Synthesis and Structural Characterization of Chitosan Nanogels

Jul 2, 2009 - Fabrice Brunel,†,‡ Laurent Véron,† Catherine Ladavi`ere,‡ Laurent David,‡ Alain Domard,‡ and. Thierry Delair*,‡. †bioMÂ...
0 downloads 0 Views 3MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Synthesis and Structural Characterization of Chitosan Nanogels Fabrice Brunel,†,‡ Laurent Veron,† Catherine Ladaviere,‡ Laurent David,‡ Alain Domard,‡ and Thierry Delair*,‡ †

bioM erieux, Chemin de l’orme, Marcy l’ etoile, France, and ‡Universit e de Lyon, Universit e Lyon 1, UMR CNRS 5223 ‘IMP’, Laboratoire des Mat eriaux Polym eres et Biomat eriaux, 15 bd. Andr e Latarjet B^ at. ISTIL, F-69622 Villeurbanne Cedex, France Received January 22, 2009. Revised Manuscript Received June 4, 2009 Colloidal physical gels of pure chitosan were obtained via an ammonia-induced gelation in a reverse phase emulsion. The water weight fraction and the chitosan concentration in the water phase were optimized so as to yield nanogels with controlled particle size and size distribution. The spherical morphology of the nanogels was established by transmission electron microscopy with negative staining. Wide-angle X-ray scattering experiments showed that these gels were partially crystalline. The electrophoretic mobilities of the particles remained positive up to pH 7, above which the particles aggregated due to the charge neutralization. From the investigation on the colloidal stability of these nanogels in various conditions (pH, salt concentration, temperature), an electrosteric stabilization process of the particles was pointed out, related to the conformation of mobile chitosan chains at the gel-liquid interface. Therefore, the structure of the nanogels was deduced as being core-shell type, a gelified core of neutralized chitosan chains surrounded by partially protonated chains.

1. Introduction The elaboration of materials from natural polymers is under intense investigation, as many applications can be addressed such as drug delivery,1 tissue engineering,2 diagnostics and imaging,3,4 or for more fundamental aspects of biology.5 Polysaccharides are prone to form gels via various approaches, and two major gelation strategies can be used. First, the polysaccharides can be chemically altered to yield amphiphilic polymers as shown for instance by Akiyoshi et al. for pullulans6 and Yoo et al. for chitosan.7 However, the chemical derivatization of these natural polymers definitely modifies their safety profiles, which is a crucial issue for in vivo applications. Hence, the second approach toward gels relies on noncovalent interactions such as the polyelectrolyte complex method,8 the tripolyphosphate-induced ionotropic gelation of chitosan,9 the calcium gelation of alginates,10 or by adding additives such as phosphoric and oxalic acids, or β-glycerophosphate11,12 for a temperature-induced gelation. However, there again, the presence of the gelation agent should be taken into account in the safety issue of the final material. Hence, an attractive method toward materials devoid of gelation agent is *Corresponding author. E-mail: [email protected]. (1) Agnihotri, S. A.; Mallikarjuna, N. N.; Aminabhavi, T. M. J. Controlled Release 2004, 100(1), 5–28. (2) Suh, J.-K. F.; Matthew, H. W. T. Biomaterials 2000, 21(24), 2589–2598. (3) Ravi Kumar, M. N. V. React. Funct. Polym. 2000, 46(1), 1–27. (4) Hayashi, T. Prog. Polym. Sci. 1994, 19(4), 663–702. (5) Nomura, Y.; Sasaki, Y.; Takagi, M.; Narita, T.; Aoyama, Y.; Akiyoshi, K. Biomacromolecules 2005, 6, 447–452. (6) Akiyoshi, K.; Deguchi, S.; Moriguchi, N.; Yamaguchi, S.; Sunamoto, J. Macromolecules 1993, 26(12), 3062–3068. (7) Yoo, H. S.; Lee, J. E.; Chung, H.; Kwon, I. C.; Jeong, S. Y. J. Controlled Release 2005, 103(1), 235–243. (8) Schatz, C.; Lucas, J.-M.; Viton, C.; Domard, A.; Pichot, C.; Delair, T. Langmuir 2004, 20(18), 7766–7778. (9) Calvo, P.; Remunan-Lopez, C.; Vila-Jato, J. L.; Alonso, M. J. J. Appl. Polym. Sci. 1997, 63, 125. (10) Grant, G. T.; Morris, E. R.; Rees, D. A.; Smith, P. J. C.; Thom, D. FEBS Lett. 1973, 32(1), 195–198. (11) Hamidine, M.; Heuzey, M.-C.; Begin, A. Rheol. Acta 2006, 45, 659–675. (12) Cho, J.; Heuzey, M.-C.; Begin, A.; Carreau, P. Biomacromolecules 2005, 6, 3267–3275.

