Hybrid rigid/soft and biologic/synthetic materials: Polymers grafted onto cellulose microcrystals

Article abstract of DOI:10.1021/bm101506j

Rigid nanoscale polymer rods were prepared by grafting preformed amine-terminated poly(styrene) and poly(tert-butyl acrylate) onto oxidized cellulose microcrystals. Low polydispersity polymers, grown using atom transfer radical polymerization, were characterized and purified prior to cellulose attachment. Oxidation of the cellulose microcrystal led to the formation of carboxylic acids on the surface of the microcrystals. Covalent attachment of the polymers onto the cellulose microcrystals was achieved via a carbodiimide-mediated amidation reaction. The length and diameter of the polymer-cellulose composites increased upon surface modification. Typically, polymer-cellulose composites are synthesized by a grafting-from method because it can be difficult to obtain sufficient graft density using a grafting-to preparation. However, the composites reported here comprised 60-64% grafted polymer by mass. This degree of grafting-to allowed the composite to form stable suspensions in organic solvents.

Full text of DOI:10.1021/bm101506j

ARTICLE
pubs.acs.org/Biomac
Hybrid Rigid/Soft and Biologic/Synthetic Materials: Polymers
Grafted onto Cellulose Microcrystals
Simon Harrisson,*,†,§ Glenna L. Drisko,†,^ Eva Malmstr€om,‡ Anders Hult,‡ and Karen L. Wooley*,†,
||
†Center for Materials Innovation and Department of Chemistry, Washington University in St. Louis, St. Louis, Missouri 63130,
United States
‡KTH Royal Institute of Technology, Department of Fibre and Polymer Technology, SE-100 44 Stockholm, Sweden
ABSTRACT: Rigid nanoscale polymer rods were prepared by grafting prefor-
med amine-terminated poly(styrene) and poly(tert-butyl acrylate) onto oxi-
dized cellulose microcrystals. Low polydispersity polymers, grown using atom
transfer radical polymerization, were characterized and purified prior to
cellulose attachment. Oxidation of the cellulose microcrystal led to the forma-
tion of carboxylic acids on the surface of the microcrystals. Covalent attachment
of the polymers onto the cellulose microcrystals was achieved via a carbodii-
mide-mediated amidation reaction. The length and diameter of the polymer-cellulose composites increased upon surface
modification. Typically, polymer-cellulose composites are synthesized by a grafting-from method because it can be difficult to
obtain sufficient graft density using a grafting-to preparation. However, the composites reported here comprised 60-64% grafted
polymer by mass. This degree of grafting-to allowed the composite to form stable suspensions in organic solvents.
’ INTRODUCTION
have been studied since 1959,7 with interest largely focused on their
ability to form colloidal liquid crystallinephases.8-12 More recently,
interest in nanocomposite materials has prompted research into the
use of cellulose microcrystals as inexpensive, biorenewable, rigid,
nanoscale fillers in composites with synthetic and natural polymers.1,5
Current limitations of unmodified cellulose microcrystals include
poor solubility in common solvents, a lack of thermoplasticity, and
poor dimensional stability. Their inability to form stable colloidal
suspensions in nonaqueous solvents forces the use of matrices
that are either water-soluble (e.g., starch,13-15 polyethylene
glycol,16-19 ethylenediamine/thiocyanate salt20) or water-dis-
persible (e.g., styrene-butyl acrylate emulsions,21 epoxy emulsions,22
poly(β-hydroxyoctanoate) latices23). The highly hydrophilic nature
of the cellulose microcrystal surfaces will result in poor adhesion
to hydrophobic matrices. By grafting polymers with differing
hydrophobicity and elasticity onto the surface of cellulose, their
dispersibility can be increased.
Many attempts have been made to modify the surface of
cellulose microcrystal suspensions to improve their dispersibility
in nonaqueous solvents.4,5 The most successful strategies for
doing so have been (a) grafting PEG chains onto the cellulose
surface,24,25 (b) silylation of the surface cellulose chains,26-28 (c)
polymerization from the cellulose surface,29-38 and (d) adsorp-
tion of a polyethylene oxide nonyl phenyl phosphate surfac-
tant.39,40 In all cases, the electrostatic stabilization of the original
microcrystals is converted to steric stabilization, by either grafted
or adsorbed polymer chains (a, c, and d) or by partially solvated
silylated cellulose chains (b). Silylation has limited utility as a
The preparation of objects with defined shape and submic-
rometer size domains is a major goal of nanotechnology. An
impressive degree of geometric control is attainable when work-
ing with crystalline materials: rods, cubes, spheres, wires, and
more complex 3D shapes have been reported.1 The inherent lack
of rigidity of soft materials places severe restrictions on the
variety of shapes that can be produced. As a result, the vast majo-
rity of discrete nanoscale structures formed from soft materials
exhibit spherical symmetry (e.g., micelles, vesicles). When cy-
lindrical (1D) or lamellar (2D) structures are obtained, they are
highly deformable, which may limit their use as building blocks in
more complex superstructures.2
An attractive alternative to the de novo production of 1D or
2D nanostructures is to make composite materials that incorpo-
rate a rigid scaffold that can impart its shape to the total assembly.
