28, 2009

A method allowing the evaluation of the solvophilic/solvophobic character of polyelectrolytes from their conformation in solution is discussed. Analyzed systems are salt-free aqueous solutions of natural copolysaccharides with controlled chemical structures. Small-angle X-ray scattering diagrams revealed their conformation by the “polyelectrolyte peak”. The study of this peak allowed the determination of cb, the crossover concentration associated with the transition between the two structural organization regimes predicted by the scaling model of hydrophobic polyelectrolytes developed by Dobrynin and Rubinstein.1 A structural law of behavior as a function of the chain primary structure is built for chitosan, showing an increasing hydrophobic character when the fraction of N-acetyl-D-glucosamine residues (DA) increases. The results concerning this random copolymer are compared with those obtained for hyaluronan. Consistently, in the case of alginates, the relative content of the constitutive units is shown not to influence the polymer hydrophobicity.

I. Introduction Polyelectrolyte polymers are macromolecular species bearing ionizable groups along their backbone. In polar solvents (generally water), polyelectrolyte salts dissociate into charged polymer chains and counterions dispersed in the solution or condensed close to their surface.2 The considerable interest about these typical polymers during the past decades3,4 is due not only to the need to better understand their specific behavior but also to their numerous and diversified applications.5-8 The reference synthetic system widely studied is the poly(styrene-co-styrenesulfonate, sodium salt) (PSSNa).9 Natural polyelectrolytes have been less explored due to the difficulty to obtain series of homogeneous samples in chemical structure and dimension. Nevertheless, cationic and anionic polysaccharides like chitosans,10 carrageenans,11 hyaluronans,12 and alginates13 have been investigated. Such amphiphilic polymers must be considered as relatively hydrophobic since water constitutes a poor solvent in which they *Corresponding author. E-mail: [email protected]. (1) Dobrynin, A. V.; Rubinstein, M. Macromolecules 1999, 32, 915–922. (2) Manning, G. S. J. Chem. Phys. 1969, 51(3), 924–934. (3) Barrat, J. L.; Joanny, J. F. Theory of Polyelectrolyte Solutions. In Advances in Chemical Physics, Polymeric Systems; Progine, S. A., Ed.; John Wiley and Sons: New York, 1996; Vol. XCIV. :: (4) Dautzenberg, H.; Jaeger, W.; Kotz, J.; Philipp, B.; Seidel, C.; Stscherbina, D. Polyelectrolytes. Formation, Characterization and Application; Hanser Publishers: Munich, 1994. (5) Bolto, B.; Gregory, J. Water Res. 2007, 41(11), 2301–2324. (6) Leemans, L.; Jerome, R.; Teyssie, P. Macromolecules 1998, 31(17), 5565– 5571. (7) Goddard, E. D.; Chandar, P. Colloids Surf. 1989, 34(3), 295–300. (8) Boucard, N.; Viton, C.; Agay, D.; Mari, E.; Roger, T.; Chancerelle, Y.; Domard, A. Biomaterials 2007, 28, 3478–3488. (9) Baigl, D.; Ober, R.; Qu, D.; Fery, A.; Williams, C. E. Europhys. Lett. 2003, 62(4), 588–594. (10) Boucard, N.; David, L.; Rochas, C.; Montembault, A.; Viton, C.; Domard, A. Biomacromolecules 2007, 8, 1209–1217. (11) Rochas, C.; Rinaudo, M. Biopolymers 1980, 19(9), 1675–1687. (12) Villetti, M.; Borsali, R.; Diat, O.; Soldi, V.; Fukada, K. Macromolecules 2000, 33, 9418–9422. (13) Wang, Z.-Y.; White, J. W.; Konno, M.; Saito, S.; Nozawa, T. Biopolymers 1995, 35(2), 227–238.

6460

DOI: 10.1021/la900061n

are not soluble in the nonionized state. Their chain conformation in aqueous solution is rather complex, and a deep study of the balance between hydrophobic and hydrophilic interactions seems necessary. Indeed, the control of the latter is the key point in the processing of polysaccharide materials such as aerogels,14 physical hydrogels,15 hollow fibers,16 nanoparticles,17,18 or multimembrane bioreactors.19 The difficulty of having an overview of the effects of this balance arises from the multiscale organization of such systems20 and from the numerous parameters involved.15,21 In polyelectrolyte solutions both short (excluded volume, hydrogen bonding, hydrophobic interactions) and long-range (electrostatic) interactions occur. Compared to the structure of neutral polymer solutions, the screening of electrostatic interactions implies the existence of an intermediate ordering scale comparable to the distance between chains or the correlation length. Furthermore, detailed characteristics of the polysaccharide chain (residue and charge distributions, type of glycosidic bonds) are also important as they control the phenomenon of counterion condensation.2,22 The latter decreases the effective charge carried out by the macroion due to a screening effect and influences the electrostatic (14) Robitzer, M.; David, L.; Rochas, C.; Di Renzo, F.; Quignard, F. Langmuir 2008, 24(21), 12547–12552. (15) Montembault, A.; Viton, C.; Domard, A. Biomaterials 2005, 26, 933–943. (16) Rivas, R. N.; David, L.; Domard, A. Chitosan Hollow Fibers for Tissue Engineering. In Advances in Chitin Science; Domard, A., Guibal, E., Varum, K. M., Eds.; Montpellier: France, 2006; Vol. IX, pp 616-618. (17) Schatz, C.; Pichot, C.; Delair, T.; Viton, C.; Domard, A. Langmuir 2003, 19, 9896–9903. (18) Sorlier, P.; Rochas, C.; Morfin, I.; Viton, C.; Domard, A. Biomacromolecules 2003, 4(4), 1034–1040. (19) Ladet, S.; David, L.; Domard, A. Nature (London) 2008, 452(7183), 76–79. (20) Popa-Nita, S.; David, L.; Rochas, C.; Domard, A. Analysis of the structure of solutions of chitosan with controlled degrees of acetylation and polymerisation In Advances in Chitin Science; Domard, A., Guibal, E., Varum, K. M., Eds.; Montpellier: France, 2006; Vol. IX, pp 261-267. (21) Sorlier, P.; Denuziere, A.; Viton, C.; Domard, A. Biomacromolecules 2001, 2, 765–772. (22) Manning, G. S. J. Chem. Phys. 1969, 51(3), 934–938.

