Cellulose Nanocrystals Grafted with Polystyrene Chains through

Apr 6, 2009 - Langmuir , 2009, 25 (14), pp 8280–8286. DOI: 10.1021/la900452a. Publication Date (Web): April 6, ... Cite this:Langmuir 25, 14, 8280-8...
0 downloads 0 Views 4MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Cellulose Nanocrystals Grafted with Polystyrene Chains through Surface-Initiated Atom Transfer Radical Polymerization (SI-ATRP) Gaelle Morandi, Lindy Heath, and Wim Thielemans* School of Chemistry and Process and Environmental Research Division;Faculty of Engineering, The University of Nottingham, Nottingham NG7 2RD, U.K. Received February 9, 2009. Revised Manuscript Received March 13, 2009 This paper reports the synthesis of cellulose nanocrystals grafted by polystyrene chains via surface-initiated ATRP. Naturally occurring cellulose was first hydrolyzed to obtain cellulose nanocrystals. Their surface was then chemically modified using 2-bromoisobutyryl bromide to introduce initiating sites for ATRP. A varying extent of surface modification was achieved by changing reaction conditions. Further initiation of styrene polymerization from these modified nanocrystals with a CuBr/PMDETA (N,N,N0 ,N0 ,N00 -pentamethyldiethylenetriamine) catalytic system and in the presence of a sacrificial initiator produced polysaccharide nanocrystals grafted by polystyrene chains. A range of nanocrystals-g-polystyrene with different graft lengths (theoretical polymerization degree = 27-171) was synthesized through this method and characterized by elemental analysis, XPS, FT-IR, TEM, and contact angle measurements. We are thus able to produce cellulose nanoparticles with varying grafting densities (by altering extent of initiator surface modification) and varying polymer brush length (through polymerization control). The nanocrystals-g-polystyrene (NC-g-PS) particles were tested for their capacity to absorb 1,2,4-trichlorobenzene from water. The results obtained show that they can absorb the equivalent of 50% of their weight in pollutant compared to 30 wt % adsorption for nonmodified nanocrystals, while also displaying faster absorption kinetics.

Introduction The development of materials derived from renewable resources is an important pillar in the drive to sustainability of our current society. Cellulose is the world’s most abundant natural, renewable, and biodegradable polymer and is a very promising raw material available at low cost for the preparation of various functional polymers. This naturally occurring polymer exists in nature in a semicrystalline state, and aqueous acids can be employed to selectively hydrolyze the less dense amorphous sections of the polymer, consequently releasing individual monocrystalline polysaccharide nanoparticles.1 The small size of those nanocrystals, reported to be around 200 nm in length and having cross-sectional dimensions of 8.8-18.2 nm for nanocrystals isolated from cotton,1,2 results in an inherently large specific surface area (∼300 m2/g). Because of the chemical nature of cellulose, the nanocrystal surface bears numerous hydroxyl functions which gives rise to highly active particles and can be used for grafting specific groups. Cellulose nanocrystals offer the benefits of being a nanofiller derived from renewable biomass in addition to possessing exceptional mechanical properties.3 Therefore, there exists a significant literature on the use of cellulose nanoparticles in nanocomposite materials.4 The high aspect ratio of cellulose nanoparticles and ease of dispersion in water has resulted in various studies on their rheological and assembly behavior.2 However, only limited work has been published on their surface modification, and reported work has mainly focused on the *Corresponding author: Tel +44 (0) 115 951 3507, Fax +44 (0) 115 951 3563, E-mail: [email protected]. (1) Samir, M. A. S. A.; Alloin, F.; Dufresne, A. Biomacromolecules 2005, 6(2), 612–626. (2) De Souza Lima, M. M.; Borsali, R. Macromol. Rapid Commun. 2004, 25(7), 771–787. (3) Sturcova, A.; Davies, G. R.; Eichhorn, S. J. Biomacromolecules 2005, 6(2), 1055–1061. (4) Dufresne, A. Can. J. Chem. 2008, 86(6), 484–494.