This approach allows the wide variety of functionalities asso-
ciated with soft materials to be married to the diversity of shapes
exhibited by hard materials.3 Cellulose is a ubiquitous crystalline
material that can be readily broken down into rod-shaped nano-
crystals.4 It is an excellent stiff substrate because it contains
surface hydroxyl groups that can be used as attachment points for
an assortment of organic functionalities.5
Cellulose is biosynthesized in the form of microfibrils with
nanoscale diameters and alternating crystalline and noncrystal-
line (amorphous) domains.5 Treatment with warm, concentrated
acid results in the preferential hydrolysis of the amorphous
domains, leading to the formation of rigid, rod-shaped crystalline
particles, known as cellulose microcrystals.4 The dimensions of
these particles depend on the source of the cellulose but are typi-
cally tens of nanometers in diameter and range from 100 nm to
several micrometers in length.6 Cellulose microcrystal suspensions
Received: December 13, 2010
Revised:
January 18, 2011
Published: March 07, 2011
r
2011 American Chemical Society
1214
dx.doi.org/10.1021/bm101506j Biomacromolecules 2011, 12, 1214–1223
|

Biomacromolecules
ARTICLE
means of dispersion: the cellulose can partially dissolve as the
silylation progresses, which may limit the degree of silylation that
can be achieved.27 The crystallinity and stiffness of the silylated
microcrystals are substantially decreased. By contrast, stabiliza-
tion by grafting polymers or by adsorbing surfactant proceeds
without detrimental effects on the morphology of the cellulose
microcrystals. The vast majority of the polymers have been
grafted-from, rather than grafted-to, the cellulose backbone.5
This approach allows high graft densities to be achieved, but the
polymerization is difficult to control and the resulting materials
are poorly characterized. The grafted-to method is a modular
approach in which the polymers are fully characterized and
purified before attachment to cellulose. Moreover, the grafted-
to approach allows the polymers to be prepared via a wider
variety of techniques, as long as there is a linkable functionality.
In this Article, we present a method for the production of
cellulose microcrystals grafted with hydrophobic polymers, such
as poly(styrene) and poly(t-butyl acrylate). The materials are
prepared via amidation reactions between the polymers’ terminal
amine functionalities and carboxylic acid groups on the surface of
the oxidized cellulose microcrystals. The polymer-grafted micro-
crystals are readily dispersed in nonpolar solvents such as acetone
and toluene and form an interesting class of rigid, rod-shaped,
polymeric nanoparticles, which may have applications in the
production of nanocomposite materials with hydrophobic ma-
trices such as natural rubber. The hydrophobic polymers are
produced via a quasi-living radical polymerization process, which
can be used to form well-defined polymers with a wide range of
functionalities and architectures. Polymer grafting could allow
this rigid biomaterial to be integrated into an assortment of
composites for use in packaging, coatings, adhesives, biomedical
polymers, tissue engineering, active surfaces, and engineering
applications.
Chromatography model 150-CV, equipped with a model 410 DRI
detector, a Precision Detectors PD2040 dual angle (15 and 90ꢀ) light
scattering detector, and a three-column series of Polymer Laboratories
PLgel 10 μm mixed B 300 ꢀ 7.5 mm columns. The system was
equilibrated at 35 ꢀC in anhydrous THF, which served as the polymer
solvent and eluent (flow rate 1.00 mL min-1). Data collection and
analysis were performed using Precision Detectors software. Interde-
tector delay volume and the light scattering detector calibration constant
were determined using a nearly monodisperse poly(styrene) calibrant
(Pressure Chemical, Mp = 90 000 g mol-1, Mw/Mn < 1.04). The
differential refractometer was calibrated with standard poly(styrene)
reference material (SRM 706 NIST) of a known specific refractive index
increment (dn/dc = 0.184 mL g-1). The dn/dc values of the analyzed
polymers were then determined from the differential refractometer
response.
Transmission Electron Microscopy (TEM). Carbon-coated
copper grids were prepared by oxygen plasma treatment to make the
surface hydrophilic. Samples were diluted 9:1 in water. A drop of diluted
sample was deposited onto a grid; after 1 min excess solvent was wicked
away. A drop of 1% PTA stain was then added and left for 1 min before
excess stain was wicked away. Particle diameters and standard deviations
were calculated from measurements of at least 100 particles from three
TEM micrographs from different regions.
Atomic Force Microscopy (AFM). Tapping-mode atomic force
microscopy measurements were conducted in air with a Nanoscope III
BioScope system (Digital Instruments) operated under ambient condi-
tions with standard silicon tips (type: OTESPA-70; L: 160 μm; normal
spring constant: 50 N m-1; resonance frequency 246-282 kHz). Sam-
ples were prepared by spin coating a dilute suspension of cellulose
microcrystals in water (pregraft) or acetone (postgraft) onto a freshly
cleaved mica surface.
Synthesis of N-(2-(2-Bromoisobutyryl)ethoxy)ethyl ph-
thalimide (1). 2-(2-Aminoethoxy)ethanol (10.5 g, 0.10 mol), phtha-
lic anhydride (14.8 g, 0.10 mol), and THF (50 mL) were mixed and
refluxed together for 1 h, by which time all phthalic anhydride had
dissolved. THF was removed in vacuo, and the product was refluxed
overnight in toluene using a Dean-Stark trap. The biphasic mixture
became homogeneous as the reaction proceeded. The solution was
cooled to 0 ꢀC, triethylamine (10.1 g, 0.10 mol) and 2-bromoisobutyryl
bromide (23.0 g, 0.10 mol) were added, and the solution was allowed to
warm to room temperature and stirred for 1 h. Triethylamine hydro-
bromide was precipitated and was removed by filtration. The solution
was washed with water (3 ꢀ 30 mL) and saturated NaCl solution (1 ꢀ
30 mL), then filtered through a short plug of basic Al2O3. Removal of
solvent in vacuo gave the desired product, which was further dried under
high vacuum to remove the last traces of toluene. Yield: 32.93 g (85.7
mmol, 85.7%) as a yellow oil.
’ EXPERIMENTAL SECTION
Materials. Fibrous cellulose powder (Whatman CF11 chromatog-
raphy grade) was obtained from Whatman International (Clifton, NJ).
