Article pubs.acs.org/Biomac

Preparation of Sterically Stabilized Chitin Nanowhisker Dispersions by Grafting of Poly(ethylene glycol) and Evaluation of Their Dispersion Stability Jun Araki* Faculty of Textile Science and Technology, and Institute for Fiber Engineering (IFES), Interdisciplinary Cluster for Cutting Edge Research (ICCER), Shinshu University, Tokida 3-15-1, Ueda, Nagano Prefecture 386-8567, Japan

Mari Kurihara Graduate School of Science and Technology, Shinshu University, Tokida 3-15-1, Ueda, Nagano Prefecture 386-8567, Japan S Supporting Information *

ABSTRACT: Sterically stabilized chitin nanowhiskers (ChNWs) were prepared by surface grafting monomethoxy poly(ethylene glycol) (mPEG) via reductive amination of primary amino groups on ChNWs and terminal aldehydes on mPEG. The amount of grafted mPEG was determined to be 0.2−0.3 g/g ChNWs, by conductometric titration, from the decrease in amino groups after grafting. ChNWs with controlled amounts of surface amino groups were obtained by deacetylation; however, this did not cause a drastic change in the amount of grafted mPEG. Grafting was confirmed by Fourier-transform infrared spectroscopy; however, X-ray diffractometry indicated no sign of mPEG. Thermogravimetry indicated a higher amount of mPEG than that from titration, suggesting an overestimation due to the facilitated combustion of grafted samples. In contrast to ungrafted samples, all grafted samples were stable in the presence of electrolytes. However, liquid− crystalline phase separation of grafted ChNWs was not observed, possibly owing to the high viscosity of the concentrated sample.

■

INTRODUCTION In recent years, analyses and applications of structural biopolymers of skeletal building blocks in living systems, especially crystalline polysaccharides, have received significant attention. Two major candidates for this are cellulose, which is a dominant constituent of cell walls of higher plants,1−3 and chitin, which is found in skeletal shells of crustaceans such as crabs and shrimp, insect shells, abyssal creatures, and diatoms.4 Cellulose and chitin, the two most abundant biopolymers on earth, are linear homopolysaccharides comprising D-glucose and 1−4 D-N-acetylglucosamine, respectively. In living structures, both of these polysaccharides associate in a parallel manner to form fine crystalline fibers called microfibrils. Their width ranges from 2 to 3 nm for wood cellulose microfibrils1,5 to ca. 20 nm for algal cellulose microfibrils6 and diatom chitin microfibrils.7 Depending on the cross-sectional size, between several tens8 and 12002,9 polysaccharide molecules regularly align in a single microfibril. With the aid of strong intra- and intermolecular hydrogen bonding between neighboring polysaccharide molecules, highly crystalline microfibrils, which are insoluble in a wide variety of solvents, are formed. Microfibrils can be separated into individual fine nanofibers through acid hydrolysis,5,10−14 oxidation,15 and attrition;14 further disintegration into shorter fragments in the case of acid hydrolysis yields © XXXX American Chemical Society

aqueous, colloidal dispersions of crystalline, rod-like particles known as nanowhiskers.1,3,4 One of the outstanding characteristics of nanowhiskers is their remarkably high mechanical performances, i.e., modulus and strength.1,3 Through a bending test using an atomic force microscope, the Young’s modulus of a single cellulose nanofiber obtained from tunicin is estimated to be 150 GPa,16 which agrees with that obtained by crystalline modulus measurements using X-ray diffraction.17 A value for the stress at break of single cellulose nanowhiskers was reported to be ∼6 GPa by sonication-induced fragmentation measurements. 18 The Young’s modulus of β-chitin microfibrils was evaluated to be 1−2 × 1011 N m−2 by an AFM-based bending test19 and 150 GPa by calculation from the mechanical modulus of nanocomposites containing chitin nanowhiskers.20 Although no actual measurement of the modulus for α-chitin nanowhiskers have been reported, the crystalline modulus of α-chitin was reported to be 41 GPa from the tensile measurement of crystalline α-chitin fibers.21 Other desirable characteristics of cellulose/chitin nanowhiskers, such as their relatively low Received: November 3, 2014 Revised: December 2, 2014

