Biomacromolecules 2004, 5, 1899-1907
1899
Comparative Study of the Second and Third Heterogeneous Deacetylations of r- and β-Chitins in a Multistep Process Guillaume Lamarque, Christophe Viton, and Alain Domard* Laboratoire des Mate´ riaux Polyme` res et des Biomate´ riaux, UMR CNRS 5627, Domaine scientifique de la Doua, Baˆ timent ISTIL, 15, Boulevard A. Latarjet, 69622 Villeurbanne Cedex, France Received April 14, 2004; Revised Manuscript Received May 26, 2004
Second and third heterogeneous deacetylations in a multistep process under argon atmosphere of R- and β-chitins in the presence of 50% (w/v) NaOH, for temperatures ranging from 80 to 110 °C, were comparatively studied in order to optimize the multistep process of deacetylation. Along with the successive reactions, we observed important changes of chemical behavior with the crystalline state related to R- and β-chitins, amorphous and partially deacetylated chitin, and chitosan. Thanks to the full reacetylation of all the deacetylated samples, we succeeded in estimating the oxidoreductive alkaline degradation occurring during deacetylation, whatever the degree of acetylation (DA) of the copolymer. It clearly appeared that the crystalline state of the samples was the key parameter on which depended the rate constants of both alkaline hydrolysis and deacetylation and, consequently, the activation energy Ea and the preexponential factor A. We may now propose optimal conditions allowing the production of well-defined chitosans with low DAs and higher molecular weights than those usually reported in the literature. Introduction Chitin and chitosan are linear copolymers of (1f4)-linked2-acetamido-2-deoxy-β-D-glucan (GlcNAc) and 2-amino-2deoxy-β-D-glucan (GlcN). Though chitin is fully insoluble in both aqueous and usual organic solvents, as long as the copolymer is soluble in dilute acidic media, it is termed chitosan. Chitin exists according to two main polymorphic forms named R and β. In the R-crystallographic structure (arranged according to an orthorhombic cell with a P212121 space group)1, if b represents the fiber axis, the chain segments are antiparallel inside a polymer sheet, along the c axis, although they are parallel between two successive sheets of chains packed along the a axis. (Since we choose b as the fiber axis for β-chitin, the same assignment has been made for R-chitin in the current work and the literature has been transposed, where necessary, to conform to this.) The packing structure is strongly stabilized by intrachain, intrasheet, and intersheet hydrogen bonds in the three unit cell directions. In contrast, β-chitin crystallizes in a monoclinic cell with a P21 space group.2,3 Because chain segments are all parallel along the a and c axes, there is no hydrogen bond between two successive chain segments along the c axis. β-Chitin consequently exhibits a better reactivity,4 swelling,5,6 and solubility7 than R-chitin. Despite a lower reactivity, R-chitin is preferentially used by industrials due to its higher relative abundance.8 In contrast, chitosan, essentially produced from the deacetylation of chitin, is much less widespread in biomass where it is almost exclusively produced in some cell walls of fungi.9 This polysaccharide finds numerous applications8,10 in agriculture,11 biomedicine,12-14 papermaking, water treat* Corresponding author. E-mail:
[email protected].
ment, and the food industry.15,16 Its properties strongly depend on its degree of acetylation (DA)17 and molecular weight,14,16 which influence not only its physicochemical behavior18-20 but also its biological activity.21 The crystalline structure of chitosan has been demonstrated as mainly dependent on its water content22,23 and sample preparation before X-ray analysis.24,25 In the unit cell, the chain segments are antiparallel inside a sheet of polymer chains, along the c axis, as for R-chitin. If chitosan under the free amine form always crystallizes in an orthorhombic cell with the space group P212121, Yui et al.22 and then Okuyama et al.23 reported that the unit cell dimensions vary from its anhydrous form (a ) 8.28, b (fiber axis) ) 10.43, and c ) 8.62 Å) to its usual form (water content of 10.7%) (a ) 8.95, b (fiber axis) ) 10.34, and c ) 16.97 Å). The presence of water also had an important impact on the structure stability of chitosan. Indeed, although there was no direct interaction between two successive sheets of polymer chains along the a axis in an anhydrous chitosan, several hydrogen bonds via two or three water molecules were observed in the hydrated chitosan, thus stabilizing the structure. Several ways were proposed to deacetylate chitin. Though some enzymatic processes showed good efficiency,26 their use was restricted to low-molecular-weight chitin,27 on a laboratory scale, in view of the high cost of extraction and low productivity of chitin deacetylases.28,29 Chemical homogeneous and heterogeneous deacetylations of chitin in alkaline media remain the most studied and used. Sannan et al.30 reported on the possibility of carrying out a homogeneous deacetylation after immersion of chitin in a 40% (w/w) NaOH solution for at least 3 h at 25 °C, under reduced pressure. It was demonstrated that the deacetylated chitin with a DA of about 50%31 exhibited an unexpected solubility