Styrene (99%) and tert-butyl acrylate (99%) were purchased from Sigma
Aldrich (St. Louis, MO); inhibitor was removed by passing the mono-
mer through a column of basic alumina prior to use. Copper(I) bromide
(99.9995%), N,N,N0,N00,N00-pentamethyl diethylene triamine (PMDETA,
99%), 2-(2-aminoethoxy) ethanol (98%), phthalic anhydride (99%)
triethylamine (99%), 2-bromoisobutyryl bromide (99%), 2,2,6,6-tetra-
methyl-1-piperidinyloxy (TEMPO, 98%), 1% phosphotungstic acid
(PTA), methanol (MeOH), ethanol (EtOH), tetrahydrofuran (THF),
dimethyl formamide (DMF), 2-(2-aminoethoxy)ethanol, phthalic an-
hydride, N-hydroxy succinimide (NHS), 1,3-diisopropyl carbodiimide
(DIC), and toluene were obtained from Sigma Aldrich and used without
further purification. Sodium hypochlorite solution (available chlorine
10-13%) was obtained from Sigma Aldrich. Its concentration was
determined by the addition of excess potassium iodide in the presence of
sulfuric acid, followed by back-titration with sodium thiosulfate solution
using a starch indicator.24
1H NMR (CDCl3): δ 1.85 (s, 6H, -(CH3)2CBr), 3.72 (t, 2H, J = 4.8
Hz, -CH2CH2OC(O)-), 3.76 (t, 2H, J = 5.8 Hz, -OCH2CH2N-),
3.90 (t, 2H, J = 5.8 Hz, -NCH2-), 4.27 (t, 2H, J = 4.8 Hz, -
CH2OC(O)-), 7.70-7.86 (m, 4H, aromatic protons). 13C NMR
(CDCl3): δ 30.8 (CH3-), 37.5 (>NCH2-), 55.8 (BrC(CH3)2-),
65.2 (-C(O)OCH2-), 68.2 (>NCH2CH2O-), 68.4 (-C(O)OCH2-
CH2O-), 123.5, 132.3, 134.2 (aromatic carbons), 168.4 (phthalimide
CdO), 171.7 (ester CdO). IR (diffuse reflectance): 2945, 2871, 1775
(phthalimide CdO stretch), 1725 (ester CdO stretch), 1467, 1430,
1400, 1281, 1174, 1028, 725, 530 cm-1. Elemental analysis: C: 49.85%
H: 4.70% N: 3.44% Br: 21.03%. Theoretical C: 50.02% H: 4.72% N:
3.65% Br: 20.80%. MS (HRFAB, 3-NBA/Li matrix): m/z 390.0528
([M þ Li]þ, theoretical mass 390.0528).
Measurements. 1H NMR (300 MHz) and 13C NMR (75 MHz)
spectra were recorded on a Varian Mercury 300 MHz spectrometer.
Infrared (IR) spectra were obtained on a Perkin-Elmer BX FT-IR system
equipped with a diffuse reflectance accessory. Elemental analysis was
performed by M-H-W Laboratories (Phoenix, AZ). Thermogravimetric
analyses (TGAs) were performed at a heating rate of 10 ꢀC min-1 on a
Mettler Toledo TGA822e using Mettler Toledo Star SW 7.01 software.
Gel Permeation Chromatography (GPC). Polymer molecular
weight distributions were determined by GPC, conducted on a Waters
Preparation of r-Phthalimidyl poly(styrene). Kinetics: In-
itiator (1, 503 mg, 1.31 mmol), styrene (14.6 g, 140 mmol), CuBr (200
mg, 1.4 mmol), and PMDETA (290 μL, 1.4 mmol) were mixed in a
round-bottomed flask and freeze-thaw degassed (three cycles) before
1215
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Scheme 1. Preparation of Polymer-Grafted Cellulose Microcrystals
being backfilled with N2 and placed in an 80 ꢀC oil bath. Samples (∼0.5
mL) were taken every 30 min to track conversion (1H NMR) and
molecular weight (GPC) (Figure 4). At 210 min, the reaction was
stopped by pouring the reaction mixture into 200 mL of MeOH. The
polymer was dissolved in THF and filtered through a short plug of basic
Al2O3 to remove copper, then reprecipitated from MeOH and dried
overnight in a vacuum oven at 70 ꢀC. We recovered 3.07 g of white
polymer (53% based on 36% final conversion). Mn = 4840 g mol-1. PDI =
1.08 (GPC, PS stds). Theoretical Mn: 4390 g mol-1. Poly(styrene) for
grafting: initiator (1, 500 mg, 1.31 mmol), styrene (7.70 g, 68 mmol), CuBr
(200 mg, 1.4 mmol), and PMDETA (280 μL, 1.4 mmol) were degassed
and heated at 100 ꢀC for 2 h before precipitation from MeOH and drying
under vacuum. Conversion 97% (1H NMR). Yield: 5.32 g (71%). Mn =
mmol) was dissolved in 20 mL of THF. Hydrazine hydrate (1 mL) was
added, and the solution was refluxed for 2 h. A white precipitate of phthal-
hydrazide formed. The polymer was precipitated in 100 mL of MeOH and
dried overnight in a vacuum oven (70 ꢀC). Yield: 5.22 g (88.9%). Mn =
9870 g mol-1. PDI = 1.32.
Deprotection of r-Phthalimidyl Poly(t-butyl acrylate).
Phthalimidyl-functionalized poly(t-butyl acrylate) (5.00 g, Mn = 10
530 g mol-1, PDI = 1.10, 0.48 mmol) was dissolved in 20 mL of EtOH.
Hydrazine hydrate (1 mL) was added, and the solution was heated at
reflux for 2 h. A white precipitate of phthalhydrazide formed. The polymer
was precipitated in 50 mL of water and dried in a vacuum oven (70 ꢀC,
overnight). Yield: 4.48 g (89.6%). Mn = 10 800 g mol-1. PDI = 1.09.
Preparation of Cellulose Microcrystal Suspension. Cellu-
lose microcrystal suspensions were produced according to the method of
Dong et al.41 In a typical procedure, 10 g of powdered cellulose was
dispersed in 100 mL of 64% H2SO4 and stirred for 45 min at 45 ꢀC. The
resulting viscous, slightly yellow suspension was poured in 1 L of H2O to
quench the reaction. The cellulose was separated by centrifugation
(3000 rpm, 5 min), then resuspended in H2O. This treatment was
repeated until the supernatant remained cloudy after centrifugation. At
this point, the entire mixture was homogenized using an Osterizer food
processor set to “liquefy” for 30 min, then centrifuged once more. The
cloudy supernatant was poured into dialysis bags and dialyzed against DI
H2O for 3 days to remove the last traces of H2SO4. A white suspension
was obtained. Volume: 535 mL. Solids content: 1.08 w/w% (determined
by lyophilization). pH 4. Yield: 58%.
