Professional Documents
Culture Documents
The shape and size distribution of crystalline nanoparticles resulting from the sulfuric acid hydrolysis of cellulose
from cotton, Avicel, and tunicate were investigated using transmission electron microscopy (TEM) and atomic
force microscopy (AFM) as well as small- and wide-angle X-ray scattering (SAXS and WAXS). Images of
negatively stained and cryo-TEM specimens showed that the majority of cellulose particles were flat objects
constituted by elementary crystallites whose lateral adhesion was resistant against hydrolysis and sonication
treatments. Moreover, tunicin whiskers were described as twisted ribbons with an estimated pitch of 2.4–3.2 µm.
Length and width distributions of all samples were generally well described by log-normal functions, with the
exception of tunicin, which had less lateral aggregation. AFM observation confirmed that the thickness of the
nanocrystals was almost constant for a given origin and corresponded to the crystallite size measured from peak
broadening in WAXS spectra. Experimental SAXS profiles were numerically simulated, combining the dimensions
and size distribution functions determined by the various techniques.
Figure 1. TEM micrographs of negatively stained cellulose whiskers obtained by sulfuric acid hydrolysis of cotton (a), Avicel (b), and tunicate
(c-e) cellulose. Insets: enlarged views of some characteristic particles. The arrows in (d) indicate zones where the whiskers are seen edge-on.
roretzi, a sea animal, was purified by alternating treatments with KOH Plus microscope operating in air and intermittent contact mode with a
and NaClO2.22 Micromash NC36 tip.
Preparation of Cellulose Whiskers Suspensions. Cotton linters Wide-Angle X-ray Scattering. Concentrated cellulose suspensions
were hydrolyzed according to the method described by Revol et al.4 (3–4 wt %) were allowed to dry onto flat Teflon surfaces. At this
They were treated with 65% sulfuric acid during 30 min at four different concentration, the suspension of Avicel particles was more viscous than
temperatures, namely 45, 54, 63, and 72 °C. In the following, the those containing cotton or tunicin microcrystals. The resulting films
resulting samples will be referred to as Cot45, Cot54, Cot63, and Cot72, were X-rayed with a Ni-filtered Cu KR radiation (λ ) 1.542 Å), using
respectively. The suspensions were washed by centrifugation, dialyzed a Philips PW3830 generator operating at 30 kV and 20 mA. The films
to neutrality against distilled water, and ultrasonicated for 4 min with were positioned either parallel or perpendicular to the X-ray beam.
a rod-type sonicator (Branson Sonifier B12). This treatment heated up Diffraction patterns were recorded on Fujifilm imaging plates, read with
the suspension by about 20 °C. After these treatments, the suspensions a Fujifilm BAS-1800II bioimaging analyzer. Diffraction profiles were
were filtered on 8 µm then 1 µm cellulose nitrate membranes (Sartorius) obtained by radially integrating the intensity over fan-shaped portions
and stocked with mixed bed resin (Sigma TMD-8) in order to eliminate of the spectra. The diffraction peaks were fitted with pseudo-voigt peak
residual electrolytes. Avicel was hydrolyzed at 72 °C following the functions, assuming a linear background. The dimension of the crystal
same procedure. The resulting sample will be referred to as Avi72. perpendicular to the diffracting planes with hkl Miller indices, Dhkl,
Tunicin was treated with 48% sulfuric acid during 13 h at 55 °C. Acid- was evaluated by using Scherrer’s expression:
free microcrystalline cellulose suspensions were obtained after repeated
centrifugations, dialysis to neutrality against distilled water, and 0.9 × λ
Dhkl ) (1)
ultrasonication for 1 min. This suspension, referred to as Tun55 in the β1/2 × cos θ
following, was stocked with mixed bed resin.
where θ is the diffraction angle, λ the X-ray wavelength and β1/2 the
Transmission Electron Microscopy. Drops of 0.001 wt % cellulose
peak width at half of maximum intensity.24
microcrystal suspensions were deposited on glow-discharged carbon-
Small-Angle X-ray Scattering. SAXS experiments were performed
coated TEM grids. The specimens were then negatively stained with
on the BM02 beamline of the European Synchrotron Radiation Facility
2% uranyl acetate, prior to complete drying, and observed using a
(Grenoble, France). Dilute cellulose suspensions (<0.5 wt %) were sealed
Philips CM200 electron microscope operating at 80 kV. Images were
in circular sample holders equipped with 25 µm thick mica windows. The
recorded on Kodak SO163 films. Selected negatives were digitized using
path length was 1 mm. Fourteen keV (λ ) 0.089 nm) X-rays were used.