10.1021/bm049780k CCC: $27.50 © 2004 American Chemical Society Published on Web 07/10/2004
1900
Biomacromolecules, Vol. 5, No. 5, 2004
in distilled water, ascribable to a random distribution of the GlcNAc/GlcN units along the polymer chains. Nevertheless, the pretreatment of the starting chitin as well as the long reaction time (several days) necessary to observe an important deacetylation made this technique unsuitable for an industrial scale. To achieve an effective reaction, deacetylation under heterogeneous conditions is preferentially used. In this case, highly concentrated aqueous solutions of sodium or potassium hydroxide (at least 40% (w/w)32-34) were processed at high temperatures, for several hours. Unfortunately, with such severe conditions, chitin underwent an important oxidoreductive alkaline hydrolysis responsible for a strong depolymerization. Several options were proposed to minimize the degradation such as changes of the reaction atmosphere (use of an inert gas such as N2 or Ar, instead of air), reduction of the amount of alkaline solution and use of a diluting agent such as acetone,35 and changes in the processing conditions with the use of reactive extrusion,36 sodiumthiophenolate,37 etc. The role of successive treatments which may involve either washing with water33,37-39 or dissolution/reprecipitation35,38 of the chitosan between successive deacetylation treatments was also investigated. This route yielded polymers with higher molecular weights and was necessary to produce fully deacetylated chitosans.37 However, no deep fundamental study on the heterogeneous multistep deacetylation was reported in the literature. We recently reported that the crystalline network was a key parameter for understanding the different chemical behaviors observed for R- and β-chitins during a first heterogeneous deacetylation.40 Thanks to the systematic separation between water-soluble and water-insoluble materials (at pH 8.5) from each deacetylated chitin recovered after neutralization of the reaction medium, we could determine that heterogeneous deacetylation of chitins led to important composition and structure heterogeneities. The structure and microstructure of the water-soluble and water-insoluble fractions (at pH 8.5) of R- and β-chitins were performed by X-ray diffraction and 1H NMR spectroscopy. Concerning the insoluble fractions of R-chitin, it was demonstrated that amorphous regions were first deacetylated before the crystalline network, which was more or less preserved. This particular mechanism led to block copolymers of GlcNAc and GlcN units. After the collapse of the R-crystalline network under the action of the alkaline solution, deacetylation could occur randomly on the whole sample. On the other hand, thanks to a weaker and more penetrable crystalline network, the water-insoluble fraction of β-chitin became fully amorphous after only a few minutes of reaction. This resulted in a one-step mechanism of deacetylation leading to copolymers of GlcNAc and GlcN units randomly distributed. On another hand, studies on water-soluble fractions (pH 8.5) revealed the importance of alkaline hydrolysis on the solution properties of these samples. This degradation emphasized the complexity of the reaction mechanisms and structural heterogeneities of the deacetylated samples. Nevertheless, we revealed that deacetylation rates of both watersoluble and water-insoluble fractions of β-chitin were very close to each other, proving that activation energies (Ea) and preexponential factors (A) were trustworthy thermodynamic
Lamarque et al.
parameters if we considered the only water-insoluble β-fraction (pH 8.5). Unfortunately, because of a denser crystalline network, deacetylation of the insoluble fraction of R-chitin was less effective compared to its soluble fraction. Thus, the parameters Ea and A for the insoluble fractions could be nonrepresentative of the whole sample as the deacetylation proceeded, if the proportion of soluble material became relatively high. We previously highlighted the chemical behavior of the copolymers under the chitin form during a first deacetylation, i.e., for DA ranging from 90 to 40%. The present paper focuses on the comparative study of the two following deacetylation steps under the same heterogeneous conditions (50% (w/v) NaOH, T > 80 °C, under argon atmosphere). We first compared the rate constants of heterogeneous deacetylations of R- and β-chitins obtained during the first reaction to those of the second and third. Activation energies and preexponential factors deduced from these kinetics were hereafter connected to the crystalline nature of the samples and the accessibility of the acetylated sites confined in the crystalline domains of the copolymer. The alkaline hydrolysis was also studied all along the deacetylation process. We propose a quite new method to estimate both its kinetics and its influence on the deacetylation. The combined studies of both deacetylation and degradation kinetics allow us to propose optimal conditions to produce chitosans of higher molecular weights and lower DAs than those reported in the literature, whatever the crystalline form of the starting chitin. Materials and Methods Raw Materials