8950 g mol-1 (GPC, PS stds), PDI = 1.14. Theoretical Mn: 5540 g mol-1
.
Preparation of r-Phthalimidyl poly(t-butyl acrylate). In-
itiator (1, 503 mg, 1.31 mmol), tert-butyl acrylate (16.99 g, 133 mmol),
CuBr (200 mg, 1.4 mmol), andPMDETA(290μL, 1.4 mmol) were mixed,
degassed and placed in an oil bath at 60 ꢀC. After 3 h (81% conversion), the
reaction was stopped by precipitating the polymer in 200 mL of 75:25
MeOH/H2O. The polymer was redissolved in THF, filtered through a
short plug of basic Al2O3, and dried thoroughly in vacuo to remove THF
and residual monomer. We recovered 7.01 g of polymer (47.3% based on
81% conversion). Mn = 10 530 g mol-1. PDI = 1.10 (GPC, molecular
weight by multiangle light scattering). Theoretical Mn: 10960 g mol-1
.
Deprotection of r-Phthalimidyl Poly(styrene). Phthalimidyl-
functionalized poly(styrene) (5.87 g, Mn = 8950 g mol-1, PDI = 1.14 0.66
1216
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 1. TEM micrographs and length and diameter population distributions for cellulose microcrystals (A, unfilled bars) and oxidized cellulose
microcrystals (B, filled bars).
Oxidation of Cellulose Microcrystals. Cellulose microcrystal
suspensions were oxidized according to the method of Araki, Kuga, and
Wada.24 In a typical procedure, 1.73 g TEMPO and 4.32 g NaBr were
added to 400 mL of cellulose suspension (1.08% solids, 4.32 g cellulose).
The pH was adjusted to approximately 11 through the addition of a few
drops aqueous KOH and 22.6 mL of NaOCl solution (86 g L-1, 2.16 g
NaOCl). The yellow-orange mixture was stirred at room temperature
for 4 h. After this time, 30 g NaCl was added to flocculate the suspension,
and the cellulose was separated by centrifugation (3000 rpm, 5 min).
The cellulose was washed three times with 1 M NaCl (by suspension and
centrifugation), three times with 0.1 M HCl, and a final time with DI
H2O before being transferred to dialysis bags and dialyzed against DI
H2O for 3 days. Yield: 183 mL of white suspension, solids content 1.63%
(69% yield). Carboxylic acid content was determined by titration to be
0.76 mmol g-1 cellulose by following the procedure in ref 24. IR analysis
showed a strong absorption at 1732 cm-1, indicating the formation of
carboxylic acid groups.
Suspension of Oxidized Cellulose in DMF. In a typical
procedure, 50 mL of oxidized cellulose suspension (1.63%, 0.82 g
cellulose) was flocculated by the addition of 3 mL of saturated NaCl,
separated by centrifugation and rinsed (by redispersion, followed by
centrifugation at 3000 rpm for 5 min) three times with 50 mL of acetone,
then two times with 50 mL of DMF. The pellet was dispersed in 50 mL
of DMF and subjected to 5 min of sonication, then centrifuged (5 min,
3000 rpm). The turbid supernatant was decanted. The residue was
dispersed in a further 50 mL of DMF, given a further 5 min ultrasound
treatment and centrifuged (5 min, 3000 rpm). All of the cellulose
remained in suspension. The two portions were combined, yielding 100
mL of DMF suspension containing 0.82 g oxidized cellulose.
mmol) was dissolved in 10 mL of DMF and added to a 50 mL
suspension of oxidized cellulose in DMF (0.41 g cellulose, 0.31 mmol
COOH). NHS (100 mg, 0.87 mmol) and DIC (172 mg, 1.37 mmol)
were added to this mixture. The suspension was stirred at room
temperature for 16 h. The cellulose and poly(styrene) were precipitated
by the addition of 50 mL of H2O and separated by centrifugation. The
pellet was washed three times with acetone by resuspension, ultrasoni-
cation (5 min), and centrifugation to dissolve the excess poly(styrene).
The pellet was then dispersed in toluene (50 mL) using an iterative
process: 10 min of ultrasound treatment, centrifugation at 3000 rpm for
5 min, decantation of the liquid, and addition of 25-50 mL toluene.
Each time, a turbid supernatant was collected. After three iterations, the
small amount of solid material remaining was discarded. The final
volume was 121 mL. The toluene was removed through evaporation
from a 10 mL portion of the suspension, and the solid residue was heated
to 60 ꢀC for 2 h under vacuum to remove any remaining solvent. Final
mass: 63 mg (7.3 mg g-1, equivalent to 0.76 g total solids). Mass yield:
186% relative to oxidized cellulose.
Poly(t-butyl acrylate) Grafted (PtBA-Grafted) onto Oxi-
dized Cellulose. Amine-functionalized poly(t-butyl acrylate) (3.0 g,
0.28 mmol) was dissolved in 10 mL of DMF and added to 50 mL of a
suspension of oxidized cellulose in DMF (0.65 g cellulose, 0.52 mmol
COOH). DIC (320 μL, 2 mmol) and NHS (140 mg, 1.2 mmol) were
added to the resulting suspension, which was then stirred for 13 h at room
temperature. The mixture was centrifuged, and the pellet was twice suspen-
ded in acetone and separated by centrifugation. It was then resuspended in
acetone and subjected to a 10 min ultrasound treatment. This resulted in a
cloudy suspension that did not completely separate on centrifugation. The
cloudy supernatant was decanted, and the ultrasound and centrifugation
treatment was repeated until no further cloudy suspension was obtained. A
total of 100 mL (86 g) of suspension was obtained, with solids content of
Poly(styrene) Grafted (PS-Grafted) onto Oxidized Cellu-
lose. Amine-functionalized poly(styrene) (2.0 g, 9 870 g mol-1, 0.2
1217
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 2. (A,C) AFM micrographs and sections of cellulose microcrystals and (B,D) oxidized cellulose microcrystals. White lines in A and B indicate the
position of sections C and D. The vertical distance between the arrowheads in C and D is 18 and 19 nm, respectively.