Langmuir 2009, 25(16), 8935–8943

the physical gelation of solutions of polysaccharides, leading to physical hydrogels composed exclusively of water and polymers. For the elaboration of material for life sciences, chitosan is of particular interest because it is widely available from renewable sources, as it is obtained by deacetylation of chitin, one of the most abundant natural polymers. Moreover, its bioactivity, biodegradability and biocompatibility make this polymer quite appropriate for in vivo applications in medicine, and so it has been used for drug delivery,1 tissue engineering,2,13 or bioactive fibers.14 Chitosan, a β (1f4)-linked linear copolymer of 2-amino-2-deoxy-β-D-glucan (GlcN) and 2-acetamido-2-deoxyβ-D-glucan (GlcNAc), is a unique natural polyelectrolyte whose conformation and resulting properties depend on various structural and physicochemical parameters.15-22 Physical hydrogels of pure chitosan were obtained from chitosan solutions at concentrations superior to C*, the critical concentration of chain entanglement, by decreasing repulsive electrostatic interactions between macromolecules, that is, by lowering the charge density of chitosan. Two methods can be used: (i) acetylation in solution up to 70%23,24 or (ii) neutralization of the polymer by increasing (13) Montembault, A.; Tahiri, K.; Korwin-zmijowska, C.; Chevalier, X.; Corvol, M.-T.; Domard, A. Biochimie 2006, 88(5), 551–564. (14) Rivas Araiza, R. N.; Rochas, C.; David, L.; Domard, A. Macromol. Symp. 2008, 266(1), 1–5. (15) Anthonsen, M. W.; Varum, K. M.; Smidsrod, O. Carbohydr. Polym. 1993, 22, 193. (16) Berth, G.; Dautzenberg, H. Carbohydr. Polym. 2002, 47(1), 39. (17) Schatz, C.; Viton, C.; Delair, T.; Pichot, C.; Domard, A. Biomacromolecules 2003, 4(3), 641. (18) Schatz, C.; Viton, C.; Delair, T.; Pichot, C.; Domard, A. Langmuir 2003, 19, 9896. (19) Sorlier, P.; Denuziere, A.; Viton, C.; Domard, A. Biomacromolecules 2001, 2(3), 765. (20) Sorlier, P.; Rochas, C.; Morfin, I.; Viton, C.; Domard, A. Biomacromolecules 2003, 4(4), 1034–1040. (21) Varum, K. M.; Ottoy, M. H.; Smidsrod, O. Carbohydr. Polym. 1994, 25(2), 65. (22) Lamarque, G.; Lucas, J.-M.; Viton, C.; Domard, A. Biomacromolecules 2005, 6, 131. (23) Vachoud, L.; Zydowicz, N.; Domard, A. Carbohydr. Res. 1997, 302, 169– 177. (24) Moore, G. K.; Roberts, G. A. F. Int. J. Biol. Macromol. 1980, 2, 78.

Published on Web 07/02/2009

DOI: 10.1021/la9002753

8935

Article

Brunel et al.

the pH of the solution with ammonia gas25 or sodium hydroxide,17 for instance. These two approaches were described for macrogels but were not appropriate for the formation of colloids, despite the fact that two authors reported nanogels that could not be purified from the remaining polymer in solution and so were not adapted for the elaboration of colloids of hydrogels of chitosan.18,26 So, this lack of methods for producing colloids of pure chitosan hydrogels prompted us to investigate a gelation process in dispersed medium. The chitosan solution was emulsified in a mixture of triglycerides, a biocompatible organic phase, and the gelation was induced with ammonia. This method allowed the formation of colloids of pure chitosan ever reported.27 These new colloids open up numerous potential applications for bioactive and fully biocompatible drug-carrier systems. The present work is devoted to the physicochemical characterization of the dispersion in terms of morphology, structural organization, and colloidal stability. To this end, processing parameters such as the water fraction in the emulsion and the polysaccharide concentration in the water droplets were first optimized to yield dispersions controlled in size and size distribution. The morphology of these nanogels was examined by transmission electron microscopy (TEM) and their internal structure by small- and wide-angle X-ray scattering (SAXS and WAXS). Finally, the impact on colloidal stability of external parameters such as pH, salt concentration, and temperature was also investigated and discussed in relation with the structure of the nanoparticles.

2. Materials and Methods 2.1. Materials. The oil phase was composed of mediumchain triglycerides from capric/caprilic acid (Miglyol 812N) purchased from SASOL (Germany) and a surfactant: sorbitan monooleate (Span 80) from Fluka (Germany). Chitosan was provided by Mahtani chitosan PVT. Ltd. (India), batch 124, and its structural parameters such as the acetylation degree and average molar mass were determined in our laboratory: DA ∼ 5%, Mw ∼ 400 000 g 3 mol-1. 2.2. Chitosan Preparation. Prior to use, the polymers were purified as follows: dissolving in a 0.1 M acetic acid solution, filtration through Millipore membranes of decreasing porosity (from 3 to 0.22 μm), precipitation with an ammonia/methanol mixture, washing with deionized water until neutrality, and freeze-drying. A nitrous depolymerization was carried out to produce low molar mass polymers.28,29 Thus, chitosan was dissolved at 0.5% (w/v) in a 0.2 M acetic acid/0.1 M sodium acetate buffer. A 0.15 M sodium nitrite solution was added to the solution to obtain a nitrite/glucosamine unit molar ratio of 0.5. The reaction was performed under moderate magnetic stirring for various times (1-24 h). Low molar mass chitosans were recovered by precipitation with an ammonia/methanol mixture, then purified by several washings with deionized water until neutrality, and finally lyophilized. The molecular features of the chitosan samples used in this work are presented in Table 1. 2.3. Chitosan Characterization. The degree of acetylation (DA) was determined on purified chitosans by 1H NMR (25) Montembault, A.; Viton, C.; Domard, A. Biomacromolecules 2005, 6, 653– 662. (26) Domard, A.; Viton, C.; Lamarque, G. Novel Method for Preparing Chitin Nanaoparticles. FR Patent WO/2007/007014, 18 May 2007. (27) Brunel, F.; Veron, L.; David, L.; Domard, A.; Delair, T. Langmuir 2008, 24 (20), 11370. (28) Graham Allan, G.; Peyron, M. Carbohydr. Res. 1995, 277, 273–282. (29) Graham Allan, G.; Peyron, M. Carbohydr. Res. 1995, 277, 257–272.