Published on Web 4/23/2009

Langmuir 2009, 25(11), 6460–6468

Popa-Nita et al.

Article

contribution of the chain stiffness together with the long-range conformation of polyelectrolyte chains by modifying electrostatic interactions. At the intermediate scale and in the dilute regime of a salt-free solution, the conformation of a polyelectrolyte corresponds to an extended succession of segments termed electrostatic blobs.23 Inside a blob, the chain conformation is quasi-unperturbed by electrostatic interactions but depends on the solvent quality as for neutral polymers. At larger length scales, electrostatic interactions dominate, and the different blobs repel each other. At higher polymer concentrations, i.e., in the semidilute regime, the major feature is the existence of a correlation length ξ, which also corresponds to the distance between chain portions. For distances shorter than ξ, the dilute solution organization is still valid, although for length scales higher than ξ, the chain is illustrated by a random walk of correlation blobs (governed by Gaussian statistics).24 Experimentally, in salt-free polyelectrolyte solutions (i.e., with no added salt), the pseudoperiodic order is illustrated by a correlation peak of the scattered intensity in light25 (SLS), small-angle X-ray9,12,26 (SAXS), or neutron27-29 (SANS) scattering measurements. The scattering vector at the maximum of the correlation peak is related to the interchain or correlation distance according to9 qmax ≈ 2π=ξ

ð1Þ

In good solvents, the scaling law of ξ with the polymer concentration cp is23 ξµcp -1=3 for cp < c ξµcp -1=2 for cp > c

ð2Þ

where c* is the critical concentration of chain entanglement. In fact, this transition was never experimentally demonstrated since c* is usually too low to allow conformational studies even with the most recent techniques. In the case of polyampholytes, thanks to a parallel with the classical Rayleigh problem of the instability of a charged droplet, Kantor and Kardar30 proposed that a polymer involving both short-range attractions and long-range repulsions may also form a necklace of globules for a critical linear charge density. Dobrynin et al.1,31 developed a different scaling theory describing the transitions between necklace-like conformations that a hydrophobic polyelectrolyte can undergo. Indeed, in the semidilute regime, the expected dependence of the correlation length with the polymer concentration, ξ µ cp-1/2, is only valid in the concentration range where ξ is larger than the length of the string between two neighboring pearls. In this regime, the hydrophobic character of the polyion does not influence the chain conformation, and hydrophilic strings control the organization. When the polymer concentration increases, the order of magnitude between ξ and (23) De Gennes, P. G.; Pincus, P.; Velasco, R. M.; Brochard, F. J. Phys. (Paris) 1976, 37, 1461–1473. (24) Dobrynin, A. V.; Colby, R. H.; Rubinstein, M. Macromolecules 1995, 28, 1859–1871. (25) Borsali, R.; Rinaudo, M.; Noirez, L. Macromolecules 1995, 28, 1085–1088. (26) Essafi, W.; Lafuma, F.; Williams, C. E. J. Phys. II 1995, 5, 1269–1275. (27) Nierlich, M.; Williams, C. E.; Boue, F.; Cotton, J. P.; Daoud, M.; Farnoux, B.; Jannink, G.; Picot, C.; Moan, M.; Wolff, C.; Rinaudo, M.; De Gennes, P. G. J. Phys. (Paris) 1979, 40, 701–704. (28) Essafi, W.; Lafuma, F.; Williams, C. E. Eur. Phys. J. B 1999, 9, 261–266. (29) Milas, M.; Lindner, P.; Rinaudo, M.; Borsali, R. Macromolecules 1996, 29 (1), 473–474. (30) Kantor, Y.; Kardar, M. Europhys. Lett. 1994, 27(9), 643–648. (31) Dobrynin, A. V.; Rubinstein, M. Macromolecules 2001, 34(6), 1964–1972.

Langmuir 2009, 25(11), 6460–6468

the string length changes, and the chain conformation is governed by the beads, with ξ varying as cp-1/3. Hence, the crossover concentration, cb, from the hydrophilic string-controlled to the hydrophobic bead-controlled regime is a significant parameter characterizing the balance between solvophilic/solvophobic interactions. In this work, synchrotron SAXS was used to study the scaling of the polyelectrolyte peak location and thus of the correlation distance in different polysaccharide solutions by varying the polymer concentration. This allowed us to estimate the value of the crossover concentration cb. First, we report on the behavior of chitosan solutions of different degrees of polymerization and molar fractions of N-acetyl-D-glucosamine residues (DA). The generalization of this approach to other kinds of copolymer chemical structures is then tested with sodium hyaluronate and sodium alginates of various relative contents of mannuronic and guluronic moieties. In addition, the evolution of the intensity and width of the polyelectrolyte peak is also shown to reflect the transition between the string- and bead-controlled regime.