A

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

density, low thermal expansion coefficient, high heat resistivity, and biodegradability as well as their remarkable mechanical properties, have resulted in their recent extensive applications as reinforcing nanofillers in nanocomposite films,20,22 fibers,23 and electrospun nonwoven fabrics.24 Having a high degree of (or ideally independent) dispersion of nanowhiskers in nanocomposite matrices is highly significant because aggregations of nanowhiskers may act as defects in the nanocomposites as well as result in a semimicroscopic discontinuity of matrices. Nanowhiskers are rod-like colloidal particles whose dispersion stabilities are strongly dependent on two stabilization mechanisms: electrostatic stabilization caused by surface charge groups and steric stabilization contributed by surface-adsorbed or grafted polymeric chains, similar to that for other colloidal systems.1 The latter may be of practical interest because its stabilization ability is still operative when electrostatic stabilization is ineffective, i.e., in the presence of electrolytes or in organic systems with a lower dielectric constant. Since pioneering works on the adsorption of surfactants on cellulose nanowhisker surfaces25 and terminal grafting of monomethoxy poly(ethylene glycol) (mPEG) on cellulose nanowhiskers26 were reported, there is now a vast amount of literature on the steric stabilization of cellulose nanowhiskers using various types of polymers.1 Chitin nanowhiskers (ChNWs) have a crystal structure very similar to that of cellulose nanowhiskers; however, with the exception of graft polymerization of acrylic acid onto chitin nanofibers for improving stability in aqueous basic solutions,27 almost no investigations on the steric stabilization of ChNWs have been reported. Achieving steric stabilization of ChNWs will provide several advantages over sterically stabilized cellulose nanowhiskers; for example, preparation of a novel type of nanocomposites will be expected by contributions of ChNWs properties, including diverse biodegradability by a wide variety of enzymes or antibacterial activities.28 Although the above-mentioned study by Ifuku et al.27 achieved ChNWs stabilization at high pH, the grafted poly(acrylic acid), a polyelectrolyte, might shrink because of changes in pH or solvent exchange to lower the effect of stabilization. Therefore, grafting of nonionic polymers that are soluble in both organic solvents and water could be advantageous. In the present study, steric stabilization of ChNWs was examined by grafting mPEG to surface amino groups of ChNWs via reductive amination. PEG is a versatile synthetic polymer that is soluble in water and a wide variety of organic solvents, and it has been widely applied in the steric stabilization of various molecules, especially proteins.29 As stated above, mPEG grafting has been previously reported to effectively improve the dispersion stability of cellulose nanowhiskers.26 As a grafting strategy, we chose reductive amination between terminal aldehyde groups of oxidized mPEG and surface amino groups of ChNWs, which is depicted in Scheme 1. Deacetylation of the raw chitin starting material was also examined to control the number of surface amino groups on the ChNWs and that of resultant grafted mPEG. Grafting was evaluated by several characterization methods: quantitative determination of amino groups by conductometric titration, Fourier-transform infrared spectroscopy (FT-IR), wide-angle X-ray diffractometry (WAXD), time-dependent changes in sedimentation heights in various electrolyte solutions, and formation of liquid crystals in concentrated suspensions.

Scheme 1. Strategy for mPEG Grafting to the Surface of ChNWs

■

EXPERIMENTAL SECTION

Materials. Commercial α-chitin powder from crab shells (Wako Pure Chemical Industries, Osaka, Japan) was used as the chitin starting material. mPEG with a number-averaged molecular weight of 2000 (mPEG2000) was purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). [Bis(acetoxy)iodo]benzene (BAIB) was purchased from either Sigma-Aldrich Co. (St. Louis, MO, USA) or Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan) and was used similarly. 2,2,6,6-Tetramethyl-1piperidinyloxyl free radical (TEMPO) and sodium cyanoborohydride (NaCNBH3) were purchased from Tokyo Chemical Industry Co. Ltd. (Tokyo, Japan). All other chemicals were purchased from Wako Pure Chemical Industries (Osaka, Japan). All reagents were used without further purification unless otherwise mentioned. Dichloromethane used in mPEG oxidation was purified by distillation. Deionized water was used throughout the study. Deacetylation of Chitin Powder. Deacetylation of the chitin powder was performed according to a previous study30,31 with slight modifications. In brief, α-chitin powder (20 g) was stirred in a 50 wt % aqueous NaOH solution (40 mL) at 40 °C for 1, 3, or 5 h, followed by dilution to 250 mL with water and careful addition of 3 M HCl to pH 11. The solution was then filtered and thoroughly washed with water, followed by freeze drying of the sample. Preparation of ChNWs Suspension. Aqueous suspensions of ChNWs were prepared from either untreated or deacetylated chitin powder, according to previous studies.13,31 In brief, chitin powder (2 g), either untreated or deacetylated, was hydrolyzed with 3 M HCl (50 mL) at 105 °C (reflux) for 1.5 h, followed by dilution with 50 mL water. The solution was then repeatedly centrifuged at 3000 rpm (ca. 1600g) for 5 min until aliquots of turbid supernatant were obtained. The combined turbid supernatant was dialyzed against deionized water using dialysis tubing (Viskase Companies, Darien, IL, USA, molecular weight cutoff 12−14 kg mol−1) until the surrounding water remained neutral. The resultant ChNWs suspension was further concentrated by osmotic compression, i.e., concentration by immersing the dialysis tubing containing ChNWs suspension in a 5−10 wt % aqueous solution of PEG with a molecular weight of 20 000. The concentrated suspension was finally sonicated using a 150 V/T unit (Biologics, Inc., VA, USA) for a few minutes. The nanowhisker samples obtained by acid hydrolysis (untreated, deacetylated for 1 h, and deacetylated for 3 h) were designated as ChNWs, DaC1hNWs, and DaC3hNWs, respectively. Preparation of mPEG-CHO. mPEG containing a terminal aldehyde group was prepared by TEMPO-mediated oxidation of terminal hydroxyl groups in mPEG, according to a previous study.32 In brief, TEMPO (30 mg, 1.92 × 10−4 mol) and BAIB (970 mg, 3.01 × 10−3 mol) were added to mPEG2000 (2 g, 1.00 × 10−3 mol −OH), which was vacuum-dried at room temperature for 2 h and dissolved in dry dichloromethane (10 mL), followed by overnight stirring. The reaction mixture was added dropwise to vigorously stirring diethyl B