a Kodak Megaplus CCD camera. For each preparation, the width and
Scattered intensities were recorded during 20 s exposures on a CCD
length of about 1000 particles (except tunicin, for which 200 particles
detector (Ropper Scientific) placed about 1.60 m behind the sample. The
were counted) were measured from the TEM images by using the
radial intensity distribution was obtained by integrating the whole diagram
AnalySIS software. Thin vitrified films of 0.1 wt % cellulose micro-
after a correction of camera distortion. Scattering data recorded from a
crystal suspensions were also prepared by quench-freezing in liquid
cell filled with water was subtracted as background.
ethane by using a procedure described elsewhere.23 They were observed
by cryo-TEM, at low temperature (-180 °C), in a Gatan 626 cryoholder.
Atomic Force Microscopy. After a short sonication to prevent the
Results
formation of aggregates, drops of dilute cellulose microcrystals
suspensions were deposited onto freshly cleaved mica. After 30 min, Particle Morphology. Figure 1 shows TEM micrographs of
the excess liquid was removed and the remaining film allowed to dry. elongated nanoparticles prepared by sulfuric acid hydrolysis of
AFM observations were carried out using a Molecular Imaging Pico the three types of cellulose. Particles from cotton (Figure 1a)
Crystalline Nanoparticles from Native Cellulose Biomacromolecules, Vol. 9, No. 1, 2008 59
and Avicel (Figure 1b) have a length between 100 and 300 nm,
whereas those from tunicin are several micrometers long and
have a whiskerlike morphology (Figure 1c,d). A significant
fraction of tunicin whiskers exhibit a kinked aspect (Figure 1e),
the defects being likely due to localized damage created during
the sonication treatment. For all samples, the negative staining
clearly reveals that most particles are composed of a few parallel
subunits. To check if this association was due to an artifactual
aggregation during staining and/or drying of the specimens on
the carbon film, cryo-TEM experiments were performed on
suspensions Cot45 and Tun55. The micrographs in Figure 2
show cellulose particles embedded in thin films of vitreous ice.
Both cotton (Figure 2a) and tunicin (Figure 2b) nanoparticles
are distributed in the film with a rather well-defined interparticle
distance, which indicates a significant electrostatic repulsion due Figure 3. Width (left column) and length (right column) distribution
to the negative charge of sulfate groups present on the surface histograms of Cot45, Cot54, Cot63, Cot72, Avi72, and Tun55
of the microcrystals after the sulfuric acid hydrolysis. Moreover, suspensions obtained by measuring the width and length on TEM
the images confirm that both cotton and tunicin nanocrystals micrographs of negatively stained preparations. The continuous
are composed of a varying number of parallel subunits. curves correspond to a fit of the data with a log-normal function.
Regular variations in width can often be seen along tunicin
whiskers. For negatively stained preparations, the accumulation larly to its long axis. For tunicin, the existence of a geometrical
of stain is more important around the narrower parts of the twist observed on TEM images had to be taken into account.