II. Experimental Section Polysaccharide Samples. The initial chitosan produced by chemical heterogeneous N-deacetylation of chitin from squid pens was purchased from Mahtani Chitosan (Veraval, India). This sample of DA = 1.5% was subjected to a first step of purification, which consisted in dissolving chitosan at 0.5% (w/v) in an aqueous solution with stoichiometric amount of acetic acid. The solution obtained was filtered successively on membranes (Millipore) of porosity 3, 1.2, 0.8, and 0.45 μm; then, the polymer was precipitated using aqueous ammonia. After repeated washings with deionized water, the precipitate was finally freeze-dried. Chitosans with different degrees of polymerization were prepared from the initial sample by ultrasound treatments. To this aim, chitosan was redissolved in acetic acid aqueous solutions (or in hydrochloric acid aqueous solutions in order to obtain the chitosan hydrochloride form) in stoichiometric conditions. Ultrasounds with a maximal electrical power of 300 W and a frequency of 20 kHz were applied through a 3 mm microprobe (Sonics Vibra Cell, Fisher Scientific Bioblock). The sonication time determined the final chain length. This depolymerization method allowed the production of chitosans with well-defined DPs and narrow molecular weight distributions (Table 1). Chitosans of different DAs were also prepared from the reacetylation under soft conditions of the initial chitosan in a fresh solution of acetic anhydride in a water/propanediol mixture (50%/50% w/w), thus allowing the preservation of a statistical distribution of residues within the chains.21 The products were isolated by precipitation on adding aqueous ammonia followed by repeated washings with deionized water. Reacetylated chitosans at DAs between 40% and 70% were soluble in water whatever the pH. Therefore, acetone was used to precipitate and wash the samples. In this case, acetic acid and acetate salt were removed by dialysis. The obtained DAs were determined by 1 H NMR spectroscopy on a Bruker DRX 300 spectrometer using the method developed by Hirai et al.32 The weight-average degree of polymerization (DPw) and polydispersity index (Ip) of the samples were determined by size exclusion chromatography (SEC) coupled on line with a differential refractometer (Waters R 410, from Waters-Millipore) and a multiangle laser-light scattering detector operating at 632.8 nm (Wyatt Dawn DSP). A 0.15 M ammonium acetate and 0.2 M acetic acid buffer (pH = 4.5) was used as eluent. For each chitosan, the refractive index increment dn/dc was evaluated from previous studies.33 Results are listed in Table 2. (32) Hirai, A.; Odani, H.; Nakajima, A. Polym. Bull. 1991, 26(1), 87–94. (33) Schatz, C.; Viton, C.; Delair, T.; Pichot, C.; Domard, A. Biomacromolecules 2003, 4, 641–648.

DOI: 10.1021/la900061n

6461

Article

Popa-Nita et al.

Table 1. Weight-Average Degrees of Polymerization and Polydispersity Indexes of the Depolymerized Initial Chitosan (DA = 1.5%) by Ultrasound Treatment (Average Values from at Least Three Independent Measurements)

Table 2. Degrees of Acetylation (DA), Weight-Average Degrees of Polymerization (DPw), and Polydispersity Indexes (Ip) of the Different Reacetylated Chitosans (Average Values from at Least Three Independent Measurements)

DPw

Ip

DA (%)

DPw

Ip

650 ( 40 320 ( 20 25 ( 4

1.3 ( 0.1 1.2 ( 0.1 1.6 ( 0.2

1.5 ( 0.1 2.6 ( 0.1 9.0 ( 0.1 20.1 ( 0.2 28.0 ( 0.5 37.3 ( 0.4 54.9 ( 0.8 68.7 ( 0.9

3100 ( 60 2080 ( 80 3360 ( 50 2330 ( 70 3200 ( 50 2680 ( 70 2830 ( 70 2360 ( 80

1.6 ( 0.1 1.6 ( 0.2 1.6 ( 0.1 1.7 ( 0.2 1.7 ( 0.1 1.5 ( 0.1 1.6 ( 0.1 1.7 ( 0.2

Sodium hyaluronate of cosmetic grade obtained by bacterial fermentation with Streptococcus zooepidemicus was purchased from HTL SAS (Javene, France). The weight-average degree of polymerization and polydispersity index were determined by SEC using a phosphate buffer of 0.15 M NaCl and 0.01 M Na2HPO4 as eluent (European Pharmacopeia protocol). We found DPw = 13 200 ( 1200 and Ip = 1.5 ( 0.1, values in good agreement with those specified by H.T.L. Sodium alginates prepared by selective extraction from Durvilea Antarctica were kindly donated and characterized by Danisco (Copenhagen, Denmark). The obtained linear copolymers contain different ratios of mannuronic (M) and guluronic (G) moieties. NMR spectroscopy was used to study the composition and sequence of uronate residues;34 in order to obtain comparable values, the relative integrals of the signals were calculated using a specific procedure.35 The degrees of polymerization and polydispersity indexes of the samples were determined by SEC using the same eluent as for hyaluronate. The characteristics of the alginates studied in this paper are presented in Table 3. Preparation of Solutions. Chitosan was dispersed in water, and pure acetic acid or hydrochloride acid aqueous solution was added to achieve the stoichiometric protonation of -NH2 sites. Sodium hyaluronan and sodium alginates were directly dissolved in pure demineralized water. All the solutions were stirred overnight to allow a complete dissolution of the polymers. Small-Angle Synchrotron X-ray Scattering. SAXS measurements were performed at the ESRF (European Synchrotron Radiation Facility) on the BM2-D2AM beamline. A synchrotron source was necessary since the intensity scattered by polyelectrolyte solutions is not easily accessible by conventional laboratory sources. The incident photon energy was tuned to 16 keV, and a two-dimensional detector (CDD camera from Ropper Scientific) positioned at about 1.6 m from the sample was used. Cylindrical capillaries of roughly 2 mm internal diameter were filled with the polyelectrolyte solutions and sealed with Parafilm. The primary data corrections (dark current, flat field response, and tapper distortion) were carried out using the software available on the beamline. Silver behenate was used for the q-range calibration. 2D images were converted into radial averages over the image center to yield the scattered intensity I vs the scattering vector q. As an additional correction, the contribution of the cells filled with water was subtracted to the scattering intensity of the samples. Data were fitted using the following relation10,36 IðqÞ ¼ D þ

C þ qγ

Imax  2 1 þ 4 q -qwmax

ð3Þ

This equation takes into account the contribution of long-range electronic density fluctuations in the C/qγ term and fits the polyelectrolyte peak by a modified Lorentzian function of intensity Imax, a position of the maximum qmax, and a full width at half-maximum (fwhm) w. D is a constant baseline parameter.

(34) Grasdalen, H. Carbohydr. Res. 1983, 118, 255–260. (35) Salomonsen, T.; Jensen, H. M.; Stenbæk, D.; Engelsen, S. B. Carbohydr. Polym. 2008, 72(4), 730–739. (36) Wang, D.; Moses, D.; Bazan, G. C.; Heeger, A. J.; Lal, J. J. Macromol. Sci., Pure Appl. Chem. 2001, 38(12), 1175–1189.