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

ether (100 mL) to precipitate the polymer, which was subsequently collected by filtration and repeatedly washed with diethyl ether. Vacuum drying the precipitate yielded the mPEG2000 with terminal aldehyde groups (mPEG2000-CHO) in an almost quantitative yield. The degree of terminal oxidation was measured by 1H NMR spectroscopy32 to be 78.2 ± 2.3%. mPEG Grafting onto ChNWs. To an aqueous suspension of ChNWs, DaC1hNWs, or DaC3hNWs (200 g, 1 wt %, containing 2 g of ChNWs), mPEG2000-CHO (2 mol equiv against the total surface amino groups of the ChNWs) was added, followed by stirring for 2 h at room temperature until complete dissolution. After further dissolution of NaCNBH3 (equivalent to mPEG2000-CHO), the mixture was stirred at room temperature for 24 h, adjusting the pH of the system to 5−6 by careful addition of 2.5 M HCl. The resultant mixture was dialyzed against deionized water, using dialysis tubing with a molecular weight cutoff of 100 000 (SpectraPor CE, Spectrum Laboratories Inc., CA, USA), for 7 days, changing the water twice a day. The grafted ChNWs samples are designated hereafter by an addition of the prefix “g-”; for example, g-DaC1hNWs were prepared from DaC1hNWs by grafting reactions. Characterization. Surface amino group content of the ChNWs was determined by conductometric titration, according to a previous study,33 with a slight modification. A mixture of an aqueous ChNWs suspension (50 mL, 1.0 wt %) and 0.5 M aqueous HCl (1 mL) was titrated using 0.01 M aqueous NaOH at an addition rate of 0.5 mL/30 s. Conductivity and pH values of the system were simultaneously monitored using a pH/conductivity meter (Seven Go Duo-SG78, Mettler-Toledo, OH, USA). The obtained conductivity curve plotted against added NaOH typically indicates two reflection points, the first and second of which correspond to the neutralization of an added amount of HCl (= 5 × 10−4 mol) and that of surface amino groups, respectively. Surface amino group content of the ungrafted ChNWs, Aungra (μmol/g ChNWs), was calculated from the following eq 1 A ungra =

0.01 × v wChNWs

WTG =

wTG =

A ungra − x 0.01 × v = wgraft 1 + 2000 × x × 10−6

Cr =

x 1−x

(5)

I020 − Iam × 100 I020

(6)

Sizes and shapes of the ChNWs were observed using transmission electron microscopy (JEOL JEX-2100, JEOL Ltd., Tokyo, Japan). A drop of very dilute aqueous suspensions of the samples was deposited on a grid covered with a carbon-coated Formvar film, which was previously treated with 0.2% aqueous bacitracin solution. After drying, the samples were observed at 80 kV using the defocus-contrast technique. Evaluation of Suspension Dispersion Stability. The dispersion stability of the grafted and ungrafted samples, in the presence or absence of various electrolytes, was evaluated by visual observation of the suspensions, light transmittance measurements of the suspensions, and measurement of the sediment heights. All prepared samples were conditioned to different levels of final electrolyte concentration (0.05, 0.1, and 0.5 M) and the same sample solid content (0.05 wt %, including the weight of mPEG for the grafted samples) using various electrolytes (NaCl, NaOH, HCl, and CaCl2). Flow birefringence of the obtained ChNWs−electrolyte mixtures were observed between two crossed polarizers. UV−vis transmittance spectra of the suspensions were measured with a UV−vis spectrophotometer (UV-2700, Shimadzu Corporation, Kyoto, Japan) in the range of 300−700 nm using deionized water as a blank sample. Semiquantitative assessments were performed by sedimentation height measurements over time in electrolyte solutions. Suspensions of the grafted and ungrafted samples were mixed with aqueous electrolyte solutions, including NaCl or CaCl2, to obtain final electrolyte concentrations of 0.05−0.5 M and final sample concentrations of 0.05 wt % (weight ratio of ChNWs samples, including mPEG, if grafted). Changes in the sediment height of the samples, normalized by the heights of the starting suspensions, with or without electrolytes, were recorded up to 10 h after vigorous mixing. Evaluation of Liquid Crystal Formation. One of the samples (gDaC1hNWs) was observed for liquid crystal formation upon concentration. The sample was concentrated to up to 15.2 wt % sample solid content (including mPEG) by osmotic compression:35 dialysis tubing containing the suspension was immersed in an aqueous solution of PEG20000 until the suspension attained the desired concentration. The concentrated suspension was then diluted with deionized water to various concentrations. The aliquots were then sonicated using a rod-type ultrasonic homogenizer (150 V/T, Biologics Inc., VA, USA), and they were allowed to stand at room temperature to observe phase separation.

