COMMUNICATION pubs.acs.org/Biomac
Viscoelastic Evaluation of Average Length of Cellulose Nanofibers Prepared by TEMPO-Mediated Oxidation Daisuke Ishii,*,† Tsuguyuki Saito,‡ and Akira Isogai‡ †
Department of Materials Chemistry, Faculty of Science and Technology, Ryukoku University, 1-5 Yokotani, Seta Oe-cho, Otsu-shi, Shiga 520-2194, Japan ‡ Department of Life Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan
bS Supporting Information ABSTRACT: Dynamic viscoelasticity measurements were performed for aqueous dispersions of cellulose nanofibers prepared by TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl radical)-mediated oxidation and subsequent mechanical disintegration in water. The frequency dependence of the storage and loss moduli of 0.02% (w/v) dispersions of TEMPO-oxidized cellulose nanofibers in water showed terminal relaxation behavior at relatively lower angular frequencies. This strongly suggests that each cellulose nanofiber in the dispersion behaves as a semiflexible rod-like macromolecular chain or colloidal particle. Furthermore, a clear boundary was observed between the terminal relaxation and rubbery plateau regions. The longest viscoelastic relaxation time, τ, was estimated from the angular frequency, corresponding to the boundary point, and the average length of the cellulose nanofibers, L, was estimated using the equation τ = πηsL3/[18kBT ln(L/d)]. The equation gave a value of L = 2.2 μm, which was in good agreement with TEM observations.
’ INTRODUCTION Native cellulose is an assembly of cellulose microfibrils, a nanofibrous material of a width ranging from 4 to 20 nm, depending on its origin. Individualization of the cellulose microfibrils has long been considered to be difficult because they are bundled via numerous hydrogen bonds and van der Waals interactions. However, several research groups have recently succeeded in the preparation of nanofibrous cellulosic materials based on the dissociation of each cellulose microfibril.1 Saito et al. recently showed that TEMPO-mediated oxidation of native celluloses, followed by gentle mechanical disintegration of the oxidized celluloses in water, leads to individualization of cellulose microfibrils.2-5 Individualized cellulose nanofibers maintain a stable dispersion state in aqueous media through electrostatic repulsion because they possess negatively charged surfaces on which one C6 carbon in every two glucopyranose units is oxidized to a carboxyl group.6 The individualized cellulose microfibrils can be processed to novel cellulosic films with high transparency, high mechanical strength, and high gas-barrier capability,7 which opens up new potential for cellulosic materials in general and specialized areas. Structural aspects, such as average diameter and length, affect the processing conditions and the resultant properties of nanosized materials. Microscopic methods such as optical, electron, or scanning probe microscopy offer clear and instant understanding of structural features; these have been used to estimate the size of cellulosic nanocrystals.8 However, observation of numerous r 2011 American Chemical Society
particles is required for statistical reliability in size evaluation. In addition, in the case of TEMPO-oxidized cellulosic nanofibers, a high aspect ratio (length/diameter) prevents precise estimation of their length; because of this shortcoming, alternative methods such as scattering experiments have been employed to complement microscopic evaluation.9 Rheological measurements have been used to characterize colloidal systems as well as molecular solution systems.10-14 Structural parameters such as length, diameter, and aspect ratio define the hydrodynamic properties of the colloidal particles and thus affect the rheological properties. These are also affected by interfacial properties of the particles, such as surface charge density and affinity with the surrounding media (hydrophilicity or hydrophobicity), which control particle-particle and particle-solvent interactions. This in turn enables the characterization of the structural features of colloidal particles, with some prerequisites, for example, stable and monodisperse distribution of the particles without aggregation. In this Communication, we show that information on the average length of TEMPO-oxidized cellulose nanofibers may be extracted from viscoelasticity data obtained from aqueous dispersions. The effect of the surface charge of the nanofibers on the estimation of their length is also briefly discussed. Received: November 19, 2010 Revised: January 10, 2011 Published: January 24, 2011 548
dx.doi.org/10.1021/bm1013876 | Biomacromolecules 2011, 12, 548–550
Biomacromolecules
COMMUNICATION
parameters such as length and diameter. However, simulation of whole viscoelasticity spectra requires consideration of all factors mentioned above and complicated mathematical treatment. Alternatively, as a simple treatment, we focus on the longest relaxation time estimated from the spectra, because the dependence of G0 against ω exhibits a rather distinct boundary between terminal relaxation and rubbery plateau regions. In the theory of linear viscoelasticity, the inverse of angular frequency at the boundary corresponds to the longest relaxation time, which reflects the rotational diffusivity of a polymer chain. The length of a polymer chain thus affects the longest relaxation time. A boundary is observed in polymeric or colloidal systems in which the polymer chains or colloidal particles are of monodisperse length. The distinct boundary observed for the TEMPO-oxidized cellulose nanofibers suggests uniformity of fiber length. We estimated the average fiber length of the nanofibers by applying the theory of linear viscoelasticity of semiflexible rodlike polymers in a dilute solution. The longest relaxation time of a semiflexible polymer chain, τ, is estimated as follows17,18
Figure 1. Viscoelastic storage (G0 ) and loss (G00 ) moduli, plotted logarithmically against angular frequency of oscillatory shear, ω, of TEMPOoxidized cellulose nanofiber aqueous dispersions with different concentrations.