6462

DOI: 10.1021/la900061n

For the calculation of qmax, a visual estimation of the maximum of the polyelectrolyte peaks directly on the scattering diagrams was in good agreement with the values deduced from eq 3. On the other hand, for the estimation of the maximum of the polyelectrolyte peak intensity, the contribution of the low angle upturn of the scattered intensity had to be subtracted, and the use of relation 3 was essential. The same numerical procedure also yielded the relative peak width (w/qmax).

Numerical Method for the Determination of the Crossover Concentration cb. For each studied polyelectrolyte, the experimental data representing the variation of the position of the maximum of the polyelectrolyte peak as a function of the polymer concentration, qmax vs cp, were analyzed using a numerical method developed in the Matlab/Octave programming language environment (script available upon request). Starting from the hypothesis of two scaling laws with exponents 1/2 and 1/3 at low and high cp values, respectively, the data were separated in two subsets, and the linear least-squares computation method was applied to determine the optimal limit between these two intervals associated with the considered scaling laws. The limit value of the concentration related to the transition between these two intervals was then identified as cb. The results were in good agreement with a manual construction of two scaling laws of exponents 1/2 and 1/3 adjusting the data visually.

III. Results and Discussion The parameters acting on the balance between hydrophilic and hydrophobic interactions in polyelectrolyte solutions are numerous. In the present paper, physicochemical factors, including the pH, added salt concentration, the nature of the solvent, and temperature, were kept constant. We then investigate the role of the solution concentration and polymer structural parameters such as the apparent linear charge density, the distribution of the repeating units along the chains, the type of glycosidic bonds, and the degree of polymerization. The chosen polyelectrolytes are the chitosan (random copolymer), the hyaluronan (alternated copolymer), and the alginate (block copolymer); their chemical structures are shown in Table 4. As shown in Table 4, chitosan is a linear copolymer constituted of randomly distributed 2-acetamido-2-deoxy-D-glucan (GlcNAc) and 2-amino-2-deoxy-D-glucan (GlcN) residues linked together via β-(1f4) glycosidic bonds. The present study mainly deals with chitosan acetate, a salt form chosen because the corresponding polyelectrolyte peak is usually well-defined in the experimental conditions described above. Nevertheless, previous studies of the polyelectrolyte microstructure in chitosan solutions10 pointed out the weak influence of the nature of the counterion on the balance between hydrophobic and hydrophilic interactions. In order to distinguish clearly the role of the polymer concentration from that of the counterion, some partial results concerning the hydrochloride form will also be presented. Langmuir 2009, 25(11), 6460–6468

Popa-Nita et al.

Article

Table 3. M:G Ratios, Relative Contents of Homopolymeric Blocks (MM and GG) and Alternating Sequences (MG or GM), Weight-Average Degrees of Polymerization, and Polydispersity Indexes of the Studied Alginates (Average Values from at Least Three Independent Measurements) M:G ratio

% MM blocks

% GG blocks

% MG or % GM

DPw

Ip

0.39 1.01 1.96 7.20

1 30 43 78

44 29 11 2

56 42 47 20

1320 ( 30 1330 ( 30 1420 ( 40 1600 ( 50

1.4 ( 0.1 1.6 ( 0.1 1.7 ( 0.1 1.8 ( 0.1

Table 4. Chemical Structures of Chitosan (a), Hyaluronan (b), and Alginate (c)

Figure 1. Scaling of qmax, the position of the maximum of the polyelectrolyte peak as a function of cp, the concentration of residues per unit volume of chitosan acetate solutions, for different degrees of polymerization, and a constant DA = 1.5%. The black line represents the cp1/2 law fitting the results obtained for high and intermediate DPs. The gray line displays the cp1/3 laws fitting low DP results. The dotted line represents the cubic pseudolattice model (eq 4).

Role of the Chain Length. The depolymerization method using high-intensity ultrasounds allowed us to tailor a first homogeneous series of chitosan chains with the same degree of acetylation (1.5%), of precise degree of polymerization and low Ip (Table 1). The narrow molecular weight distribution is particularly useful for the study of the lowest DPs where the presence of oligomers strongly influences the polyelectrolyte organization. For the different DPs considered, a study was carried out for polymer concentrations, expressed in terms of number of residues per unit volume, ranging from 0.02 to 0.25 mol/L. This concentration unit will be used everywhere in the present paper. For each DPw: 3100 (initial sample), 650, 320 and 25, the SAXS diagram revealed a polyelectrolyte peak, and the position of its maximum (qmax) was studied as a function of the polymer concentration (Figure 1). The degree of acetylation of the samples, which are fully ionized, is low (1.5%), and hence the apparent charge density is high. Therefore, the ionic condensation described by Manning2 and Oosawa37 is (37) Oosawa, F. Biopolymers 1968, 6(1), 135–144.

Langmuir 2009, 25(11), 6460–6468

effective (DA is lower than 28%21,38), and electrostatic forces promote solubility. For high and intermediate DPs, qmax is scaling as cp1/2 as predicted by theory.23 The hydrophilic character of the polymer is predominant, meaning that in terms of the pearl-necklace model1 the chain conformation is structurally governed by the strings (stringcontrolled regime). Nevertheless, for high polymer concentrations, a deviation from this scaling law is observed (gray rhombus in Figure 1); this phenomenon will be largely discussed in the next subsection. Obviously the situation is different for DPw = 25, where short molecules are in a nonentangled dilute regime. If the experimental correlation distance between chains is ξ, then qmax = 2π/ξ is found to scale as cp1/3 (gray line in Figure 1). These results can be compared with the rough estimation deduced from the pseudolattice model.23 The periodicity distance (or unit cell parameter) d of the cubic pseudolattice can be written as 1 d ¼ cp NA

!1=3 or qmax ¼ 2πðcp NA Þ1=3

ð4Þ

where cp represents the number of residues per unit volume (mol/m3) and NA Avogadro’s number. This scaling relation is also reported in Figure 1 (dotted line) and fairly agrees with experimental results (empty triangles). To conclude, as long as the contour length of the chain is high enough, totally ionized chitosans of low DA exhibit the same relation between qmax and cp, with qmax µ cp1/2 corresponding to hydrophilic polyelectrolyte conformation in solutions; a deviation from this behavior was observed only at high polymer concentrations at DPw = 320 for which the preparation of the concentrated solutions is easier. For low DPs with contour lengths of the order of the correlation distance (DP = 25), chitosan short chains act as individual scattering objects with qmax µ cp1/3. Variation of cb with the Degree of Acetylation (DA). A second homogeneous series of samples (Table 2) was prepared by reacetylating the starting material. Thus, chitosans with a statistical distribution of repeating units, with the same DPw and Ip but different DAs, were obtained. Solutions of polymer (38) Lamarque, G.; Lucas, J.-M.; Viton, C.; Domard, A. Biomacromolecules 2005, 6, 131–142.