Table 1. Dimensions of PS- and PtBA-Grafted, Oxidized, and Untreated Microcrystals from Analysis of TEM Data
PS-grafted
s.d. (nm)b
PtBA-grafted
XN (nm)a
XW (nm)c
PDId
XN (nm)a
s.d. (nm)b
XW (nm)c
PDId
cellulose microcrystals
oxidized cellulose
length
300
17
110
7.0
91
7.0
341
20
1.15
1.16
1.13
1.09
1.11
1.07
340
24
140
12
390
31
1.16
1.27
1.15
1.13
1.10
1.05
diameter
length
250
23
280
25
310
18
120
350
20
diameter
length
6.0
polymer-grafted cellulose
301
44
99
12
334
47
440
79
136
17
482
82
diameter
a Number average. b Standard deviation. c Weighted average. d Polydispersity index (XW/XN).
30 mg g-1 (total solids 2.58 g, 397% mass yield relative to cellulose). A
small amount of material remained precipitated. The suspension was stable
for more than 6 months of storage under ambient conditions.
’ RESULTS AND DISCUSSION
The synthetic strategy for preparation of the polymer-grafted
cellulose microcrystals is shown in Scheme 1. It involves (a) the
preparation of a cellulose microcrystal suspension, (b) the
oxidation of the cellulose to produce carboxylic acid groups at
the microcrystal surface, (c) the polymerization of amine-func-
tionalized polymers, and (d) the attachment of the polymers to
the cellulose surface via carbodiimide-mediated amidation. The
amine-functionalized polymers were synthesized via atom trans-
fer radical polymerization (ATRP) using initiator 1, which carries
latent amine functionality in the form of a phthalimidyl group.
This group is quantitatively converted to a primary amine by
reaction with hydrazine.
Figure 3. Typical potentiometric titration curve for a suspension of
oxidized cellulose microcrystals (gray line) and its first derivative (black
line). The weak (cellulose-COOH) and the strong (HCl) acid end
points are indicated with arrows. The first derivative curve is a three-
point smoothed curve of point estimates of first derivative.
birefringence when shaken between crossed polarizers, providing
further confirmation of the rod morphology.
Formation and Oxidation of Cellulose Microcrystal Sus-
pensions. Cellulose microcrystal suspensions were produced
according to the method of Dong et al.41 by hydrolysis of
Whatman CF11 powdered cellulose (derived from cotton) in
warm 64% H2SO4. During this process, the amorphous regions
of the cellulose microfibrils were converted to soluble monomers
and oligomers, leaving a suspension of crystalline material. The
suspended cellulose consisted of rigid rod-shaped particles, which
were observed by TEM (Figure 1A) and AFM (Figure 2A). By
removing the flexible amorphous matter, we produced individual
crystalline rods. Suspensions of the microcrystals exhibited flow
The suspensions were subsequently oxidized, to convert some
of the primary alcohols attached to the C6 carbon of the
anhydroglucose units into carboxylic acids.24 Oxidation pro-
ceeded without loss of the rod-shaped morphology or significant
changes in the nanoparticle dimensions, as measured by TEM
(Figure 1B, Table 1) and AFM (Figure 2B). The diffuse reflec-
tance IR spectrum of the lyophilized particles showed a strong
absorption at 1732 cm-1 due to the CdO stretch. The con-
centration of carboxylic acid groups could be determined by
1218
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 4. (A) Conversion versus time and (B) molecular weight/PDI versus conversion plots for the atom transfer radical polymerization of styrene
initiated by 1. Straight line in (A) shows linear fit to -ln(1-conversion) data, demonstrating first-order kinetics after a short induction period. Straight
line in (B) represents the theoretical molecular weight (= MW(1) þ MW(STY) ꢀ [M]/[I] ꢀ conversion). The line connecting the polydispersity data
is a guide for the eye only. Data points represented by closed circles relate to the y axis on the left, where open circles correspond to the y axis on the right.
Figure 5. 1H NMR spectra of R-phthalimido (bottom) and R-amino (top) (A,C) poly(styrene) and (B,D) poly(t-butyl acrylate).
potentiometric titration of the microcrystals (Figure 3) and was
found to be 0.76 mmol g-1 of oxidized cellulose.
using N-(2-bromoisobutyryl)ethyl phthalimide initiator. The
initiator used in this work, 1, was designed with a longer spacer
between the phthalimide and ester moieties to prevent the
irreversible isomerization of aminoethyl ester to hydroxyethyl
amide upon deprotection.43 The utility of 1 as an initiator for
ATRP was demonstrated here in the polymerization of styrene.
The rate of polymerization was first-order in monomer concen-
tration after a short induction period, showing that insignificant
Formation and Deprotection of Amine-Terminated Poly-
mers. Amine-terminated polymers were formed in a two-step
process from a phthalimidyl-terminated initiator, N-(2-(2-bro-
moisobutyryl)ethoxy)ethyl phthalimide (1). The use of the
phthalimide protecting group as a latent amine functionality in
ATRP has previously been suggested by Haddleton and coworkers42
1219
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 6. Molecular weight distributions of (A) poly(styrene) and (B) poly(t-butyl acrylate), before (- - -) and after (—) hydrazinolysis.
3
levels of termination took place. Low polydispersity polymer was
produced, with a molecular weight proportional to conversion
and close to the theoretical value (Figure 4).