whiskers (Figure 2e), which means that these regions are also Assuming a rectangular cross-section and a twisted ribbon
thicker. The alternation of narrow and wide parts along the geometry, the width was defined as the largest dimension
whiskers suggests that they have a ribbonlike shape and that a measured perpendicularly to the whisker axis, while the thick-
homogeneous twist occurs. The same effect was observed in ness e was defined as the smallest dimension measured along
cryo-TEM images of vitrified specimens, proving that the twist the same nanocrystal. The corresponding size distribution
existed in solution and was not a drying/staining artifact (Figure histograms are shown in Figure 3. All distributions are asym-
2b). Using the longer whiskers, we measured the distance metrical, extending toward the longer and wider particles. Most
between two narrow or two wide regions and deduced a half- distributions were fitted using a log-normal function D whose
helical pitch of 1.2–1.6 µm. This distance is larger than the expression is:
( )
average length of the whiskers, which explains why, in TEM
- (ln(x) - ln(m))2
images, only one narrow region is seen along most whiskers. D ) A exp (2)
Length and Width Distributions. The length L of about s2
1000 particles from the three sources of cellulose was measured where m is the mode and s2 is the variance. Log-normal
directly from the TEM images of negatively stained preparations. functions are classically used to describe the size-distribution
For cotton and Avicel particles, the width was defined as the of objects obtained by fragmentation.25–27 The parameters m
largest dimension measured along the nanoparticle, perpendicu- and s2 calculated for the distributions of all suspensions are
60 Biomacromolecules, Vol. 9, No. 1, 2008 Elazzouzi-Hafraoui et al.
Figure 9. Observed (red +) and simulated (blue –) small-angle X-ray scattering profiles (cross-section plots ln(QI) ) f(Q2)) from dilute suspensions
of cotton, Avicel, and tunicate cellulose particles.
would be proportional to the contact area, thinner or shorter particles. The contribution of a rectangular particle can be
particles should be separated more easily. calculated by integrating its form factor F over all orientations:
Shape Distribution of the Particles. The width and length
measured from TEM micrographs of cotton nanocrystals
seemed to be somewhat correlated: in general, the longer F )
sin ( Lx2 )sin( ly2 )sin( ez2 ) (6)
particles were also the wider. To verify this assumption, we Lx ly ez
plotted two-dimensional histograms as a function of width 2 2 2
and length. The histograms were fitted with several empirical
functions, and the following expression was found to be the where L, l, and e are the length, width, and thickness of the
most satisfactory: particle, respectively. A series of SAXS profiles were calculated
for a range of width and length, assuming a constant thickness.
Ω(L,I) ) All contributions were finally sum-weighted using the two-
(( ))
(ln (l sin (θ) + L cos (θ)) - ln b1 2 dimensional probability function Ω(L,l) calculated from the
a exp - × TEM images, with thickness e as the adjustable parameter.
m1
(( ))
Figure 9 shows the experimental SAXS data collected from
l cos(θ) - L sin(θ) - b2 2
Cot45, Cot72, Avi72, and Tun55 suspensions, plotted as ln(IQ2)
exp - (4) ) f(Q2). As previously mentioned, all plots are convex and do
m2 ln(l sin(θ) + L cos(θ))
not show any linear region in the Q range, satisfying QRg , 1.
where a, θ, b1, b2, m1, m2 were the parameters to refine. SAXS curves were simulated for different particle thicknesses,
Moreover, we assumed that the particle length and width were using the two-dimensional function Ω(L,l) determined from
linearly related, according to the expression: particle width and length distributions, and superimposed to the
experimental profiles. The thickness was adjusted to obtain
l ) L tan θ + k (5) the better fit between simulated and experimental curves. The
where k is a constant. Ω(L,l) contains two terms: (i) a log-normal particle thickness was estimated to be about 6 nm for cotton,
function representing the size distribution of particles with an 3.8 nm for Avicel, and 10 nm for tunicin. These values are in
aspect ratio following expression 5; (ii) a Gaussian function good agreement with the AFM height measurements and with
that takes into account the deviation of the length and width of the crystallite size estimated from the peak broadening of WAXS
the particles from expression 5. A small m2 means that the diagrams.
relation between width and length is more strictly respected.
Using TEM images of negatively stained preparations, a
couple (L,l) was measured for each particle. As an example, Conclusion
the two-dimensional histogram corresponding to Cot45 particles For the three sources of cellulose considered in this study,
is shown in Figure 8 along with the Ω(L,l) function fitting the the crystalline nanoparticles resulting from a hydrolysis/soni-
data. The values of θ, b1, b2, m1, m2 obtained for Cot45, Cot72, cation treatment were rather flat and constituted of a few laterally
Avi72, and Tun55 samples are listed in Supporting Information bound elementary crystallites that were not separated by
Table S2. conventional acid hydrolysis and sonication process. In par-
Simulation of the SAXS Profiles. At low concentration, ticular, the particles from cotton and Avicel cellulose, which
assuming that there is no interparticle correlation, the intensity are classic samples from secondary cell wall, showed width
scattered by the suspension is a sum of contributions from all distributions following log-normal functions. We have shown
Crystalline Nanoparticles from Native Cellulose Biomacromolecules, Vol. 9, No. 1, 2008 65
that the description of cellulose nanoparticles as monodisperse (12) Araki, J.; Kuga, S. Langmuir 2001, 17, 4493–4496.