DOI: 10.1021/la900061n

6463

Article

Popa-Nita et al.

Figure 2. Log-log plots of scattered intensity I as a function of the scattering vector q for chitosan acetate solutions of different DAs, same DPw and Ip (Table 2), and at a constant concentration of residues per unit volume of 0.035 mol/L. The scattering curves are shifted vertically for clarity. Full curves represent the data fit using eq 3.

concentrations ranging from 0.01 to 0.15 mol/L were analyzed for each DA. An example of scattering curves is presented in Figure 2. At high DA values, changes in the structural state induce a broader polyelectrolyte peak and a shift of qmax toward lower q values. The variation of this maximum is the expression of the increase of the correlation length ξ with DA. In this example, ξ varies from 10 up to about 16 nm for DA = 1.5 to 69%, respectively. The decrease of qmax with DA can also be observed from a vertical reading (i.e., for a constant polymer concentration) of the qmax(cp) plot10 (Figure 3). When DA increases, the density of hydrophobic N-acetyl-D-glucosamine moieties along the chain increases, and the interaction balance is progressively displaced in favor of more hydrophobic interactions.9,10 Furthermore, for a given DA, when the polymer concentration increases, the correlation distance decreases. In the framework of the pearlnecklace model,1,31 the chains in the solution exhibit a structural transition from a hydrophilic to a more hydrophobic regime. Consistently, the scaling law of qmax with the polymer concentration changes from a power law of exponent R = 1/2 (black lines in Figure 3) to R = 1/3 (gray lines in Figure 3). Similar results were obtained in the case of synthetic polyelectrolytes of different fractions of ionizable units9,26 or in solvents of various quality and polarity.39 A crossover concentration, cb, can be associated with this structural transition (labeled by dotted lines in Figure 3). The transition in the solution morphology implies a change in the nature of the scattering entities and thus a variation of the form and the structure factor leading to a different profile of the scattered intensity peak. Assuming a Lorentzian shape for all polyelectrolyte peaks (eq 3), we determined the evolution of the relative peak width, w/qmax (where w is the full width at half-maximum), with DA (Figure 4). Here again, the two main structural regimes can be identified. At low DA, the stringcontrolled regime is characterized by low values of the relative peak width. In the higher DA range, the structural transition between the two regimes is evidenced by a steep increase of the value of w/qmax. As expected from the results of Figure 3, an increase of cp induces a decrease in the value of DA at the structural transition point. As a result, the value of the relative peak width can also be considered as a probe of the polyelectrolyte conformation. (39) Waigh, T. A.; Ober, R.; Williams, C. E.; Galin, J.-C. Macromolecules 2001, 34, 1973–1980.

6464

DOI: 10.1021/la900061n

Figure 3. Variation of qmax (position of the maximum of the polyelectrolyte peak) with cp (polymer concentration) in chitosan acetate solutions of different degrees of acetylation. The concentration-induced structural transition at cb between the string-controlled (black lines with qmax ∼ cp1/2) and the bead-controlled regimes (gray lines with qmax ∼ cp1/3) is labeled by dotted lines. Solid lines are least-squares fits (see IIExperimental Section).

Figure 4. Variation of the relative polyelectrolyte peak width, w/qmax, as a function of DA for chitosan acetate solutions of concentrations 35  10-3 (9) and 85  10-3 mol/L (b). Highlight of the three variation domains pointed out by the general law of variation of chitosan physicochemical parameters in aqueous solutions with DA.

Even if the concentration interval investigated here is different from previous studies,33,40 the plots in Figure 4 are in good agreement with the general law of variation of chitosan physicochemical parameters (such as dn/dc, second virial coefficient, Mark-Houwink-Khun-Sakurada equation parameters, etc.) in aqueous solutions as a function of DA.33,38,40,41 The three distinct domains generally pointed out by such a physicochemical law can be easily identified on the graphs representing the variation of the relative polyelectrolyte peak width. For DA values below 28%, in the ionic condensation regime, chitosan behaves as a polyelectrolyte of high charge density. For DA above 50%, it becomes a more hydrophobic polymer with a low charge density, and between these two extreme behaviors, a transition range is observed. For each DA, a systematic study of the variation of qmax with the polymer concentration allowed the identification of the stringand bead-controlled organization regimes and the critical crossover concentration associated with the transition between them (Figure 3). The influence of DA and the polymer concentration on the structural regimes was qualitatively described in previous studies.10 Furthermore, in the case of the chitosan gelation, the (40) Sorlier, P.; Viton, C.; Domard, A. Biomacromolecules 2002, 3(6), 1336– 1342. (41) Montembault, A.; Viton, C.; Domard, A. Biomacromolecules 2005, 6(2), 653–662.

Langmuir 2009, 25(11), 6460–6468

Popa-Nita et al.