8936 DOI: 10.1021/la9002753

spectroscopy (Varian, 500 MHz), according to the method developed by Hirai et al.30 The weight-average molecular weight (Mw), the z-average root mean-square of the gyration radius (RG,z) and the polydispersity index (Ip) were measured by gel permeation chromatography (3000 and 6000 PW TSK gel columns, Inner Diameter=7.8 mm and Length = 300 mm) coupled on line with a differential refractometer (Waters 410) and a multiangle-laser-light-scattering spectrometer (MALLS, Wyatt, Dawn DSP, Santa Barbara CA) equipped with a 5mW He/Ne laser operating at λ = 632.8 nm. Analyses were performed using the K5 flow cell. Light intensity measurements were derived following the classical RayleighDebye equation allowing us to deduce Mw with an expected error of about 10%. A degassed 0.2 M acetic acid/0.15 M ammonium acetate buffer (pH = 4.5) was used as eluent. The flow rate was maintained at 0.5 mL/mn. Refractive index increments (dn/dc= 0.196 mL 3 g-1) were determined from a calibration curve previously established under the same conditions, that is, in the same solvent and with an interferometer (NFT Scan ref) operating at λ = 632.8 nm.18 The water content in chitosan samples was determined by thermogravimetric analysis (DuPont Instrument 2950, Twin Lakes WI). 2.4. Nanoparticle Synthesis. Nanoparticles were prepared in a reverse emulsion process with an aqueous chitosan phase emulsified in an oil phase of Miglyol 812N containing the surfactant, as presented in Scheme 1. For a typical recipe with 1% surfactant, chitosan (0.05 g) was dissolved under moderate stirring by adding a stoichiometric amount of acetic acid with respect to the free amine functions, for each degree of acetylation. This aqueous solution (5 mL of chitosan at 1% (w/v)) was emulsified in 25 mL (23.75 g) of Miglyol 812N containing Span 80 (0.25 g, 1% (w/v)) with a sonoprobe (Bandelin KE76, Berlin, Germany) operating at 20 kHz with a peak-to-peak amplitude of 260 μm, under magnetic stirring. The ultrasonic probe (Ø=6  L = 118 mm) was immersed into the liquid down to 2 cm. The reactor (volume 40 mL) was thermostatted at 20 °C to prevent flocculation and destabilization of the emulsion due to the heating induced by sonication. After 5 min of emulsification, a stream of ammonia gas increased the pH of the medium up to 9, resulting in a gelation of the chitosan droplets. The ultrasound treatment was maintained for an additional 10 min to prevent droplet coalescence and/or particle aggregation. Finally, the dispersion was centrifuged at 1000g for 15 min. After elimination of the supernatant, the crude particles were dispersed in 40 mL of ethanol. This washing cycle was repeated twice with ethanol and twice with water, and then particles were dispersed in an ammonium acetate buffer (50 mmol 3 L-1, pH=4.5) by slow stirring overnight.

2.5. Characterization of Nanoparticles. 2.5.1. QuasiElastic Light Scattering (QELS). The size and size distribution of the nanogels were determined at 173° with a Zetasizer NanoZS instrument (Malvern instrument Ltd., Worcestershire, U.K.) equipped with a 4 mW He/Ne laser beam operating at λ=633 nm. The advantages for using backscatter detection at 173° arise from the fact that the light beam does not have to go through the sample but passes through a shorter path length, eliminating or reducing the multiple scattering effect. As a consequence, (i) the sensitivity is improved enabling the size measurement of dilute solutions and dispersions of sub-nanometer particles, (ii) a wider range of sample concentrations can be employed, and (iii) the effects of dust are greatly reduced as contaminants (since they scatter most light in the forward direction). All measurements were performed at 25.0 ( 0.2 °C. The self-correlation function was expanded in a power series (Cumulants methods).31 The polydispersity value provided by the software is a dimensionless parameter PDI defined as the ratio μ2/(Γ)2, where μ2 is the second cumulant of the correlation function and Γ is the average decay rate. (30) Hirai, A.; Odani, H.; Nakajima, A. Polym. Bull. 1991, 26, 87–94. (31) Koppel, D. E. J. Chem. Phys. 1972, 57, 4814.

Langmuir 2009, 25(16), 8935–8943

Brunel et al.

Article

Table 1. Molecular Mass Characterization of Chitosans and the Corresponding Nanogel Particle Size Distribution Evaluation: Weight-Average Molecular Weight (Mw), Number-Average Molecular Weight (Mn), and Polydispersity Index (Ip) molar mass distribution of all chitosan samples Mw (g 3 mol-1)

Mn (g 3 mol-1)