cylinders was not satisfactory to account for the X-ray scattering (13) Furuta, T.; Yamahara, E.; Konishi, T.; Ise, N. Macromolecules 1996,
29, 8994–8995.
properties of the suspensions. A more sophisticated approach,
(14) Orts, W. J.; Godbout, L.; Marchessault, R. H.; Revol, J.-F. Macro-
taking into account the data collected with a large panel of molecules 1998, 31, 5717–5725.
imaging and scattering techniques, was developed, allowing the (15) Belamie, E.; Davidson, P.; Giraud-Guille, M. M. J. Phys. Chem. B
statistical description of the assembly of cellulose crystalline 2004, 108, 14991–15000.
nanoparticles. However, the relation between the biosynthetic (16) Heux, L.; Chauve, G.; Bonini, C. Langmuir 2000, 16, 8210–8212.
(17) Araki, J.; Wada, M.; Kuga, S. Langmuir 2001, 17, 21–27.
pathways and the morphology remains an open question. Besides
(18) Favier, V.; Chanzy, H.; Cavaillé, J. Y. Macromolecules 1995, 28,
the biological relevance, the precise knowledge of the morphol- 6365–6367.
ogy and size distribution of the particles is crucial to analyze (19) Nishiyama, Y.; Kuga, S.; Wada, M.; Okano, T. Macromolecules 1997,
quantitatively statistical phenomenon like phase separation in 30, 6395–6397.
particle suspensions or percolation of rodlike objects. Indeed, (20) De Souza Lima, M. M.; Wong, J. T.; Paillet, M.; Borsali, R.; Pecora,
the determination of the particle density strongly depends on R. Langmuir 2003, 19, 24–29.
(21) Elazzouzi, S. Doctoral Thesis, Université Joseph Fourier, Grenoble,
the precision of the geometrical description of the objects. 2006.
Together with polydispersity, density is a key parameter that (22) Sugiyama, J.; Persson, J.; Chanzy, H. Macromolecules 1991, 24, 2461–
influences the phase separation and the formation of chiral 2466.
nematic organizations in cellulose nanocrystal suspensions. (23) Harris, J. R. NegatiVe Staining and Cryoelectron Microscopy: The
Thin Film Techniques; BIOS Scientific Publishers: Oxford, 1997.
Acknowledgment. We thank ESRF and the Innovation (24) Klug, H. P.; Alexander, L. E. In X-Ray Diffraction Procedures for
Center for Nanobiotechnologies (Nanobio, Grenoble) for grant- Polycrystalline and Amorphous Materials; Wiley Interscience: New
York, 1954, p 491.
ing access to SAXS and AFM facilities, respectively. We are (25) Epstein, B. Franklin Inst. 1947, 244, 471–477.
also grateful to H. Chanzy and B. Jean (CERMAV) for (26) Ishii, T.; Matsushita, M. J. Phys. Soc. Jpn. 1992, 61, 3474–3477.
stimulating discussions and I. Pignot-Paintrand for assistance (27) Cheng, Z.; Redner, S. J. Phys. A: Math. Gen. 1990, 23, 1233–1258.
with TEM observations. (28) Sugiyama, J.; Vuong, R.; Chanzy, H. Macromolecules 1991, 24, 4168–
4175.
Supporting Information Available. Mode (m) and squared (29) Araki, J.; Wada, M.; Kuga, S.; Okano, T. Colloids Surf. A 1998, 142,
75–82.
standard deviation (s2) of the log-normal functions describing (30) Terech, P.; Chazeau, L.; Cavaillé, J. Y. Macromolecules 1999, 32,
the width (l) and length (L) distributions of particles from acid- 1872–1875.
hydrolyzed cotton, Avicel, and tunicate cellulose. Parameters (31) van Daele, Y.; Revol, J.-F.; Gaill, F.; Goffinet, G. Biol. Cell 1992,
of the function defined in expression 4 and used to fit 76, 87–96.
experimental width and length distribution histograms shown (32) Helbert, W.; Nishiyama, Y.; Takeshi, O.; Sugiyama, J. J. Struct. Biol.