Figure 5. Structural law of behavior, i.e., variation of the crossover concentration cb as a function of the degree of acetylation for salt-free stoichiometric chitosan acetate solutions in the fully protonated state. For DAs over 65%, the crossover point (qmax, cb) escapes from the accessible window of SAXS measurements. The average values of cb of chitosan hydrochloride (2) for three different DAs are added for comparison.

study of the role of cp on different physicochemical parameters42 (the zero-shear viscosity, pH at the gel point, the time to reach the gel point) evidenced the existence of a transition in the behavior of solutions taking place at a concentration c**. This phenomenon was found to be essential in the development of chitosan materials (especially hydrogels and fibers), and later on, c** was associated with cb.10 In this work, the experimental determination of the values of cb as a function of DA was possible thanks to numerous collected data analyzed with a linear least-squares fit method (see IIExperimental Section); some examples of estimation of cb are presented in Figure 3. The obtained results (including those concerning chitosan hydrochloride) are gathered in Figure 5, which represents the “structural” law of behavior of salt-free stoichiometric chitosan acetate solutions and defines the stringand bead-controlled regions. In the theory of the ionic condensation in solutions of hydrophobic polyelectrolytes,31 cb is explicitly related to the apparent linear charge density ( f ) and to the solvent quality parameter (τ) through the scaling relation: 0 !1=2 1 -1 τ@ τ3 A ð5Þ cb ≈ 3 1 þ b uf 2 where b is the structural length of the repeating unit (close to 5.1 A˚ for both N-acetyl-D-glucosamine and D-glucosamine moieties), u = (1 - DA)lB/b is the dimensionless electrostatic interaction parameter, and lB is the Bjerrum length (7.2 A˚ in pure water at 25 °C). The solvent quality parameter τ is related to the θ temperature of the polymer/solvent system by τ = (θ - T)/θ, where T is the experimental temperature. As long as the ionic condensation occurs (i.e., u > 1 or DA < 28%), the change in DA does not influence the effective charge density. Indeed, the charge renormalization occurs inducing a more or less constant effective value of the parameter f comparable to its maximum feff ∼ b/lB. Thus, the variation of cb with DA in this domain should only be due to the change of the solvent quality parameter τ. Above the ionic condensation limit (u < 1), the definition relation of cb (eq 5) can be simplified and becomes1 !1=2 2 -3 uf cb ≈ b ð6Þ τ (42) Montembault, A.; Viton, C.; Domard, A. Biomaterials 2005, 26, 1633– 1643.

Langmuir 2009, 25(11), 6460–6468

Article

where f = 1 - DA at full ionization. Unfortunately, the above scaling relations (eqs 5 and 6) cannot be directly used to extract a consistent value of the solvent quality parameter τ from the value of cb. Experimentally, the increase of the degree of acetylation is responsible for a decrease of the crossover concentration (Figure 5) and should reflect an increase of the parameter τ (as in eq 6). The loss in solvent quality with DA can be related to previous studies38 showing the decrease of chitosan solubility in terms of the second virial coefficient with increasing DA. Such behavior is attributed to the decrease of hydrophilicity of the polymer chains in favor of hydrophobic interactions brought about by GlcNAc residues.33 Figure 5 provides an interpretation of the concentration dependence of the general physicochemical law of behavior of chitosan33,40 in terms of structural regimes as a function of the degree of acetylation since the different variation domains are clearly observed: the ionic condensation regime, the more hydrophobic polymer organization region at high DA, and the transition range in between. Furthermore, still in Figure 5, results concerning the values of cb determined in the case of chitosan hydrochloride for three different DAs have been added. These values are somewhat lower than those of chitosan acetate, suggesting that chitosan hydrochloride is a slightly more hydrophobic polyelectrolyte (as also qualitatively evidenced in previous works10). Molecular modeling analyses43 have shown that acetate ions can interact via both electrostatic and hydrogen bonding with glucosamine residues, thus forming a complex which enhances the solubility of the polymer. As hydrogen bonding cannot be involved in the case of chloride anions, a difference in the hydrophobicity of the polyelectrolyte/counterion system is then observed related to the formation of strong ion pairs.43 To conclude, experimental results of a detailed study of chitosan conformation in salt-free aqueous solutions as a function of cp and DA are explained by the pearl-necklace model for hydrophobic polyelectrolytes.1 Thus, with increasing concentration, the structure of the solution is gradually controlled by the solvophobic beads. This also should favor interchain associations, responsible for the formation of large size aggregates in relation with the evolution of macroscopic physicochemical behavior at c**. Case of a Perfectly Alternated Copolymer: Sodium Hyaluronan. Chemical and enzymatic methods44,45 for the synthesis of a perfectly alternated chitosan with a high enough chain length are not yet available. Thus, the analysis was checked with hyaluronan, a perfectly alternated copolysaccharide. This appeared as a good alternative since both copolymers contain N-acetyl-D-glucosamine moieties. However, in chitosan chains, this residue is linked via β-(1f4) glycosidic bonds while in hyaluronan, alternating β-(1f4) and β-(1f3) linkages are found between the N-acetyl-D-glucosamine and the D-glucuronic acid constituting the repeating disaccharide unit (Table 4). At pH > pK0 = 3, its intrinsic pK, hyaluronate becomes soluble in water and behaves as a weakly charged polyelectrolyte due to the presence of an ionized carboxylate group on each disaccharide unit (u = lB/2b ∼ 0.7). As in the case of chitosan, the structure of sodium hyaluronate solutions was analyzed for concentrations (43) Terreux, R.; Domard, M.; Domard, A. Dynamic Study of the Interaction Between Ions and a 30 Monomer Chitosan Chain. In Advances in Chitin Science; Domard, A., Guibal, E., Varum, K. M., Eds.; Montpellier: France, 2006; Vol. IX, pp 219-226. (44) Akiyama, K.; Kawazu, K.; Kobayashi, A. Carbohydr. Res. 1995, 279, 151– 160. (45) Aly, M. R. E.; Ibrahim, E.-S. I.; El Ashry, E. S. H.; Schmidt, R. R. Carbohydr. Res. 2001, 331(2), 129–142.

DOI: 10.1021/la900061n

6465

Article

Popa-Nita et al.