(1)

(2)

The weight of the grafted mPEG2000 obtained by titration, wtitr, can then be calculated as follows wtitr = 2000 × x × 10−6

(4)

where WTG is the weight of grafted mPEG per weight of grafted samples (g/g) and wTG is the weight of grafted mPEG per weight of ChNWs (g/g, excluding mPEG). Wgraft, WChNWs, and WmPEG are residue ratios (in %) of the grafted samples, ChNWs, and mPEG2000CHO at 500 °C, respectively, against their weight at 100 °C. Wide-angle X-ray diffractometry (WAXD) of the grafted and ungrafted ChNWs was performed using the freeze-dried samples pressed into tablets (8 mm diameter and 0.3 mm thick) with a rotating anode X-ray generator (Rigaku RU-200BH, Rigaku Corporation, Tokyo, Japan). Samples were irradiated with Ni-filtered Cu Kα radiation (λ = 0.1542 nm) generated at 40 kV and 150 mA, and diffraction patterns were recorded in the range of 2θ = 5−35°. Silicone was added to each sample as an internal standard, and diffraction profiles were calibrated to the diffraction of silicone (2θ = 28.46°) and that of crystalline chitin (2θ = 9.3°). According to a previous study,34 the degree of crystallinity of the samples, Cr, was calculated from an intensity of (020) plane (I020, at 2θ = 9.24°), which is indicative of an amorphous halo (Iam, at 2θ = 16.0°)

where v (mL) is the volume of 0.01 M NaOH required for the neutralization of surface amino groups (i.e., volume between the first and second reflection points) and wChNWs (g) is the weight of ChNWs used for titration. Although a similar equation will give an amino content for the grafted samples, mere subtraction of the obtained value from the ungrafted ChNWs will not yield the amount of grafted mPEG2000. This is due to the fact that the weight of the grafted samples, wgraft, also contains the weight of the grafted mPEG; therefore, the amino content for the grafted samples, Agraft, is per total sample weight and not per ChNWs weight. Therefore, the calculated Agraft values can be related to the molar numbers of the grafted mPEG2000, x, by the following eq 2

Agraft =

Wgraft − WChNWs WmPEG − WChNWs

(3)

Attenuated total reflectance Fourier-transform infrared (ATRFTIR) measurements of the freeze-dried samples were performed using a Shimadzu IRPrestage-21 spectrometer (Shimadzu Corporation, Kyoto, Japan), equipped with a diamond ATR accessory (SensIR Technologies DurasampleIR II, Smiths Detection, UK), in air at 4 cm−1 resolution and 32 scans. Thermogravimetry (TG) was performed using ThermoPlus EVOII TG-8120 (Rigaku Corporation, Tokyo, Japan). Freeze-dried samples packed in an aluminum pan were subjected to heating under air from room temperature to 500 °C at 10 °C/min. Residual weight values of heated mPEG-CHO, ungrafted ChNWs (including deacetylated samples), and grafted samples at 500 °C were converted to the value of grafted mPEG using the following equations C

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

■

Article

Evaluation of Amino Group Content by Conductometric Titration. The amount of grafted mPEG was evaluated from the decrease in surface amino groups of ChNWs consumed by grafting rather than weight variation of the samples because the latter sometimes decreases after grafting, presumably owing to handling losses. Surface amino groups on ChNWs act as weak acids at pH lower than their pKa (6.3)33 and can be titrated by a standard basic solution to give the amino content. Figure 1 shows conductometric titration curves