Ip

405 400 ( 8800 91 700 ( 530 43 950 ( 80 11 350 ( 30

304 800 ( 5800 56 260 ( 230 24 970 ( 80 5630 ( 40

1.33 1.63 1.76 2

Scheme 1. Scheme of the Nanoparticles Synthesis Process

Each value was the average of three series of 10 measurements. For a monodisperse colloidal suspension, the polydispersity index should be below 0.05, but values up to 0.5 could be considered for comparisons.32 The pH of the suspension was varied by addition of small amounts of aqueous ammonia or acetic acid. Zeta potentials were derived, using Smoluchowski’s equation, from electrophoretic mobility measurements obtained from the same instrument. 2.5.2. Transmission Electron Microscopy (TEM). Particle micrographies were obtained with a Philips CM120 instrument. The particles could not be observed directly because of the low contrast of the chitosan hydrogel and the vacuum in the microscope sample chamber. Indeed, specimens examined by transmission electron microscopy have to be thin, dry, and contain contrast usually from a heavy metal stain. To this end, we used a coloration technique, involving heavy metal salts, to ensure a sufficient contrast. A droplet of particle suspension in ammonium acetate buffer (50 mmol 3 L-1 at pH = 4.5) with a solid content of 0.3% (w/v) was deposited during 1 min on a Formvar-carbon-coated specimen grid. Once the particles were adsorbed onto the film surface, the excess sample was blotted off and the grid covered with a small drop (5 μL) of staining solution (see below). The latter was left on the grid for a few minutes and then blotted off. The sample was then dried and examined by TEM. The staining solutions used to observe our samples correspond to a 1% (w/w) solution of sodium silicotungstate dissolved in water (pH=4.5), a 2% (w/w) solution of phosophotungstic acid dissolved in water (pH=6.8), and a 1% (w/w) solution of uranyl acetate (pH=4.5).

particle size DPn 1850 350 150 35

average diameter (nm)

PDI

464 ( 120 386 ( 51 339 ( 29 239 ( 10

0.25 0.21 0.162 0.15

1500 cm, respectively, for WAXS and SAXS measurements. We used a two-dimensional detector (CCD camera from Ropper Scientific). All data corrections were performed with the bm2img software available on the beamline. The data were corrected for dark current, flat field response, and taper distortion. Finally, the azimuthal averages around the image center (position of the center of the incident beam) were performed for the q-range calibration standard (silver behenate33) and the samples. The scattering by the empty cell (filled with water only) was subtracted from all samples. The crude chitosan particle pellets (centrifugation 1000g, 15 min, in water) were placed between two metallic cylindrical sample holders with a 2 mm hole closed by two kapton windows. The scattering diagrams obtained with washed chitosan particles were treated qualitatively with the PowderCell software to identify the crystal structure (from the knowledge of the unit cell parameters and atomic coordinates in the crystalline structures of chitosan).

3. Results and Discussion

Wide-angle X-ray scattering was performed on the BM2-D2AM beamline at the ESRF (Grenoble, France). A synchrotron source was required for our study, since the intensity scattered by the polyelectrolyte hydrogel is not accessible by conventional sources in point collimation conditions. The data were collected at an incident photon energy of 16 keV (λ = 0.77 A˚). The size of the beam was about 225 μm width and 110 μm height at the sample location; and the sample-detector distance was about 16 and

The reverse emulsion was obtained by sonication of an aqueous solution of chitosan in miglyol as described in the Materials and Methods section. In our previous work,14 we determined that the optimum temperature of the emulsification process (i.e., the lowest size and polydispersity) was about 20 °C and the optimum surfactant concentration was 1% (w/w) versus the oil phase, which corresponds to a surfactant concentration of 5% (w/w) versus the water phase for a water fraction in the emulsion of 16.7% (w/w) (i.e., an emulsion weight ratio w/o of 1/5). The chitosan physicochemical characteristics also impacted the particle elaboration process. The particle size and polydispersity were better controlled when the DA and the molar mass were low (i.e., Mw =10-20 000 g 3 mol-1 and DA=1-5%). In this paper, we investigate the impact of the weight fraction of the water phase and of its viscosity on the course of the particle forming process. The morphology and structural organization at low scale of the particles are observed by electron microscopy and by SAXS and WAXS, respectively. Finally, the colloidal stability of these nanogels is extensively investigated and correlated to conformational aspects of chitosan chains at the colloid interface. 3.1. Impact of Processing Parameters. To investigate the impact of the water fraction in the emulsion on the gelation process, we decided to maintain a constant surfactant concentration against the aqueous phase (5% (w/w)), to ensure that the same amount of surfactant surrounded the water droplets (i.e., the surface tension remains constant). The water phase fraction in the process was varied from 15 to 33% (w/w), and the polymer concentration in the water phase was varied from 1 to 3%, as shown in Figure 1. The size and the size distribution of particles, obtained by quasi-elastic light scattering, were drastically affected when the water fraction in the oil phase reached 35% as a result of the increased coalescence rate. Another possible explanation is that the absolute amount of the surfactant increases in the

(32) Coombes, A. G. A.; Scholes, P. D.; Davies, M. C.; Illum, L.; Davis, S. S. Biomaterials 1995, 15(9), 673–680.

(33) Huang, T. C.; Toraya, H.; Blanton, T. N.; Wu, Y. J. Appl. Crystallogr. 1993, 26(2), 180–184.

2.5.3. X-ray Diffraction (Synchrotron-SAXS and WAXS).

Langmuir 2009, 25(16), 8935–8943

DOI: 10.1021/la9002753

8937

Article

Brunel et al.

Figure 1. Influence of the water fraction (w/w) in the emulsion and chitosan concentration (w/w) on the nanoparticle size (left) and size distribution index (right). Chitosan: DA=4% and Mw=20 000 g 3 mol-1. Data are the average of two independent experiments.