1H NMR spectroscopy confirms the deprotection of the amine
after polymerization of styrene and t-butyl acrylate. Resonances
due to the phthalimide and the aliphatic protons of the initiator-
derived end group are clearly visible in the 1H NMR spectra of
the polymers (Figure 5A,B). The R-phthalimido polymers were
converted to their R-amino analogs by heating with excess
hydrazine for 2 h. The reaction was accompanied by almost
complete (93-100%) disappearance of the aromatic resonances
1
of the phthalimidyl moiety in the H NMR spectrum and the
appearance of a new resonance (at 2.9 ppm for poly(styrene) or
3.0 ppm for poly(t-butyl acrylate)), assigned to the methylene
protons R to the amine group (Figure 5C,D). The resonance
position of the methylene protons R to the ester was essentially
unaffected by deprotection, indicating that this functionality
survived deprotection without isomerization to the hydroxyalkyl
amide. No other changes to the polymer spectra were observed,
indicating that the hydrazinolysis was specific to the phthalimide
group. In particular, the relative integration ratios of the initiator
to chain resonances were preserved during the deprotection,
suggesting that hydrazinolysis of the ester linkage in the initiator
did not occur to a significant extent.
The molecular weight distribution of the poly(styrene), as
observed by size exclusion chromatography, exhibited a high-
molecular-weight shoulder after deprotection, indicative of some
chain-chain coupling during deprotection (Figure 6A). The
mechanism of the coupling reaction is not well-understood.
However, as shown in Figure 5, the amine-carrying end group
is not affected, suggesting that the Br terminus is involved. Pos-
sible mechanisms for coupling might involve radical formation by
loss of bromine, followed by radical-radical coupling or reaction
of the bromine terminus with hydrazine to form a substituted
hydrazine. The molecular weight distribution of the poly(t-butyl
acrylate) was essentially unaffected by the deprotection process
(Figure 6B).
Figure 7. IR spectra of (A) oxidized cellulose, (B) cellulose-PS graft, and
(C) cellulose-PtBA graft. Characteristic absorptions of the grafted polymers
are labeled with asterisks (*): (B) aromatic C-H str., 3050 cm-1; aromatic
C-C str., 1500 cm-1, out of plane C-H bend, 700 cm-1; (C) methyl C-
H str., 3000 cm-1, CdO str., 1730 cm-1, t-butyl umbrella bend, 800 cm-1
.
The absorptions marked with @ at 1650 cm-1 are due to adsorbed DMF.
polymers and hydrophilic cellulose to coexist in solution. Graft-
ing took place via a carbodiimide-mediated amidation reaction
between the carboxylic acid moieties of the cellulose and the
terminal amines of the polymer chains. Unreacted polymer
chains were removed by repeated sonication of the grafted cellu-
lose in acetone, followed by separation of the cellulose by
centrifugation. Under normal conditions of storage, the suspen-
sions were stable for more than 6 months; however, the cellulose
could be precipitated by centrifugation due to the difference in
the specific gravity between the acetone and the microcrystals.
Both the poly(styrene)-cellulose and poly(t-butyl acrylate)-
cellulose grafts formed stable suspensions in the nonpolar
solvents tested: chloroform, acetone, and toluene. Neither poly-
mer-cellulose composite could be dispersed in water. Diffuse reflec-
Grafting Polymers onto Cellulose. Poly(styrene) and poly-
(t-butyl acrylate) were grafted onto microcrystalline cellulose. DMF
was used as a compatibilizing solvent to allow the hydrophobic
1220
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 8. Thermogravimetric analysis traces of (A) cellulose-graft-poly(styrene) (- - -), oxidized cellulose (gray - - -, normalized to 36% of graft initial
mass), poly(styrene) (—, normalized to 56% of graft initial mass) and sum of oxidized cellulose and poly(styrene) traces (gray —); and (B) cellulose-
graft-poly(t-butyl acrylate) (- - -), oxidized cellulose (gray - - -), normalized to 34% of graft initial mass), poly(t-butyl acrylate) (—, normalized to 60% of
graft initial mass) and sum of oxidized cellulose and poly(t-butyl acrylate) traces (gray —).
Figure 9. (A,B) Atomic force micrographs and (C,D) sections of (A,C) PS-grafted and (B,D) PtBA-grafted cellulose microcrystals. White lines on A
and B indicate the positions of C and D. The vertical distances between the arrowheads in C and D are 27 and 40 nm, respectively.
tance IR analysis of the polymer-cellulose grafts showed absorbances
characteristic of both cellulose and the grafted polymer (Figure 7).
Thermogravimetric analysis of the polymer-cellulose grafts
(Figure 8) provided a means of estimating the extent to which
grafting had taken place. In the case of poly(styrene)-grafted
cellulose, the cellulose and poly(styrene) decompositions were
clearly separated. Superposition of a poly(styrene) decomposi-
tion curve onto the composite decomposition curve indicated
that the graft comprised 56% by mass poly(styrene) and 38% by
mass cellulose. This corresponds to an attachment of poly(styrene)
to ca. 20% of the cellulose carboxylic acid groups, calculated
from n(polymer)/n(COOH) = m(polymer)/m(cellulose) ꢀ 1/[n-
(COOH)/g cellulose] ꢀ 1/[Mn (polymer)], where n is the
quantity (mol) and m is the mass (g). A further 6% mass loss
occurred in the range 120-200 ꢀC due to desorption of DMF,
which binds strongly to cellulose.
Poly(t-butyl acrylate) shows a very sharp mass loss at 240 ꢀC
due to conversion of the t-butyl ester to a carboxylic acid, with
accompanying loss of isobutene: this transition results in the loss
of 43.7% of the mass of poly(t-butyl acrylate). A corresponding
mass loss was observed in the cellulose-t-butyl acrylate graft and
was used to estimate the t-butyl acrylate content of the graft at
60 wt %, with 6 wt % DMF and 34 wt % cellulose. This
corresponds to the attachment of poly(t-butyl acrylate) to ca.
22% of the cellulose carboxylic acid groups. For comparison,
Araki, Wada, and Kuga24 reported grafting densities of 20-30%
(depending on the method of analysis) in the production of
PEG-grafted cellulose by amidation in aqueous suspension.
For both cellulose-polymer grafts, the maximum cellulose
decomposition rate was observed at 350 ꢀC, compared with
325 ꢀC for oxidized cellulose or a blend of oxidized cellulose and
poly(styrene). This may be a result of the covalent grafting
reaction and the intimate contact between the cellulose surface
and the grafted polymer.