and Preparation. As in a previous work,41 R-chitin was extracted from marine Parapenaeopsis stylifera shrimp shells and β-chitin from squid Loligo pens, both provided by France Chitine. The starting materials for the study of the second heterogeneous deacetylation were prepared as follows: 222 and 667 mL of a 50% (w/v) NaOH solution (prepared under argon bubbling) were added to 10 and 30 g of R- and β-chitin powder, respectively. The chitin/ alkaline solutions were stirred for 30 min at 100 °C and 20 min at 90 °C, respectively. These reaction conditions were chosen for their effectiveness on the deacetylation and the lowest alkaline degradation they induced. Deacetylations were stopped by immersion of the reaction medium in liquid nitrogen. The alkaline solutions of deacetylated chitin were neutralized by 0.1 M HCl below 0 °C up to pH 8.5 in order to definitively stop the reaction and increase the proportion of the insoluble materials. The reaction medium was then centrifuged and the supernatant separated. The insoluble fraction was extensively washed with distilled water at pH 8.5 until the conductivity reached that of water and then was lyophilized and characterized. Amounts of 9.2 and 21.4 g of these one step deacetylated R- and β-chitins were obtained and termed 1D/R and 1D/β, respectively. Studying the third deacetylation on both starting materials was not necessary, since the deacetylation of both 1D/R and 1D/β samples would lead to a similar chitosan crystalline form. Due to its higher molecular weight and degree of deacetylation (DDA) compared to 1D/R, 10 g of 1D/β was
Heterogeneous Deacetylations of R- and β-Chitins
preferentially chosen to react with 222 mL of a 50% (w/v) NaOH solution (prepared under argon bubbling) for 20 min at 100 °C. The deacetylated chitosan was isolated in the same conditions as previously described and termed 2D/β. Second and Third Heterogeneous Deacetylations. Typically, 5 mL of a 50% (w/v) NaOH solution (prepared under argon bubbling) heated at a desired temperature (80-110 °C) was added to 0.225 g of 1D/R, 1D/β, or 2D/β. Kinetics were carried out for different reaction times ranging between 5 and 120 min and at different temperatures. Deacetylations were stopped by immersion of the reaction medium in liquid nitrogen. The deacetylated alkaline chitin solutions were neutralized by 0.1 M HCl below 0 °C up to pH 8.5. Reaction media were then centrifuged and supernatants were separated from the insoluble fractions. Some of them were dialyzed through membrane tubing (Spectra/Por Biotech CE MWCO: 500 g/mol) against water at pH 8.5 (fixed with ammonia), whereas the insoluble parts were extensively washed with deionized water at pH 8.5 until the conductivity reached that of water. Soluble and insoluble fractions were finally lyophilized and characterized. Characterization of the Degree of N-Acetylation and Distribution of N-Acetyl Groups by NMR Spectroscopy. Depending on their apparent solubility in deuterated solvent, the deacetylated samples were dissolved either in DCl/D2O (20% w/w) or in dilute acidic D2O (at pD 3-4) and their DAs analyzed by 1H NMR spectroscopy. Spectra were recorded on a Bruker AC 200 spectrometer (200 MHz for 1 H) at 298 K or a higher temperature, when necessary. At least 250 scans were acquired. 1H chemical shifts were expressed from the signal of 3-(trimethylsilyl)propanesulfonate-d4 (TSP-d4) as an external reference. DA was calculated as proposed by Hirai et al.,42 from the ratio of the methyl proton signal of the (1f4)-2-acetamido-2-deoxyβ-D-glucan residues to the whole H-2 to H-6′ proton signals. Solid-state 13C CP/MAS NMR spectroscopy was also used in the determination of the DAs whenever the apparent solubility seemed too low in the previous deuterated solvents. Spectra were obtained on lyophilized samples with the CP/ MAS technique (cross-polarization, magic angle spinning) using a Bruker DSX300 instrument working at 75.5 MHz. The samples (100 mg) were spun at 10 kHz at room temperature. The number of scans per spectrum was typically 300-600, the pulse intervals were 5 s, and the contact times were 2 ms. Chemical shifts were referred to tetramethylsilane and calibrated with the carbonyl carbon of glycine at 176.03 ppm as external standard. The DA was calculated by comparison between the integrated areas of the methyl group carbon (δ 22 ppm) and the C2-C6 signals (δ 50-105 ppm).42 To study the distribution of the N-acetyl groups, chitosans were hydrolyzed and analyzed as described by Vårum et al.43 Spectra were recorded on a Bruker Spectrospin AM 300 spectrometer (300 MHz). About 200-250 scans were acquired. X-ray Diffraction Analysis. Experiments were carried out at the ESRF (synchrotron of Grenoble) on the D2AM beam line in transmission mode at 16 keV (λ ) 0.7749 Å), using a bi-dimensional detector (from Ropper Scientific). The data
Biomacromolecules, Vol. 5, No. 5, 2004 1901 Table 1. DAs, Intrinsic Viscosities and Related Molecular Weights, Water Content, and Ash Contents of the Starting Materials
chitin
DAa (%)
Mw (g/mol)
water contentb (% w/w)
ash contentb (% w/w)
R β 1D/R 1D/β 2D/β
89.28 ( 0.07 90.30 ( 0.06 52.00 ( 0.08 42.80 ( 0.03 16.96 ( 0.15
1 020 000c 1 340 000c 720 000d 773 000d 428 000e
6.34 10.46 7.72 12.28 8.96