Figure 2. Transmission electron micrographs of chitosan nanoparticles (Mw =400 000 g 3 mol-1) in the dry state, with different negative stains: (a) uranyl acetate (1% (w/w) in water, pH = 4.5), (b) phosophotungstic acid (2% (w/w) in water, pH = 6.8), and (c) sodium silicotungstate (1% (w/w) in water, pH =4.5).

emulsion with the water fraction and, as described previously, Span 80 concentrations higher than 1% (w/w) resulted in lower emulsification efficiency (higher size and size distribution) due to the stabilization of the cavitation bubbles. No colloidal dispersion could be obtained (whatever the water weight fraction) if the chitosan concentration was 4% or above, probably because the viscosity of the water phase was too high to allow an efficient emulsification to be achieved. However, the impact, on particle diameter, of an increase in the chitosan concentration up to 3% was negligible for water fractions of 15-25%, though for a water phase fraction of 25% a limited but regular increase in average diameter and size distribution was observed. Conversely, at a 35% water fraction, the chitosan concentration in the aqueous phase had to remain below 1%, otherwise only a flocculated material was obtained. In this case also, the increase in the dispersed phase viscosity with the chitosan concentration could be responsible for the reverse emulsion breaking. Indeed, miglyol droplets could be trapped in the large and viscous aqueous phase. Hence, a multiple emulsion (o/w/o) could be formed, leading to the observed macroscopic gel. Moreover, at high chitosan input, a small amount of chitosan could be solubilized or trapped in the miglyol phase (after neutralization) and could contribute to the bridging of nanoparticles. In the following, the optimized processing parameters were used to obtain the maximum amount of particles with optimum control over the size and polydispersity. The optimized water fraction and chitosan concentration were, respectively, 25 and 3% (w/w), which, to our knowledge, correspond to the highest particle fraction, by comparison with other reverse emulsion processes previously reported in the literature. Moreover, most 8938 DOI: 10.1021/la9002753

of these processes used a chemical cross-linker such as glutaraldehyde,34,35 epichlorhydrine, or ethylene diglycidyl ether,36,37 whereas our process leads to particles containing only water and chitosan under the form of physical nanohydrogels. The purity of the particles and the yields in chitosan recovery were checked for this optimized synthesis according to the method previously described.27 Particles were constituted of 90% of chitosan, and 80% of the input of chitosan was recovered in the particles. The minimum size and polydispersity index of these optimized particles were respectively in the range of 200 nm and 0.2. These sub-micrometric size chitosan hydrogels will further be called “nanogels”. 3.2. Particle Characterization. Figure 2 shows spherical particles as evidenced by TEM using a negative staining technique of nanogels prepared with a high molecular mass chitosan (Mw= 400 000 g 3 mol-1). These micrographs enable us to confirm the quasi-elastic light scattering measurements: the nanoparticles exhibit sub-micrometeric diameters with a relatively broad distribution of particle sizes. The sodium silicotungstate staining solution (Figure 2c) provides the best contrast and was used to observe particles made with different molar mass chitosan samples. TEM experiments were indeed reproduced with particles obtained with chitosans of different molar masses (data not (34) Yang, B.; Chang, S.-q.; Dai, Y.-d.; Chen, D. Radiat. Phys. Chem. 2007, 76, 968–973. (35) Jia, Z.; Wang, Y.; Lu, Y.; Luo, G. React. Funct. Polym. 2006, 66, 1552– 1558. (36) Ohya, Y.; Shiratani, M.; Kobayashi, H.; Ouchi, T. J. Macromol. Sci., Part A: Pure Appl. Chem. 1994, A31, 629. (37) Banerjee, T.; Mitra, S.; Kumar, A. S.; Sharma, R. K.; Maitra, A. Int. J. Pharm. 2002, 243, 93.

Langmuir 2009, 25(16), 8935–8943

Brunel et al.

Figure 3. SAXS relative intensities of nanoparticle pellets obtained with chitosan samples of various molar masses before (a) and after (b) washing cycles.

shown), and all the observed particles showed the same characteristics (with diameters ranging from 100 nm to 1 μm, similar polydispersity, spherical shapes). The stain precipitated onto the formavar grid after deposition of the particles. Thus, the particles appeared white on a dark background (negative staining). Some dark spots were also visible and were probably due to some heterogeneity at the surface of the grid. The high polydispersities observed in Figure 2 are quite representative of the colloids obtained with a high molar mass chitosan and clearly underline the impact of the molar mass of the polymer on the colloidal properties of the dispersion. Small- and wide-angle synchrotron X-ray scattering techniques were used to characterize the nanostructure of the particles, including the crystallinity of the polymer network within the nanogels. Figure 3 shows the relative scattering intensities as a function of the scattering vector for the particle pellets obtained after centrifugation (1000g for 15 min) before (a) and after (b) the washing cycles. A scattering peak located close to q=0.35 A˚-1 was observed for the particle pellet before the washing cycles. This peak corresponds to a Bragg length 2π/q of about 18 A˚, a repetition pattern originating from the organization of the Span 80, as various self-organization modes of sorbitan monooleate were reported in the literature (such as liquid crystal or vesicles)38,39 and as this correlation peak disappeared after washing the particles with ethanol (a solvent in which both Span 80 and miglyol (38) Kato, K.; Walde, P.; Koine, N.; Ichikawa, S.; Ishikawa, T.; Nagahama, R.; Ishihara, T.; Tsujii, T.; Shudou, M.; Omokawa, Y.; Kuroiwa, T. Langmuir 2008, 24(19), 10762–10770. (39) Kato, K.; Walde, P.; Koine, N.; Imai, Y.; Akiyama, K.; Sugahara, T. J. Dispersion Sci. Technol. 2006, 27(8), 1217–1222.