RESULTS AND DISCUSSION Strategy for Steric Stabilization of ChNWs. Most previous studies on the synthesis of chitin derivatives and surface modification of ChNWs have utilized primary amino groups at the C2 position of glucosamine residues in chitin molecules generated via deacetylation, i.e., hydrolysis of Nacetamide groups.4 To achieve steric stabilization of ChNWs, we also considered the grafting of mPEG to surface amino groups of ChNWs. A strategy employing an amidation reaction between surface amino groups on ChNWs and terminal carboxyl groups on mPEG using 1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydrochloride (EDC) was attempted because a similar method successfully yielded sterically stable suspensions of cellulose NWs in our previous examinations.26 However, the application of this strategy to ChNWs failed to yield sterically stabilized ChNWs. In addition, FT-IR analysis of the obtained sample indicated almost no grafting of mPEG (data not shown). Although the reasons for these unsuccessful results are still unclear, the following mechanism may be plausible. Most amidation reactions, including those mediated by EDC, are based on the nucleophilic attack of amino groups to electrophilic carbons in activated carboxyls. Therefore, the above-mentioned amidations may be disadvantageous because the amino groups are stabilized on ChNWs and their mobility is quite limited. An alternative strategy involving reductive amination between the primary amino groups on the ChNWs surface and terminal aldehyde groups of mPEG (Scheme 1), which was previously employed for the preparation of graft copolymers of chitosan and mPEG,36,37 was also examined. The reaction readily proceeded in an aqueous system under mild conditions, pH = 5−6 at room temperature. Oxidation of mPEG2000. The conversion of the terminal hydroxyl groups in mPEG to aldehydes was performed by TEMPO-mediated oxidation, reported by Masson et al.,32 using TEMPO as an oxidizing agent and [bis(acetoxy)-iodo]benzene (BAIB) as a co-oxidant, in dichloromethane. While contamination of dihydroxy-PEG in mPEG may sometimes result in a degree of oxidation larger than the original hydroxyl content, we previously confirmed the presence of an equal number of methoxy and hydroxyl groups through analysis using trichloroacetyl isocyanate.38 The 1H NMR spectrum (Figure S1, Supporting Information) of the oxidized mPEG, mPEG2000-CHO, possesses signals identical to those reported by Masson et al.32 The degree of oxidation was calculated as the ratio of conversion from hydroxyls to aldehydes using the integrated areas corresponding to methylene protons adjacent to aldehydes compared with those of terminal methoxy protons. This was found to be 78.2 ± 2.3% for our samples, which is relatively low because the values in the previous studies32,37 indicated almost quantitative conversion to aldehydes. Modification of reaction conditions, including prolonged reaction times of up to 2 days or the use of strictly purified solvents, did not improve the degree of oxidation. Therefore, at present, the reasons for our poor oxidation results are still unclear. For the following experiments, the grafting reactions were performed on the basis of the molar equivalence of the surface amino groups on ChNWs, measured by conductometric titration (see below), and terminal aldehyde groups and not on the basis of the molar amount of mPEG, considering the degree of oxidation obtained above.

Figure 1. Conductometric titration curves of (a) 500 mg of ChNWs and (b) 297 mg of g-ChNWs. Gray areas correspond to neutralization of surface amino groups.

for ChNWs and g-ChNWs; the results of other samples are summarized in the Supporting Information. Whereas the surface amino group content can be calculated from these results, those of the grafted samples should be recalculated because the titration results are against sample weight, including ChNWs and mPEG, and not for bare ChNWs (see the Experimental Section for details). The calculated surface amino group contents are summarized in Table 1. The amino content of ChNWs was almost comparable to that reported in previous studies.31,33,39 The values for DaC1hNWs and DaC3hNWs also clearly indicate that the pretreatment of the chitin starting material using 50% aqueous NaOH at 40 °C successfully yielded ChNWs with increased amino content, similar to previous studies.30,31 After the grafting reactions, all three ungrafted NWs indicated decreased amino content, which further corresponds to a grafting amount of 0.2−0.3 g mPEG/g ChNWs. The ratios of consumed amino groups against total amino content of the starting samples were 38.0, 20.4, and 20.1% for g-ChNWs, g-DaC1hNWs, and g-DaC3hNWs, D

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

intensified absorption at 2881 and 1114 cm−1, which is attributed to C−H stretching and C−O stretching of amorphous PEG chains, respectively. Conversely, signals attributed to crystalline PEG chains42 at 963, 1102 (C−O stretching and CH2 rocking), and 1345 (CH2 wagging and twisting) were not observed. The above-mentioned observations together with the following WAXD results suggest the amorphous state of the grafted mPEG chains in the dry samples. WAXD Measurements. WAXD profiles of all of the samples are shown in Figure 3. The diffraction pattern for

Table 1. Surface Amino Group Content, Aungra and Agraft, and Amount of Grafted mPEG, wtitr, of the Samples Determined by Conductometric Titration amino group content, μmol/g samples samples