1998, 124, 42–50.
in Figure 3. Wide-angle X-ray diffraction diagrams recorded (33) Ruben, G. C.; Bokelman, G. H.; Krakow, W. In Plant Cell Wall
on films of cotton, tunicin, and Avicel cellulose particles. Cross- Polymers; Lewis, N. G., Paice, M. G., Eds.; ACS Symposium Series,
sectional view of the structural model of cellulose I. This Vol. 399; American Chemical Society: Washington, DC, 1989; pp
material is available free of charge via the Internet at http:// 278–298.
pubs.acs.org. (34) Hanley, S. J.; Revol, J.-F.; Godbout, L.; Gray, D. G. Cellulose 1997,
4, 209–220.
(35) Kim, N.-H.; Herth, W.; Vuong, R.; Chanzy, H. J. Struct. Biol. 1996,
References and Notes 117, 195–203.
(1) Rånby, B. G. Discus. Faraday Soc. 1951, 11, 158–164. (36) Sugiyama, J.; Chanzy, H.; Revol, J. F. Planta 1994, 193, 260–265.
(2) Marchessault, R. H.; Morehead, F. F.; Koch, M. J. J. Colloid Sci. (37) French, A. D.; Johnson, G. P. Cellulose 2004, 11, 449–462.
1961, 16, 327–344. (38) Brown, R. M., Jr.; Montezinos, D. Proc. Natl. Acad. Sci. U.S.A. 1976,
(3) Mukherjee, S. M.; Woods, H. J. Biochim. Biophys. Acta 1953, 10, 73, 143–147.
499–511. (39) Ito, T.; Brown, R. M., Jr. Protoplasma 1988, 144, 160–169.
(4) Revol, J.-F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray, (40) Brown, R. M., Jr.; Willison, J. H. M.; Richardson, C. L. Proc. Natl.
D. G. Int. J. Biol. Macromol. 1992, 14, 170–172. Acad. Sci. U.S.A 1976, 73, 4565–4569.
(5) Revol, J.-F.; Marchessault, R. H. Int. J. Biol. Macromol 1993, 15, (41) Read, S. M.; Bacic, T. Science 2002, 295, 59–60.
329–335. (42) Nishiyama, N.; Kim, U.-J.; Kim, D.-Y.; Katsumata, K. S.; May, R. P.;
(6) Abeysekera, R. M.; Willison, J. H. M. Biol. Cell 1990, 68, 251–257. Langan, P. Biomacromolecules 2003, 4, 1013–1017.
(7) Roland, J. C.; Reis, D.; Vian, B.; Satiat-Jeunemaitre, B.; Mosiniak, (43) Battista, O. A. Ind. Eng. Chem. 1950, 42, 502–507.
M. Protoplasma 1987, 140, 75–91. (44) Kuga, S.; Mutoh, N.; Isogai, A.; Usuda, M.; Brown, R. M., Jr. In
(8) Dong, X. M.; Revol, J. -F.; Gray, D. G. Cellulose 1998, 5, 19–32. Cellulose, Structure and Functional Aspects; Kennedy, J. F., Philips,
(9) Li, J.; Revol, J.-F.; Marchessault, R. H. J. Colloid Interface Sci. 1997, J. O., Williams, P. A., Eds.; Ellis Horwood: Chichester, UK, 1989;
192, 447–457. pp 75–80.
(10) Dong, X. M.; Kimura, T.; Revol, J.-F.; Gray, D. G. Langmuir 1996, (45) Saito, T.; Nishiyama, Y.; Putaux, J.-L.; Vignon, M.; Isogai, A.
12, 2076–2082. Biomacromolecules 2006, 7, 1687–1691.
(11) Li, J.; Revol, J.-F.; Naranjo, E.; Marchessault, R. H. Int. J. Biol.
Macromol. 1996, 18, 177–187. BM700769P