The cellulose-polymer grafts were also characterized by AFM
(Figure 9) and TEM (Figure 10). These techniques revealed
substantial increases in the diameters and lengths of the cellulose
microcrystals on grafting, with retention of their rod-shaped mor-
phologies. Some grafting techniques lead to decreases in the par-
ticle length, even at low grafting densities.44 Because the length/
width aspect is highly related to the performance of the particles
as reinforcers, this is a desirable grafting technique because it does
not compromise the morphology of the cellulose microcrystal.
In the case of poly(styrene)-grafted cellulose microcrystals,
attachment of poly(styrene) resulted in an increase in average
diameter from 23 ( 7 to 44 ( 12 nm (a difference of three standard
deviations relative to the substrate), as measured by TEM (Table 1).
The diameter distribution of the poly(styrene)-graftedmicrocrystals
was clearly shifted toward larger diameters relative to the substrate
(Figure 10). AFM measurements were in broad agreement with
this value, with typical particle heights of ∼30 nm (Figure 9C),
compared with ca. 20 nm for oxidized cellulose (Figure 2). For
1221
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
Figure 10. TEM images and histograms (filled bars) of (A) PS-grafted and (B) PtBA-grafted cellulose. The unfilled bars in histograms represent the
substrate oxidized cellulose size distributions.
comparison, the contour length of a polystyrene molecule with
MW 9600 is ca. 20 nm. The length distribution of the poly-
(styrene)-grafted microcrystals is similar to that of the substrate,
with an increase in average length from 250 ( 91 to 300 ( 99 nm
(a difference of 0.55 standard deviations). Whereas the absolute
increase in average length is larger than the increase in diameter,
it is associated with a high degree of uncertainty due to the
inherent variability in the length of the cellulose microcrystals.
Therefore, a large proportion of the observed difference in length
may be due to sampling error.
Poly(t-butyl acrylate) grafted microcrystals displayed a larger
increase in average diameter (18 ( 6 to 79 ( 17 nm) and average
length (310 ( 120 to 440 ( 136 nm) than poly(styrene) grafted
microcrystals, as observed by TEM. Because both poly(styrene)-
grafted and poly(t-butyl acrylate)-grafted microcrystals con-
tained similar amounts of polymer (∼60% in each case) of simi-
lar molecular weight (Mn ≈ 10 000 g mol-1) and contour length
(∼20 nm), the larger increase in size observed for poly(t-butyl
acrylate)-grafted particles cannot solely be attributed to differ-
ences in intramolecular interactions. The diameter distribution of
PtBA-grafted cellulose (Figure 10) is distinctly multimodal, with
peaks at 55-60, 70-75, 85-90, and 110-115 nm, separated
by approximately the diameter of one cellulose microcrystal
(18 nm). This unusual distribution suggests that the many of
the particles observed by TEM comprise bundles of several
microcrystals. Such aggregations of a few microcrystals would
also be longer than a single microcrystal, explaining the increase
in observed average length of the poly(t-butyl acrylate)-grafted
microcrystals. Whereas bundle-like aggregates of PS-grafted
cellulose are also observed, they appear to be much more
prevalent in PtBA-grafted cellulose, to the point where a typical
particle contains several cellulose microcrystals. Typical micro-
crystal heights of ∼40 nm were obtained by AFM (Figure 9D),
similar to the heights of PS-grafted microcrystals.
1222
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223

Biomacromolecules
ARTICLE
’ CONCLUSIONS
(9) Dong, X. M.; Kimura, T.; Revol, J.-F.; Gray, D. G. Langmuir
1996, 12, 2076.
We have demonstrated a method for the production of
polymer-grafted cellulose microcrystals, making use of amine-
terminated polymers produced via ATRP. The grafting-to ap-
proach used here allows for well-defined polymers of a variety of
molecular weights and functionalities to be covalently attached to
the cellulose surface, without degradation of the support. The
success of grafting the two polymer examples on to cellulose
microcrystals suggests that this approach could be extended to
many other amine-functionalized polymers. The resulting rigid,
rod-shaped nanoparticles are readily dispersible in a range of organic
solvents, including acetone, chloroform, and toluene. Therefore, the
grafting density of 60-64% (polymer/composite) is sufficient to
overcome the dispersibility challenges of unmodified cellulose.
ATRP can be used to produce a wide range of polymers with
different functionalities and architectures, which should allow
substantial modification of the surface properties of the cellulose-
based nanoparticles while conserving the underlying rod-shaped
morphology. The particles that result may be useful as the
reinforcing phase in nanocomposites and would be expected to
exhibit greater compatibility with hydrophobic matrices than do
unmodified cellulose microcrystals. The particles are also of
interest as precursors to multifunctional 1D nanoparticles.
(10) Dong, X. M.; Gray, D. G. Langmuir 1997, 13, 2404.
(11) Edgar, C. D.; Gray, D. G. Macromolecules 2002, 35, 7400.
(12) De Souza Lima, M. M.; Wong, J. T.; Paillet, M.; Borsali, R.;
Pecora, R. Langmuir 2003, 19, 24.
(13) Anglꢁes, M. N.; Dufresne, A. Macromolecules 2000, 33, 8344.
(14) Anglꢁes, M. N.; Dufresne, A. Macromolecules 2001, 34, 2921.
(15) Mathew, A. P.; Dufresne, A. Biomacromolecules 2002, 3, 609.
(16) Samir, M. A. S. A.; Alloin, F.; Sanchez, J.-Y.; El Kissi, N.;
Dufresne, A. Macromolecules 2004, 37, 1386.
(17) Samir, M. A. S. A.; Alloin, F.; Gorecki, W.; Sanchez, J.-Y.;
Dufresne, A. J. Phys. Chem. B 2004, 108, 10845.
(18) Samir, M. A. S. A.; Alloin, F.; Sanchez, J.-Y.; Dufresne, A.
Macromolecules 2004, 37, 4839.