Figure 6. Log-log plot of the scattered intensity I as a function of the scattering vector q for sodium hyaluronate solutions of different concentrations cp in residue per unit volume.

ranging from 0.016 to 0.11 mol/L. The SAXS diagrams also revealed a polyelectrolyte peak which intensity and position of maximum increased with cp (Figure 6). The presence of a polyelectrolyte peak in salt-free hyaluronate solutions was previously shown by SAXS.12 Again, the good quality of the present experimental results provides the opportunity to estimate the value of cb. The construction of the qmax(cp) plot reported in Figure 7 displays the shift of the polyelectrolyte peak toward higher q values with increasing polymer concentration and evidence the crossover concentration cb thanks to a least-squares fit numerical procedure. At low cp values, hyaluronate behaves as a classical hydrophilic polyelectrolyte, and we find the variation qmax µ cp1/2 predicted by the theory.1 But when cp increases, the hydrophobic structural regime emerges, and we then notice qmax µ cp1/3. The crossover concentration associated with the transition between these two domains is found to be about 60  10-3 mol/L, a value that can be compared with the “chitosan equivalent” of DA = 50% (cb ∼ 30  10-3 mol/L at DA = 54.9 ( 0.9%). The similarity in the behavior of sodium hyaluronate and chitosan acetate of DA close to 50% was also revealed by the comparison of the maximum intensity of the polyelectrolyte peak, Imax (eq 3), as a function of cp. To our knowledge, the structural transition between the string- and bead-controlled regimes was not yet described from such an analysis. Data are given in Figure 8 with two different characteristic scaling behaviors. Assuming a wormlike chain polymer model and strong electrostatic repulsions between polyelectrolyte chains, Koyama46 predicted a variation of the maximum intensity approximately proportional to the root square of the polymer concentration. Furthermore, SANS results concerning the conformation of poly[5-methoxy-2-(4-sulfobutoxy)-1,4-phenylenevinylene] in aqueous solutions36 were found consistent with this model. Here also, for low polymer concentrations, the maximum intensity of the polyelectrolyte peak scales as cpR with R ∼ 0.5. As the concentration increases, experimental results are no longer concordant with Koyama’s theoretical predictions. Indeed, in this domain a linear fit of Imax(cp) seems more appropriate. Such result is expected in the case of the scattered intensity by a collection of hydrophobic pearls. Moreover, according to the scaling model used in this work, the intensity of the scattering function at the peak position is given by Imax µ cpg, where g is the number of monomers in the correlation blob. For cp < cb, g µ ξc-1/2 and for cp > cb, g is of the order of the number of monomers in a bead (mb). This (46) Koyama, R. Macromolecules 1984, 17(8), 1594–1598.

6466

DOI: 10.1021/la900061n

Figure 7. Concentration-induced structural transition at cb between the string-controlled regime (black line with qmax ∼ cp1/2) and the bead-controlled organization (gray line with qmax ∼ cp1/3) in salt-free sodium hyaluronate solutions. Solid lines are leastsquares fits (see IIExperimental Section). Dotted lines represent the precision for the determination of cb.

Figure 8. Scaling of Imax, the maximum of the scattered intensity with cp, the polymer concentration in the case of solutions of sodium hyaluronate and chitosan acetate of DA close to 50%. The black curves represent the power law Imax µ cpR for the low concentration data with R = 0.6 ( 0.1 (hyaluronate) and R = 0.5 ( 0.1 (chitosan). The gray lines correspond to the linear fit Imax µ cp of the high concentration data. Dotted lines evidence the crossover concentrations cb.

results in the scattering intensity at the peak position to be Imax µ cp g µ cp 1=2 for cp < cb Imax µ cp g µ cp mb µcp for cp > cb

ð7Þ

In the case of sodium hyaluronate, this structural transition of behavior is observed for cp > 60  10-3 mol/L, in good agreement with the value of cb extracted from the qmax vs cp analysis. The concentration dependence of Imax in the case of chitosan acetate of DA close to 50% is very similar to that of sodium hyaluronan, but with a transition between the string- and bead-controlled regimes arising at a lower concentration, as expected. Taking into account that the structural length of the different monosaccharide units are nearly the same for the two copolymers, we could conclude from eq 6 that hyaluronate is more hydrophilic than chitosan of DA close to 50%, reflecting different contributions to the solvophilic balance. The glucuronic acid residue could be considered as more hydrophilic than N-glucosamine. The different nature of glycosidic linkages in the two copolymers also influences the behavior of the N-acetyl-Dglucosamine residues. Moreover, the alternated chemical structure of hyaluronate induces a more homogeneous charge density distribution than with a statistical chitosan, which is more likely to contain several short sequences of uncharged GlcNAc residues along the chain. Such hydrophobic chain portions could act as Langmuir 2009, 25(11), 6460–6468

Popa-Nita et al.

Figure 9. SAXS Scattering diagram of solutions of sodium alginates of different mannuronic/guluronic acid ratios at a constant concentration of residues per unit volume of 0.11 mol/L. The scattering curves are shifted vertically for improved clarity.

precursors of hydrophobic beads and favor the string-to-bead transition at a lower concentration. Consequently, in order to fully understand the influence of the polyelectrolyte chemical structure on the interaction balance, the distribution of charged residues should be taken into account. In the case of chitosan, this analysis is limited since the synthesis methods for alternate and block chitosans with long chains are not available yet. Case of a Complex Block Copolymer: Sodium Alginate. Alginates are copolysaccharide polyelectrolytes with a complex primary structure consisting in β-(1f4)-linked D-mannuronic (M) and R-(1f4)-linked L-guluronic (G) moieties (Table 4). The residues are organized in three types of blocks: homopolymeric blocks of mannuronate (MM) and guluronate (GG) and blocks with an alternating sequence of both residues (MG or GM). Four different alginate samples were analyzed (see Table 3 for details). Sodium alginate is soluble at neutral pH due to the presence of carboxylate groups. The polymers were then dispersed in deionized water, and solutions of concentration ranging from 0.03 to 0.23 mol/L were analyzed by synchrotron SAXS. The polyelectrolyte peak appearing in the SAXS diagrams is displayed in Figure 9. Only slight changes are observed in the position and shape of the polyelectrolyte peaks obtained on the scattering diagrams for the samples at constant cp. The role of the polymer concentration on the scattering profile was studied for each alginate, thus obtaining qmax = f(cp) plots (Figure 10). Alginates at low concentrations in deionized water act as hydrophilic polyelectrolytes, and then, the variation qmax µ cp1/2 predicted by the theory is again verified. For more concentrated solutions, the alginate conformation becomes dominated by the hydrophobic beads, and then qmax µ cp1/3 is found. Using the same analytical and numerical procedures as before, the crossover concentrations cb were determined. Taking into account the precision of the determination of cb, we can conclude that the relative content of mannuronic and guluronic moieties does not influence the polymer hydrophilicity. This could be explained by the equivalent molecular structure of the two saccharide units constituting the alginate chain. Indeed, guluronic acid is a carbon 5 epimer of mannuronic acid. The content of mannuronic and guluronic moieties with different glycosidic bonds (β-(1f4) and R-(1f4), respectively) is considered a key parameter influencing alginate chain stiffness47,48 and properties in solution49 and in the (47) Smidsrød, O.; Glover, R. M.; Whittington, S. G. Carbohydr. Res. 1973, 27 (1), 107–118.  (48) Dentini, M.; Rinaldi, G.; Risica, D.; Barbetta, A.; Skjak-Bræk, G. Carbohydr. Polym. 2005, 59, 489–499. (49) Strand, K. A.; Boe, A.; Dalberg, P. S.; Sikkeland, T.; Smidsrød, O. Macromolecules 1982, 15, 570–579.