Langmuir 2009, 25(16), 8935–8943

Article

Figure 4. Relative intensity of wide-angle X-ray scattering diagrams of nanoparticle pellets obtained with different molar mass chitosan samples ((a) 10 000 g 3 mol-1, (b) 40 000 g 3 mol-1, (c) 90 000 g 3 mol-1, and (d) 400 000 g 3 mol-1) compared to the theoretical diffraction diagram of the hydrated chitosan crystal structures obtained with the PowderCell software (ftp://ftp.bam.de/ Powder_Cell/) and crystallographic data in ref 43. The data were shifted by 125, 250, and 375 for diffraction diagrams (b), (c), and (d), respectively, for clarity.

Figure 5. Chitosan nanoparticle size (9) and polydispersity index (b) as a function of the solid content of nanoparticles resuspended in an ammonium acetate buffer (T=20 °C). Chitosan: DA=4% and Mw=10 000 g 3 mol-1. Data are the average of three independent experiments (SD.

are soluble). This surfactant organization could be (i) at the surface of the chitosan nanoparticles under the form of monolayers or bilayers such as lipoparticles,40,41 (ii) as separated vesicles in the aqueous phase, or finally (iii) around droplets of residual miglyol (despite the washing cycles). The surface organization onto chitosan nanoparticles could result from preferential interactions between chitosan and Span 80, as demonstrated (40) Thevenot, J.; Troutier, A.-L.; David, L.; Delair, T.; Ladaviere, C. Biomacromolecules 2007, 8(11), 3651–3660. (41) Thevenot, J.; Troutier, A.-L.; Putaux, J.-L.; Delair, T.; Ladaviere, C. J. Phys. Chem. B 2008, 112, 13812–13822.

DOI: 10.1021/la9002753

8939

Article

Brunel et al.

Figure 6. Chitosan nanoparticle size (a, 9), polydispersity index (a, 2), and electrophoretic mobility (b, [) versus pH in 50 mM ammonium acetate buffer (T=20 °C and SC=1% (w/w)). Chitosan: DA=4% and Mw =10 000 g 3 mol-1. Data are the average of three independent experiments (SD.

Figure 7. Chitosan nanoparticle size (a) and polydispersity index (b) versus time with different solutions: 50 mM phosphate buffer (2), 150 mM NaCl solution (9), and 50 mM ammonium acetate buffer (b) (T=4 °C and SC=0.1% (w/w)). Chitosan: DA=4% and Mw = 10 000 g 3 mol-1. Data are the average of three independent experiments ( SD.

previously,27,42 with an expected orientation of the polar saccharidic (furanic) head of sorbitan monooleate close to the polysaccharide (chitosan) particle surface. The fact that the intensity of the correlation peak decreased when the molar mass of chitosan increased can be explained by (i) the concomitant increase in particle diameter (Table 1), inducing a decrease in the available area (at constant polysaccharide amount), and (ii) the increase of the polydispersity of the particles. Interestingly, this observation is an indirect argument in favor of the organization of sorbitan monooleate at the colloid interface rather than in the form of separate vesicles or as surrounding droplets of miglyol. Indeed, these two last systems would not be impacted by any change in the number of particles at similar mass of polysaccharide. Finally, it is worth noting that, for every chitosan sample, a sharp interface between the particles and the continuous phase (oil or water) is evidenced in the scattering diagrams of all particles, both before or after particle washing steps. Indeed, in the low q values of Figure 3, the scattered intensity is inversely proportional to the fourth power of the scattering vector (Porod’s Law). In the solid state, chitosan is a semicrystalline polymer and the degree of crystallinity is a function of DA and the molar mass. The crystallinity is maximum for both chitin (i.e., 100% acetylated) and fully deacetylated chitosan.43,44 As we used a low DA

chitosan, the presence of a crystalline phase in the particles was investigated. Figure 4 shows the diffraction patterns of particle pellets after subtraction of the contribution of water. The nanohydrogels displayed a well-defined crystalline structure, in good agreement with the theoretical diagram of the hydrated crystalline allomorph of chitosan, also known as “tendon” chitosan.43 The diffractogram displays three diffraction lines, close to 0.7, 1.4, and 1.6 A˚-1 indexed to the (200)h, (220)h, and (020)h reticular planes, respectively, of the hydrated crystalline structure.45 The same crystalline structure can be obtained in macroscopic gels by direct neutralization of a chitosan aqueous solution with ammonia gas,46 but to our knowledge the present work is the first report of nanosized semicrystalline chitosan hydrogels. The degree of crystalinity could not be determined because the hydrogels contain more than 90% water. As a result, the raw diffraction patterns are dominated by the amorphous halo of the solvent. Such a contribution could not be separated from the contribution of the amorphous chitosan, and only the diffraction peaks associated to the crystalline phase of chitosan could be clearly deconvoluted from raw data. From data in Table 1, there is no clear effect of the molar mass of chitosan on the crystallinity of the nanogel. This result rules out our previous hypothesis that the higher crystallinity of low molar

(42) Grant, J.; Cho, J.; Allen, C. Langmuir 2006, 22, 4327. (43) Ogawa, K.; Hirano, S.; Miyanishi, T.; Yui, T.; Watanabe, T.; New Polymoph of Chitosan, A Macromolecules 1984, 17, 973–975. (44) Belamie, E.; Domard, A.; Chanzy, H.; Giraud-Guille, M.-M. Langmuir 1999, 15, 1549.