’ EXPERIMENTAL SECTION Preparation of TEMPO-Oxidized Cellulose Nanofibers. Once-dried bleached hardwood kraft pulp was used as the cellulosic starting material. TEMPO-mediated oxidation and subsequent mechanical disintegration were performed according to a previously reported method.3 The resulting TEMPO-oxidized cellulose nanofiber, with the H atom of its carboxyl groups replaced by sodium (-COONa), had an average diameter of 4 nm and carboxyl content of 1.5 mmol g-1, with a zeta potential of -80 mV.4 Viscoelasticity Measurement. A stress-controlled rheometer (HAAKE Rheostress RS600, Karlsruhe, Germany) was used. Measurements were performed using a cone-plate geometric sensor with a diameter of 60 mm and cone angle of 1°. The sample temperature was adjusted to 20 °C. The strain amplitude was set to 1% for dispersions with the solid concentrations of >0.04% (w/v) or to 5% for dispersions with the solid concentrations of 0.03% (w/v) showed values of G0 and G00 that were almost independent of ω, which suggests the formation of a gel-like transient network structure of entangled nanofibers. In contrast, dispersions with concentrations of 0.02% (w/v) showed power-law type terminal relaxation behavior at ω < 0.46 s-1, with G0 and G00 obeying the following relations G0 ω2 , G} ω ð1Þ The terminal relaxation behavior revealed that each nanofiber behaves as a semiflexible rod-like macromolecular chain, to which the molecular theory of linear viscoelasticity is applicable.15-18 However, the 0.01% (w/v) dispersion did not follow the same trend; here G00 was proportional to the square of ω. Therefore, in the following sections, theoretical calculations based on linear viscoelasticity were carried out using the data for the 0.02% (w/v) dispersion alone. The viscoelastic properties of the dispersions in the terminal relaxation region reflect the characteristics of the fibrous particles, for example, stiffness, frictional force at the surface between the particle and the surrounding solvent, and dimensional
τ ¼ πηs L3 =½18kB T lnðL=dÞ
ð2Þ
where L, d, T, ηs, and kB are the chain length, chain diameter, absolute temperature, solvent viscosity, and Boltzmann constant, respectively. In this case, we used the following values: T = 293 K, ηs = 1 mPa 3 s (viscosity of water at 293 K), and d = 4 nm. The value of τ was 7.32 10-2 s, calculated from the angular frequency at the crossover point; the value of ω was obtained from the equation τ = 1/(2πω) as 0.46 s-1. Substitution of these values into eq 2 gave L = 2.2 μm, which is of the same order of magnitude as the value obtained from the TEM image.3
’ DISCUSSION The viscoelastic properties of cellulosic colloidal systems have been investigated by various research groups, using cellulosic particles with different morphologies.19,20 Tatsumi et al. found that a suspension of cellulosic pulp fiber and a molecularly dispersed solution of cellulose in lithium chloride/N,N-dimethylacetamide showed similar viscoelastic properties.20 For both the colloidal and the solution system, G0 was dependent on the mass fraction of cellulose. However, conventional cellulosic fibrous suspensions showed gel-like viscoelastic behavior due to network formation because of the high aspect ratios and strong attractive interactions of the fibrous particles. Therefore, it was impossible to extract information on a single cellulosic particle from the viscoelastic data. In our system, TEMPO-oxidized cellulose nanofibers were dispersed individually in aqueous media, which allowed structural evaluation of single nanofibrous particles. For precise size evaluation, the effect of surface charge on the viscoelastic properties of the TEMPO-oxidized cellulose nanofibers must be considered. Electrostatic properties affecting the viscoelasticity of charged colloidal systems are often described in terms of Debye-H€uckel theory.21,22 In this theory, the magnitude of electrostatic interaction between ionic species is evaluated using Debye screening length, κ-1, which is estimated as follows K - 1 ¼ ð8πNA lB IÞ - 1=2
ð3Þ
lB ¼ e2 =ð4πεkB TÞ
ð4Þ
where
549
dx.doi.org/10.1021/bm1013876 |Biomacromolecules 2011, 12, 548–550
Biomacromolecules and
COMMUNICATION
1X 2 ci zi I ¼ 2 i