Langmuir 2009, 25(11), 6460–6468

Article

Figure 10. Variation of qmax (position of the maximum of the polyelectrolyte peak) with cp (polymer concentration) in sodium alginate solutions of different M:G ratios. The concentration-induced structural transition at cb between the stringcontrolled (black lines with qmax ∼ cp1/2) and the bead-controlled regimes (gray lines with qmax ∼ cp1/3) is labeled by dotted lines. Solid lines are data fit obtained by a least-squares fit procedure.

gel phase.50,51 Nevertheless, according to our results, the influence of this parameter on the polymer solvophobicity is negligible. An additional feature concerning the SAXS diagrams of the different polyelectrolytes studied in the present paper is related to the observation of an “upturn” in the range of low scattering vector values, i.e., q < 0.02 A˚-1. The existence, at a higher scale of organization, of “domain structures” containing chain aggregates gives rise to this low-angle excess scattering.52,53 This superposed signal is more apparent for scattered intensities of highly concentrated solutions and, in the case of chitosan, for the highest DA values, which again is the signature of the displacement of the interaction balance toward more hydrophobic interactions. Such conclusions were also reached when dynamic light scattering studies were performed on chitosan solutions of varying DA and concentration.18 While the formation of such regions may be a general feature of hydrophobic polyelectrolyte solutions, their precise shape and size distributions appear to be nonuniversal and not yet fully understood. In the case of highly concentrated alginate solutions, the low angle upturn of the scattered intensity is well-defined. The data fit using relation 3 allows the calculation of the term C/qγ with a relatively good precision. A γ exponent close to 2.7 ( 0.2 is found for the four alginate samples studied in the present paper. This value is comparable to results from the literature: for instance, SANS experiments on PSS solutions of different degrees of sulfonation and counterions54 allowed the determination of the corresponding power law exponent γ, which was found to lie in the range of 2.5-4.

IV. Conclusion Small-angle synchrotron X-ray scattering allowed us to investigate the conformation of natural anionic and cationic polyelectrolytes in salt-free aqueous solutions at the nanoscale. The detailed study of the polyelectrolyte peak provided an insight into the types of interactions governing the solution morphology. The polymer concentration and structural parameters such as the (50) Draget, K. I.; Braek, G. S.; Smidsrød, O. Carbohydr. Polym. 1994, 25, 31–38. (51) Stokke, B. T.; Draget, K. I.; Smidsrød, O.; Yuguchi, Y.; Urakawa, H.; Kajiwara, K. Macromolecules 2000, 33, 1853–1863. (52) Ermi, B. D.; Amis, E. J. Macromolecules 1998, 31, 7378–7384. (53) Borsali, R.; Nguyen, H.; Pecora, R. Macromolecules 1998, 31(5), 1548– 1555. (54) Zhang, Y.; Douglas, J. F.; Ermi, B. D.; Amis, E. J. J. Chem. Phys. 2001, 114 (7), 3299–3313.

DOI: 10.1021/la900061n

6467

Article

Popa-Nita et al.

degree of polymerization, the apparent linear charge density, the type of glycosidic linkage, and the nature and distribution of the repeating units along the chains were found to influence the hydrophilic/hydrophobic interaction balance. Our results could be described by the theory of Dobrynin and Rubinstein1,31 (for the variation of the position of the maximum of the polyelectrolyte peak, qmax, with the polymer concentration, cp) and partly by the theoretical predictions of Koyama46 (for the intensity maximum, Imax, as a function of cp). In terms of the pearl-necklace model, we observed the transition between hydrophilic string-controlled and hydrophobic bead-controlled organizations with increasing cp. Thus, for low cp, qmax was found to vary as cp1/2 and Imax µ cp1/2. For high cp values, we noticed that qmax µ cp1/3 and Imax µ cp was no longer concordant with the Koyama theoretical prediction. The crossover concentration, cb, associated with the transition between these two organization regimes was determined for hyaluronan, chitosans of different DAs, and alginates of various M:G ratios. We then used cb as a way to evaluate the solvophilic/solvophobic behavior of polyelectrolytes. The structural law of behavior was built for statistical chitosan. The variation of cb with DA

6468

DOI: 10.1021/la900061n

was found similar to the general law of behavior describing the variation of the physicochemical properties of chitosan chains with this parameter.33 Taking into account the charge densities, the constitutive units with their respective glycosidic linkages, and the intrachain distribution of residues, we concluded that sodium hyaluronate was more hydrophilic than chitosan acetate of DA close 50%. Concerning alginates, no influence of the distribution and relative content of the two saccharide units on the chain hydrophobicity was observed, as expected for a copolysaccharide with residues of equivalent hydrophilicity. Acknowledgment. We thank Hans-Christian Buchholt (Danisco, Denmark) for the gift and characterization of the alginate samples. We are in debt for the technical assistance of Jean-Michel Lucas and Agnes Crepet. We thank the CRG group at ESRF for allocated beamtime. These studies are part of the NanoBioSaccharides project from the 6th European Framework Program “Nanotechnologies and nanosciences, knowledge-based multifunctional materials and new production processes and devices”.

Langmuir 2009, 25(11), 6460–6468