(45) Clark, G. L.; Smith, A. F. J. Phys. Chem. 1936, 40(7), 863–879. (46) Clayer-Montembault, A. Elaboration d’hydrogels physiques de chitosane: Applications a l’ingenierie tissulaire pour la regeneration du cartilage. Univ. C. Bernard, Lyon, 2004.

8940 DOI: 10.1021/la9002753

Langmuir 2009, 25(16), 8935–8943

Brunel et al.

Article

Figure 8. Chitosan nanoparticle size (a) and polydispersity index (b) versus time with different solutions: 50 mM phosphate buffer (2), 150 mM NaCl solution (9), and 50 mM ammonium acetate buffer (b) (T=20 °C and SC=0.1% (w/w)). Chitosan: DA=4% and Mw = 10 000 g 3 mol-1. Data are the average of three independent experiments ( SD.

mass chitosans would be a driving force toward better-defined colloids. Therefore, the better control over size and size distribution obtained with low molar mass chitosan would rather be attributed to a lower viscosity of the dispersed phase during the synthesis processing. 3.3. Colloidal Stability. 3.3.1. Effect of Dilution. After the washing cycles, the particle pellets were dispersed in an ammonium acetate buffer at 3% (w/w) solid content. At this solid content, the particle suspension presented a milky aspect, but after dilution to a solid content below 1% (w/w) it became transparent. The evolution of QELS sizing measurements as a function of the dilution of the suspension is reported in Figure 5. The measured size and polydispersity decreased with increasing dilutions, to reach a pseudoplateau below 1% solid content. At the plateau, independent particles were observed, as confirmed by transmission electron microscopy. Particle aggregation or bridging at high solid content took place during centrifugation via chain entanglement and hydrogen bonding. The dilution of the sample in an ammonium acetate buffer allowed hydrogen bond breaking22 and dissociation of the aggregated colloid into independent particles in suspension. When the dilutions from 3 to 1% were carried out in water, or phosphate or sodium acetate buffers, no dissociation was observed. The suspensions remained milky because these buffers do not induce hydrogen bond rupture. To confirm the particle association during the centrifugation steps, a diluted particle suspension was reconcentrated from 1 to 3% (w/w) by ultrafiltration. The reconcentrated suspension remained transparent with particle sizes of 220 nm. The transparent aspect of the unaggregated chitosan particles suspension is quite unique in the field of colloids. This can be the result of the very low optical contrast between the hydrogel (composed of more than 95% of water)25,47 and the continuous phase. Most of the colloids reported in the literature used a chemical or ionic cross-linker and hydrophobic interactions also occurred, which increased the density and the refractive index of the particles. In this work, the particles are highly hydrated nanogels. Indeed, the turbidity of a dispersed medium depends on the ratio of the refractive index of the dispersed phase to the refractive index of the continuous phase.48 3.3.2. Impact of pH and Nature of the Buffer and Temperature on Colloidal Stability. The variations of the particle size, polydispersity, and electrophoretic mobilities with pH are (47) Boucard, N.; David, L.; Rochas, C.; Montembault, A.; Viton, C.; Domard, A. Biomacromolecules 2007, 8(4), 1209–1217. (48) Brochette, P. Emulification: elaboration et etude des emulsions. Techniques de l’Ing enieur 1999, J 2, 150.

Langmuir 2009, 25(16), 8935–8943

Figure 9. Chitosan nanoparticle size (9) and polydispersity index

(2) versus the ionic strength or NaCl concentration (T = 20 °C and SC = 0.1% (w/w)). Chitosan: DA=4% and Mw=10 000 g 3 mol-1. Data are the average of three independent experiments ( SD.

reported in Figure 6 for particles obtained with chitosan of Mw= 10 000 g 3 mol-1 and DA=4%. For a pH value below pH 3, an instantaneous dissolution was observed, thus limiting the lower pH value to 4. On increasing pH, the particle size remained constant until pH=6, above which the average diameter started to increase before precipitation at pH=7. This is consistent with the charge neutralization observed by the decrease of the electrophoretic mobility with pH (Figure 6b). This behavior suggests a strong involvement of an electrostatic stabilization process. The colloidal stability of the particles was also studied at two different temperatures, 4 and 20 °C (Figures 7 and 8, respectively), for three different media: 150 mmol 3 L-1 NaCl solution, ammonium acetate, and phosphate buffer (both salt concentrations were 50 mmol 3 L-1 and pH = 4.5). For both temperatures, the particles remained stable for at least 3 months in the ammonium acetate buffer, but a fast aggregation was observed in the phosphate buffer. At pH = 4.5, the particles could be in the presence of divalent phosphate anions (pKa (H3PO4/H2PO4-)= 2 and pKa (H2PO4-/HPO42-) = 7), whose complexation with chitosan could lead to the aggregation of nanoparticles. Indeed, Vorlop and Klein prepared chitosan beads by ionic gelation with multivalent phosphate anions.49 Nevertheless, due to a pKa of 7 for the second acidity, compared to pH 4.5, this event was necessarily negligible. On the contrary, a possible cross-linking (49) Vorlop, K.-D.; Klein, J. Biotechnol. Lett. 1981, 3(1), 9.

DOI: 10.1021/la9002753

8941

Article

Brunel et al.