(19) Schroers, M.; Kokil, A.; Weder, C. J. Appl. Polym. Sci. 2004, 93,
2883.
(20) Hattori, K.; Abe, E.; Yoshida, T.; Cuculo, J. A. Polym. J. 2004,
36, 123.
(21) Favier, V.; Chanzy, H.; Cavaillꢀe, J. Y. Macromolecules 1995, 28,
6365.
(22) Ruiz, M. M.; Cavaillꢀe, J. Y.; Dufresne, A.; Graillat, C.; Gꢀerard, J.-
F. Macromol. Symp. 2001, 169, 211.
(23) Dubief, D.; Samain, E.; Dufresne, A. Macromolecules 1999, 32,
5765.
(24) Araki, J.; Wada, M.; Kuga, S. Langmuir 2001, 17, 21.
(25) Thielemans, W.; Belgacem, M. N.; Dufresne, A. Langmuir
2006, 22, 4804.
’ AUTHOR INFORMATION
(26) Hattori, K.; Hiwatari, M.; Iiyama, C.; Yoshimi, Y.; Kohori, F.;
Sakai, K.; Piletsky, S. A. J. Membr. Sci. 2004, 233, 169.
(27) Goussꢀe, C.; Chanzy, H.; Excoffier, G.; Soubeyrand, L.; Fleury,
E. Polymer 2002, 43, 2645.
Corresponding Author
*E-mail: (S.H.) simon.harrisson@u-psud.fr. (K.L.W.) E-mail:
wooley@chem.tamu.edu.
(28) Petzold, K.; Koschella, A.; Klemm, D.; Heublein, B. Cellulose
2003, 10, 251.
Present Addresses
§Facultꢀe de Pharmacie, Universitꢀe Paris-Sud XI, 5 rue Jean-
Baptiste Clꢀement, 92296 Ch^atenay-Malabry, France.
Department of Chemistry, Texas A&M University, P.O. Box
30012, College Station, TX 77842, United States.
^Laboratoire de Chimie de la Matiꢁere Condensꢀee de Paris,
Collꢁege de France, 11 place Marcelin Berthelot, 75005 Paris,
France.
(29) Liu, P.-S.; Chen, Q.; Wu, S.-S.; Shen, J.; Lin, S.-C. J. Membr. Sci.
2010, 350, 387.
€
(30) Nystr€om, D.; Lindqvist, J.; Ostmark, E.; Antoni, P.; Carlmark,
A.; Hult, A.; Malmstr€om, E. ACS Appl. Mater. Interfaces 2009, 1, 816.
€
(31) Hansson, S.; Ostmark, E.; Carlmark, A.; Malmstr€om, E. ACS
Appl. Mater. Interfaces 2009, 1, 2651.
(32) Roy, D.; Guthrie, J. T.; Perrier, S. Soft Matter 2008, 4, 145.
(33) Roy, D.; Knapp, J. S.; Guthrie, J. T.; Perrier, S. Biomacromole-
cules 2008, 9, 91.
€
(34) Ostmark, E.; Nystr€om, D.; Malmstr€om, E. Macromolecules
’ ACKNOWLEDGMENT
2008, 41, 4405.
The material presented in this publication is based on work
supported by the National Science Foundation under grant no.
0301833. We are grateful to Mr. G. Michael Veith for assistance
with transmission electron microscopy.
€
(35) Lindqvist, J.; Nystr€om, D.; Ostmark, E.; Antoni, P.; Carlmark,
A.; Johansson, M.; Hult, A.; Malmstr€om, E. Biomacromolecules 2008, 9,
2139.
(36) Yi, J.; Xu, Q.; Zhang, X.; Zhang, H. Polymer 2008, 49, 4406–
4412.
€
(37) Ostmark, E.; Harrisson, S.; Wooley, K. L.; Malmstr€om, E. E.
’ REFERENCES
Biomacromolecules 2007, 8, 1138.
(38) Margutti, S.; Vicini, S.; Proietti, N.; Capitani, D.; Conio, G.;
Pedemonte, E.; Segre, A. L. Polymer 2002, 43, 6183.
(39) Bonini, C.; Heux, L.; Cavaillꢀe, J.-Y.; Lindner, P.; Dewhurst, C.;
Terech, P. Langmuir 2002, 18, 3311.
(1) Lee, S.-M.; Cho, S.-N.; Cheon, J. Adv. Mater. 2003, 15, 441.
(2) Yan, Q.; Yuan, J.; Kang, Y.; Yin, Y. Polym. Chem. 2010, 1, 423.
(3) Sanchez, C.; Juliꢀan, B.; Belleville, P.; Popall, M. J. Mater. Chem.
2005, 15, 3559.
(4) 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.
(5) Roy, D.; Semsarilar, M.; Guthrie, J. T.; Perrier, S. Chem. Soc. Rev.
2009, 38, 2046.
(40) Heux, L.; Chauve, G.; Bonini, C. Langmuir 2000, 16, 8210.
(41) Dong, X. M.; Revol, J.-F.; Gray, D. G. Cellulose 1998, 5, 19.
(42) Lecolley, F.; Waterson, C.; Carmichael, A. J.; Mantovani, G.;
Harrisson, S.; Chappell, H.; Limer, A.; Williams, P.; Ohno, K.; Had-
dleton, D. M. J. Mater. Chem. 2003, 13, 2689.
(43) Phillips, A. P.; Baltzly, R. J. Am. Chem. Soc. 1947, 69, 200.
(44) Siqueira, G.; Bras, J.; Dufresne, A. Langmuir 2010, 26, 402.
(6) Lu, J.; Askeland, P.; Drzal, L. T. Polymer 2008, 49, 1285.
(7) Marchessault, R. H.; Morehead, F. F.; Walter, N. M. Nature
1959, 184, 632.
(8) Revol, J.-F.; Bradford, H.; Glasson, J.; Marchessault, R. H.; Gray,
D. G. Int. J. Biol. Macromol. 1992, 14, 170.
1223
dx.doi.org/10.1021/bm101506j |Biomacromolecules 2011, 12, 1214–1223