(4) Saito, T.; Hirota, M.; Tamura, N.; Kimura, S.; Fukuzumi, H.; Heux, L; Isogai, A. Biomacromolecules 2009, 10, 1992–1996. (5) Isogai, A.; Saito, T.; Fukuzumi, H. Nanoscale 2011, 3, 71–85. (6) Hirota, M.; Furihata, K.; Saito, T.; Kawada, T.; Isogai, A. Angew. Chem., Int. Ed. 2010, 49, 7670–7672. (7) Fukuzumi, H.; Saito, T.; Iwata, T.; Kumamoto, Y.; Isogai, A. Biomacromolecules 2009, 10, 162–165. (8) Elazzouzi-Hafraoui, S.; Nishiyama, Y.; Putaux, J.-L.; Heux, L.; Dubreuil, F.; Rochas, C. Biomacromolecules 2008, 9, 57–65. (9) Branca, C.; Magazu, V.; Mangione, A. Diamond Relat. Mater. 2005, 14, 846–849. (10) Hough, L. A.; Islam, M. F.; Janmey, P. A.; Yodh, A. G. Phys. Rev. Lett. 2004, 93, 168102. (11) Echeverria, I.; Urbina, A. Eur. Polym. J. 2006, B50, 491–496. (12) Xu, J.; Chatterjee, S.; Koelling, K. W.; Wang, Y.; Bechtel, S. E. Rheol. Acta 2005, 44, 537–562. (13) Parra-Vasquez, A. N. G.; Stepanek, I.; A. Davis, V. A.; C. Moore, V. C.; Haroz, E. H.; Shaver, J.; Hauge, R. H.; Smalley, R. H.; Pasquali, M. Macromolecules 2007, 40, 4043–4047. (14) Matsumoto, T. Nihon Reoroji Gakkaishi 2004, 32, 3–9. (15) Shankar, V.; Pasquali, M.; Morse, D. C. J. Rheol. 2002, 46, 1111–1154. (16) Kirkwood, J. G.; Auer, P. L. J. Chem. Phys. 1951, 19, 281–283. (17) Warren, T. C.; Schrag, J. L.; Ferry, J. D. Biopolymers 1973, 12, 1905–1915. (18) Ullman, R. Macromolecules 1969, 2, 27–30. (19) Lassenguette, E.; Roux, D.; Nishiyama, Y. Cellulose 2008, 15, 425–433. (20) Tatsumi, D.; Ishioka, S.; Matsumoto, T. Nihon Reoroji Gakkaishi 2002, 30, 27–32. (21) Shulka, A.; Rehage, H. Langmuir 2008, 24, 8507–8513. (22) Caplan, M. R.; Moore, P. N.; Zhang, S.; Kamm, R. D.; Lauffenburger, D. A. Biomacromolecules 2000, 1, 627–631.
ð5Þ
Here NA, I, e, ε, c, and z are the Avogadro constant, ionic strength (mol m-3), elementary charge, dielectric constant of the medium (water), electrolyte concentration, and charge number of the electrolytes, respectively. The parameter lB is known as the Bjerrum length and reflects the spatial scale at which there is a balance between thermal (kBT) and electrostatic energy. In water at 293 K, lB = 0.71 nm. In view of the fact that the major ionic species of the nanofiber aqueous dispersion are Naþ and COO(dissociated carboxylate groups), and neglecting the effect of counterion condensation, κ-1 was estimated from eq 3 to be 18 nm, which was somewhat larger than the diameter of the cellulose nanofiber itself. Nevertheless, the effect of Debye length on the length estimate was considered to be relatively small because of the large influence of L on τ in eq 2. For example, even when d is 40 nm (the sum of the nanofiber diameter and twice the Debye length), eq 2 gives a value of L = 1.9 μm at τ = 7.32 10-2 s. The problem of the effect of electrostatic interaction on the precise estimation of average length may be avoided by the addition of a small amount of electrolyte, such as a buffer salt, which should not affect the dispersion state of the nanofibers. The effectiveness of viscoelasticity measurements in the length evaluation of TEMPO-oxidized cellulose nanofibers will be further investigated using the nanofibers prepared under different conditions.3,4
’ ASSOCIATED CONTENT
bS
Supporting Information. Transmission electron micrographs (TEMs) show that the TEMPO-oxidized cellulose nanofibers are approximated by linear rods on the order of 1 μm in length. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Phone: þ81 77 543 7599. Fax: þ81 77 543 7483. E-mail:
[email protected].
’ ACKNOWLEDGMENT This research was supported as a Nanotech Challenge Program from 2007 by the New Energy and Industrial Technology Development Organization of Japan (NEDO, Grant No. 1023001) and by Grants-in-Aid for Scientific Research (Grant Nos. 18380102, 07J04140, and 18-10902) from the Japan Society for the Promotion of Science (JSPS). ’ REFERENCES (1) Eichhorn, S. J.; Dufresne, A.; Aranguren, M.; Marcovich, N. E.; Capadona, J. R.; Rowan, S. J.; Weder, C.; Thielemans, W.; Roman, M.; Renneckar, S.; Gindl, W.; Veigel, S.; Keckes, J.; Yano, H.; Abe, K.; Nogi, M.; Nakagaito, A. N.; Mangalam, A.; Simonsen, J.; Benight, A. S.; Bismarck, A.; Berglund, L. A.; Peijs, T. J. Mater. Sci. 2010, 45, 1–33. (2) Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A. Biomacromolecules 2006, 7, 1687–1691. (3) Saito, T.; Kimura, S.; Nishiyama, Y.; Isogai, A. Biomacromolecules 2007, 8, 2485–2491. 550
dx.doi.org/10.1021/bm1013876 |Biomacromolecules 2011, 12, 548–550