Aungra

Agraft

wtitr, g/g ChNWs

area covered by a single polymer, nm2

ChNWs DaC1hNWs DaC3hNWs

366 525 768

178 344 118

0.278 0.214 0.310

4.3 5.6 3.8

respectively. Considering the density of α-chitin crystals of 1.425 g cm−340 and assuming the size of a single ChNW as 8 × 8 × 200 nm3, the area of ChNWs surfaces occupied by a single mPEG2000 could also be calculated and are listed in the last column of Table 1. The amount of grafted mPEG appears to plateau with increasing surface amino group content; the grafting of g-ChNWs and g-DaC1hNWs were comparable, whereas the value for g-DaC3hNWs was slightly increased but did not correspond with a large increase in surface amino groups. This behavior may result from the steric repulsion of the unreacted mPEG with the previously grafted mPEG, resulting in a saturation of grafting sites. FT-IR Analysis. FT-IR spectra of all the samples are shown in Figure 2. Similar to a previous study,41 the spectra of

Figure 3. WAXD profiles of (a) ChNWs, (b) g-ChNWs, (c) DaC1hNWs, (d) g-DaC1hNWs, (e) DaC3hNWs, and (f) gDaC3hNWs. The asterisk indicates a reflection corresponding to silicone.

ChNWs was almost identical to that observed in a previous study.31 DaC1hNWs and DaC3hNWs profiles indicated a decrease in their degree of crystallinity (Cr). The Cr values for ChNWs, DaC1hNWs, and DaC3hNWs calculated from eq 6 were 74.6, 55.2, and 47.8%, respectively. This decrease in the degree of crystallinity, which was also reported in a previous study,31 suggests a partial swelling of ChNWs, with a 50% aqueous NaOH solution, into alkali chitin.43 The profiles observed before and after grafting of mPEG were almost identical; diffractions of crystalline PEG at 19.1° and 23.3°44 were not observed for the grafted samples (Figure 3b,d,f). This suggests the noncrystallinity of the grafted mPEG on the surface, which corresponds to the state determined by FT-IR measurements, as described above. Assuming the dimensions of ChNWs and deacetylated ChNWs to be 8 × 8 × 200 nm3, the area occupied by a single mPEG chain would be 3.8−5.6 nm2 (Table 1), whereas the contour length of mPEG2000 is 13.18 nm.45 These values imply the possibility of two scenarios: one is the formation of entangled, insufficiently crystallized, multiple grafted mPEG chains, and the other is the interaction between the ChNWs surface and the grafted mPEG, which is

Figure 2. FT-IR spectra of (a) mPEG2000-CHO, (b) ChNWs, (c) gChNWs, (d) DaC1hNWs, (e) g-DaC1hNWs, (f) DaC3hNWs, and (g) g-DaC3hNWs.

ChNWs, DaC1hNWs, and DaC3hNWs exhibited characteristic absorption at 3475, 3260, 1656, and 1554 cm−1, which is attributed to O−H stretching, N−H stretching, amide I, and amide II, respectively. Intensities of the last two absorptions did not decrease significantly after deacetylation, as shown in Figure 2d,f, probably because of the very limited numbers of deacetylation events, which maintain chitin crystallinity. The spectra of the grafted samples (Figure 2c,e,g) indicate E

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Table 2. Weight Residues of Various Samples at 500 °C and the Amount of Grafted mPEG, wTG, Estimated from TG Measurements

more favored than that between the grafted mPEG chains. The possibility of the former was predicted in a previous study26 using relatively short mPEG (mPEG1000); however, the present results appear to be surprising because the extension of the PEG to twice its length provided no mPEG crystallization. The latter scenario, which is also plausible even without direct evidence, may be speculated from the previous study. A strong hydrogen-bonding interaction between hydroxyl groups at the C6 position of cellulose and oxygen atoms in the PEG main chain was proposed by Kondo et al. from precise analyses of blend polymers comprising PEG and cellulose.46 While a similar investigation has not been performed for chitin and PEG, the structural analogy of the former to cellulose strongly suggests a similar strong interaction between them, as well as a resultant strong adsorption of grafted mPEG onto the surface of ChNWs in the present samples after drying. Thermogravimetric Analysis. Figure 4 shows the thermogravimetric (TG) curves of the grafted and ungrafted