Figure 10. Chitosan nanoparticle size (a) and polydispersity index (b) versus time at different temperatures: þ4 °C (9), 20 °C (b), and 37 °C

(2) in an ammonium acetate buffer (SC = 1% (w/w)). Chitosan: DA = 4% and Mw = 10 000 g 3 mol-1. Data are the average of three independent experiments ( SD. Scheme 2. Scheme of the Particle Stabilization Mechanism: (a) in Acidic pH and at Low Ionic Strength (Electrostatic and Steric Stabilization Occurred), (b) at Acidic pH and High Ionic Strength (Steric Stabilization Only), and (c) at Neutral pH (No Stabilization)

involving a unique phosphate anionic site and hydrogen bondings could occur between phosphate groups and chitosan. In a NaCl solution, the particles aggregated at more than 2 weeks at room temperature but remained stable for 3 months at 4 °C. Moreover, as shown in Figure 9, no instantaneous aggregation was observed up to 300 mM sodium chloride. The screening of the electrostatic forces by counterions should lead to a fast aggregation of the particles; therefore, electrostatic repulsions could not be the only mechanism involved in the colloidal stability. Since the observed stability was independent of the presence of counterions, a steric stabilization could be considered with also an important role. This role would be governed by the balance between hydrogen bonding and hydrophobic interactions of chitosan chains as illustrated by the effect of temperature discussed above. This suggests that particles consist of a hydrogel core and a partially protonated shell ensuring the colloidal stability as depicted in Scheme 2a and b. The neutralization of the polymer charge of the shell induced the collapse of the expanded polyelectrolyte chains, responsible for the steric repulsions, leading to a fast aggregation (Scheme 2c). In the presence of salt, the change in conformation of the chitosan chains of the shell is not drastic enough to induce a fast aggregation and particles remained stable for months at 4 °C thanks to the residual steric effect (Scheme 2b). The faster aggregation observed at 20 °C could arise from the displacement of the balance between hydrophobic and hydrophilic interactions between chitosan chains toward the formation of more hydrophobic junction points, as established by Vachoud et al. in their study on the physical gelation of chitosan.23 These new junction points could alter the conformation of the chitosan in the shell of the particles with an 8942 DOI: 10.1021/la9002753

adverse impact on the steric stabilization process and hence the aggregation of the dispersed system. A similar effect was observed in the ammonium acetate buffer at 37 °C (Figure 10), but in these conditions the aggregation of the particles was controlled and the average size stabilized after 1 day at around 470 nm (with a concomitant decrease in the polydispersity index). Interestingly, on cooling this dispersion from 37 to 20 °C, the average particle diameter decreased to 350 nm (PDI=0.12), indicating that this process was partially reversible (on the time scale of our experiment), as expected from a process involving both hydrogen bonding and hydrophobic interactions, both being dependent on temperature.50

4. Conclusion Colloidal polysaccharide nanogels were obtained by gelation of a reverse emulsion of chitosan in miglyol. The process only used products generally regarded as safe and was controlled, in term of size and size distribution, by both external (synthesis process) and internal (polymer characteristics) parameters. With temperature and surface tension (via the relative weight amount of Span 80) set at optimum values, the water weight fraction in the emulsion and the chitosan concentration in the dispersed phase were varied to obtain a maximal amount of particles per batch, with controlled colloidal properties. This was essential in order to provide enough material for potential applications. The investigation of the impact of chitosan intrinsic parameters, DA and molar mass, on particle size and size distribution showed that chitosan should (50) Clayer, A.; Viton, C.; Domard, A. Macromol. Symp. 2003, 200(1), 1–8.

Langmuir 2009, 25(16), 8935–8943

Brunel et al.

have both low DA and molar mass. This could be related to the viscosity of the internal phase; the lower the viscosity, the better the control over the colloidal characteristics of the dispersion. WAXS results showed that the polysaccharide was partially crystalline in the nanogel, whatever the molar mass of chitosan, excluding thus that crystallization of the short chitosan chain could be a driving force toward well-defined colloids. Chitosan nanogels exhibited a very long-term colloidal stability even in the presence NaCl up to 300 mM. This stabilization was attributed to expanded protonated chains of chitosans at the colloid interface. The neutralization of the charges by increasing pH led to the collapse of the interfacial chitosan and hence the aggregation of the colloid. Chitosan is also known to form hydrogen bonds and hydrophobic interactions. Particles could associate via hydrogen bonding when the dispersion was concentrated by centrifugation (only the hydrogen bond breaking ammonium acetate buffer could dissociate the particles upon dilution). The role of

Langmuir 2009, 25(16), 8935–8943

Article

hydrophobic interactions was evidenced by the observed aggregation of the nanogels on increasing temperature. These results are in favor of a core-shell structure of the colloidal gels with a condensed, though hydrated core surrounded by free chitosan chains. This investigation provides fundamental data on the structure and the physicochemical properties of these partially crystalline nanogels. Various applications of these new materials are currently under investigation, and they will be reported in due term. Acknowledgment. We thank J. M. Lucas for his assistance in the chitosan charaterization by gel permeation chromatography coupled to a MALLS detector, Beatrice Burdin and the members of the “Centre Technologique des Microstructures” for their assistance in transmission electron microscopy, and Cyrille Rochas and the CRG group for the Synchrotron SAXS and WAXS experiments. This work is part of the NanoBioSaccharide European project.

DOI: 10.1021/la9002753

8943