samples

weight residues at 500 °C, %

wTG, g/g ChNWs

mPEG2000-CHO ChNWs DaC1hNWs DaC3hNWs g-ChNWs g-DaC1hNWs g-DaC3hNWs

2.82 16.0 15.1 15.9 11.9 11.7 9.53

0.451 0.383 0.949

(see Figure S3 in the Supporting Information for details). While reasons for the observed overestimation of mPEG content obtained by TG measurements are still unclear at present, some overdecomposition of the grafted ChNWs together with mPEG under air might be responsible for this phenomenon. Previous studies predicted that the thermal oxidative degradation of PEG under air generates several species, including formate esters, hydroperoxy groups, and formic acid,47 which might decompose ChNWs more effectively during our measurements and result in a larger weight decrease of the grafted samples. Electron Microscopy. Figure 5 shows electron micrographs of various ChNWs. No significant change in the size and shape of ChNWs was observed before and after grafting of mPEG (Figure 5a,b), whereas significant fragmentation and jaggedness of the ChNWs contours were observed for the deacetylated samples (Figure 5c,d). These observations suggest that the grafting reactions employed in the present study did not affect the morphology of the ChNWs, whereas our deacetylation condition of treatment with 50% NaOH(aq) at 40 °C caused partial swelling of the ChNWs, which corresponds to WAXD results showing a decrease in the Cr values of ChNWs for longer deacetylations. In forthcoming investigations, we will employ different deacetylation conditions (40 wt % NaOH(aq) at 40 °C for 1−5 h), which were discovered separately abreast the present investigation and were found to have no significant effect on ChNWs morphology. Dispersion Stability Evaluation by Visual Observation. Qualitative evaluation of the dispersion stability of the grafted and ungrafted ChNWs was observed in the presence of different electrolytes: NaCl, HCl, and NaOH. The ungrafted ChNWs lost their flow birefringence immediately after addition of any electrolyte at 0.1 M, whereas all of the grafted samples indicated remarkable flow birefringence with any electrolyte at a similar concentration. Figure 6 shows the aqueous suspensions of various grafted and ungrafted ChNWs in the presence of different electrolytes, indicating marked dispersion stability of the three grafted samples in comparison with that of the ungrafted sample. The appearance in Figure 6b is especially significant because the pH of the 0.1 M NaOH system is higher than the pKa value of the surface amino groups on the ChNWs (6.3); thus, repulsion between surface amino groups is nonexistent. Therefore, the remarkable dispersion stability in NaOH, as shown in Figure 7, can be entirely attributed to surface mPEG grafting. UV−vis transmittance spectra of the ChNWs and g-ChNWs suspensions in various types of electrolyte solutions are shown in Figure 8. In any electrolyte solutions, ChNWs (dotted lines) showed reduced transmittance due to light scattering by aggregations of the ChNWs formed by addition of the

Figure 4. Thermogravimetric curves of mPEG2000-CHO, ChNWs, gChNWs, DaC1hNWs, g-DaC1hNWs, DaC3hNWs, and gDaC3hNWs.

samples. The small vibrations observed at around 330 °C arose from too rapid endotherms due to extremely fast degradations of the samples. These rapid endotherms provide a deficit of heating energy, insufficient heating of the samples, and slight transient decline in the sample temperature. A previous study26 reported the two-step decomposition (i.e., weight decrease) of cellulose nanowhiskers grafted with mPEG because of the sequential decompositions of nanowhiskers and mPEG. Although a similar decomposition behavior was expected for our grafted ChNWs, only the g-ChNWs displayed this phenomenon, whereas the g-DaC3hNWs did not. Decomposition of pure mPEG2000 was initiated at ca. 200 °C, whereas decomposition of other samples, i.e., grafted and ungrafted ChNWs, was observed above 250 °C. Weight residues of the samples at 500 °C are summarized in Table 2. The amounts of grafted mPEG were calculated from these values using eqs 4 and 5 and are also summarized in Table 2. As expected, the results obtained by TG measurements were considerably higher than those obtained by conductometric titration, as shown in Table 1. We speculate that the results from TG measurements were overestimated for unknown reasons because physical mixtures of ChNWs and mPEG2000, at the ratios specified in Table 2, did not yield the FT-IR spectra shown in Figure 2, whereas FT-IR spectra of the mixtures, at the ratio shown in Table 1, agreed with Figure 2 F

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Figure 5. Electron micrographs of (a) ChNWs, (b) g-ChNWs, (c) g-DaC1hNWs, and (d) g-DaC3hNWs. Scale bars are 1 μm.

Figure 7. Appearance of a g-ChNWs suspension in 0.1 M NaOH(aq), indicating brilliant flow birefringence. The sample was placed between crossed polarizers. (The sample shown here is the same as that in Figure 7H in ref 1.)

results indicate that the ChNWs were readily aggregated with the addition of these three types of electrolytes (0.1 M HCl, NaOH, and NaCl), whereas the dispersion states of g-ChNWs in the presence of these electrolytes are exactly identical to those without electrolytes and are unaffected by the type of electrolyte. Dispersion Stability Evaluation by Sedimentation Rate Measurements. To semiquantitatively assess and compare the stability, the rate of sedimentation of ChNWs and grafted ChNWs over time were measured in the presence and absence of two electrolytes, NaCl and CaCl2. Figure 9 and Table 3 summarize these results. The three ungrafted ChNWs samples, i.e., ChNWs, DaC1hNWs and DaC1hNWs, eventually formed aggregates whose rates of sedimentation were dependent on the valences and concentrations of electrolytes. In addition, a general trend was observed in which longer reaction times for the deacetylation of the chitin starting material gave larger amounts of surface amino groups on the ChNWs, providing higher stability and resistance to salt-induced aggregation. The detailed results are as follows: in 0.05 M NaCl, ChNWs rapidly aggregated, whereas DaC1hNWs and DaC3hNWs showed moderate (up to 100 min) and permanent stability, respectively. The same concentration of CaCl2 made these suspensions more unstable, of which only DaC3hNWs was stable up to 450 min. Increasing salt concentration to 0.1

Figure 6. Appearance of suspensions of ChNWs, g-ChNWs, gDaC1hNWs, and g-DaC3hNWs (from left to right) in (a) 0.1 M HCl, (b) 0.1 M NaOH, and (c) 0.1 M NaCl.

electrolytes. On the contrary, all of the g-ChNWs (solid lines) indicated higher transmittance spectra, which were completely identical to those of the well-dispersed ChNWs and g-ChNWs suspensions containing no electrolytes (data not shown). These G

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Figure 8. UV−vis transmittance of aqueous suspensions of ChNWs (dotted lines) and g-ChNWs (solid lines) dispersed in (a) 0.1 M HCl, (b) 0.1 M NaOH, and (c) 0.1 M NaCl. Sample concentration is 0.05 wt % for all samples.

term stability of up to 50 min. A salt concentration of 0.5 M caused rapid loss of flow birefringence of all samples, suggesting strong aggregation at this electrolyte concentration. The minimum durations until loss of flow birefringence, which are summarized in Table 3, correspond well to the sedimentation curves shown in Figure 9, showing faster sedimentation and lower sediment height of the more unstable samples. In a striking contrast, all three grafted samples indicated good dispersion without any loss of flow birefringence up to 10 h, as shown in Table 3, as well as no formation of clear supernatant in the presence of any electrolytes at any concentration, as shown in Figure 9. These observations clearly indicate the strong effect of mPEG grafting on the dispersion stability of ChNWs. Observation of Liquid Crystal Formation. The ungrafted ChNWs have been reported to form a chiral nematic liquid crystalline phase above a critical concentration,13,31,33 similar to cellulose nanowhiskers.12,26,35 The effect of mPEG grafting on the formation of chiral nematic liquid crystals in the present system is of considerable interest because some sterically stabilized nanowhisker systems have been reported to form chiral nematic liquid crystals,25,26,48−50 whereas another study employing different conditions did not exhibit this phenomenon.51 We examined the effects of the concentration of a grafted sample, g-ChNWs, to observe whether it exhibits the typical phase-separation phenomenon. However, the results were undesirable because the aqueous suspensions of g-ChNWs did not show phase separation even after concentration up to 8.59−15.2% solid content (including the grafted mPEG2000), thorough sonication treatments, and standing for several days. This was in contrast to the separation of the ungrafted ChNWs at concentrations above ca. 5 wt % 1 day after sufficient sonication. At 8.59 wt % solid content, the viscosity of the suspension was too low to easily flow, and it did not separate. At 10.2%, the sample indicated an apparent yield stress, i.e., it was gel-like but flowed upon inclination of the container. The sample was a thixotropic gel at 12.1%; it flowed after vigorous shaking but gelated after standing overnight. At 15.2%, the sample turned into a stable gel, which was quite difficult to separate. Although the reason for this absence of a liquid crystalline phase separation is still unclear, we speculate that it is due to a sharp rise in system viscosity via the following mechanism: the critical concentration for the phase separation of sterically stabilized systems is considerably higher than that of electrostatically stabilized systems because in the latter the electrostatic repulsion causes independent particles to be farther apart and renders a larger apparent particle size. It

Figure 9. Changes in sediment height over time for (a) ChNWs and gChNWs, (b) DaC1hNWs and g-DaC1hNWs, and (c) DaC3hNWs and g-DaC3hNWs. Cross and circle symbols indicate that the type of added electrolyte is NaCl and CaCl2, respectively. Colors indicate electrolyte concentrations: green for 0.05 M, orange for 0.1 M, and red for 0.5 M. Blue symbols are the results of grafted samples containing 0.5 M electrolytes, and other colors are those for ungrafted samples (in panel c, blue crosses, blue circles, and green crosses are overlapping).

M also induced rapid aggregation of most samples apart from DaC3hNWs in 0.1 M NaCl, which showed a relatively shortH

dx.doi.org/10.1021/bm5016078 | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

Table 3. Minimum Time until Grafted and Ungrafted Samples Lose Birefringence in the Presence of Various Types and Concentrations of Electrolytes 0.05 M

0.1 M

0.5 M

samples

NaCl

CaCl2

NaCl

CaCl2

NaCl

CaCl2

three g-samples ChNWs DaC1hNWs DaC3hNWs

+ −