8280 DOI: 10.1021/la900452a

attachment of organic compounds via reaction of the surface hydroxyl groups using isocyanates,5,6 isothiocyanates,7 anhydrides,5,6 or chlorosilanes.8 The effective surface modification of polysaccharide nanocrystals with larger chains (stearic acids, polytetrahydrofuran, polycaprolactone, poly(ethylene glycol), etc.) via a grafting onto approach has also been reported.9,10 Though the expected nanoparticles were obtained through this method, a limitation in graft density due to steric hindrance and the potential of blocking of reactive sites by grafted polymer chains was observed. The surface modification of cellulose nanocrystals via a grafting from approach seems an interesting alternative as this method should give access to a higher graft density and better control of the overall structure. A recent study investigated ring-opening polymerization from the surface of cellulose nanoparticles using the surface hydroxyl groups as initiating sites.11 In this work, atom transfer radical polymerization (ATRP), which has proved to be a versatile technique to synthesize welldefined polymers with complex architecture,12,13 was chosen to synthesize the grafts as it will provide a good control of the graft length and composition. Furthermore, ATRP has already been (5) Angellier, H.; Molina-Boisseau, S.; Belgacem, M. N.; Dufresne, A. Langmuir 2005, 21(6), 2425–2433. (6) Nair, K. G.; Dufresne, A.; Gandini, A.; Belgacem, M. N. Biomacromolecules 2003, 4(6), 1835–1842. (7) Dong, S.; Roman, M. J. Am. Chem. Soc. 2007, 129(45), 13810–13811. (8) Gousse, C.; Chanzy, H.; Excoffier, G.; Soubeyrand, L.; Fleury, E. Polymer 2002, 43(9), 2645–2651. (9) Thielemans, W.; Belgacem, M. N.; Dufresne, A. Langmuir 2006, 22(10), 4804–4810. (10) Labet, M.; Thielemans, W.; Dufresne, A. Biomacromolecules 2007, 8(9), 2916–27. (11) Habibi, Y.; Goffin, A.-L.; Schiltz, N.; Dusquesne, E.; Dubois, P.; Dufresne, A. J. Mater. Chem. 2008, 18, 5002–5010. (12) Kamigaito, M.; Ando, T.; Sawamoto, M. Chem. Rev. (Washington, DC, U. S.) 2001, 101(12), 3689–3745. (13) Matyjaszewski, K.; Xia, J. Chem. Rev. (Washington, DC, U.S.) 2001, 101 (9), 2921–2990.

Published on Web 04/06/2009

Langmuir 2009, 25(14), 8280–8286

Morandi et al.

Article

Figure 1. (A) Representation of the acid hydrolysis of cellulose and (B) chemical structure of cellulose.

extensively employed to graft polymers from various surfaces such a glass,14 gold,15,16 magnetite17 and silica18-20 and has proved to be an efficient technique to graft polymers from macroscopic cellulose fibers21-26 and cellulose powder.27 Applying this approach to cellulose nanocrystals will give access to a new range of applications in the field of nanomaterials. To the best of our knowledge, only one example of surface-initiated ATRP from cellulose nanocrystals has been reported in the literature.28 This study describes a single reaction grafting polystyrene from cellulose nanoparticles and is focused on chiralnematic self-ordering of the grafted nanocrystals. Herein we report the synthesis of a range of polysaccharide nanocrystals grafted by polystyrene chains of different lengths using controlled SI-ATRP. We also report various conditions to control the level of initiator grafted on the nanocrystal surface, thereby allowing control over the grafting density of the polymer brush. The efficiency of the final nanoparticles to remove a typical persistent organic pollutant from water is also reported as a useful application. The latter is an important finding given the current emphasis on the limited character of water resources around the globe and the existing and increasing level of pollution in water streams.

Experimental Section Materials. Styrene (99%) was distilled over 2,5-di-

tert-butylhydroquinone under vacuum and stored at 4 °C after purification. N,N,N0 ,N0 ,N00 -Pentamethyldiethylenetriamine (PMDETA, 99+%) and copper(I) bromide (99.99%) were purchased from Aldrich. All other reagents and solvents were obtained from Aldrich and used without further purification. Instrumentation. Molecular weight and molecular weight distribution of polymers were obtained by gel permeation (14) Zhang, Z.; Chao, T.; Chen, S.; Jiang, S. Langmuir 2006, 22(24), 10072–10077. (15) Li, D.; Cui, Y.; Wang, K.; He, Q.; Yan, X.; Li, J. Adv. Funct. Mater. 2007, 17(16), 3134–3140. (16) Wei, Q.; Ji, J.; Shen, J. Macromol. Rapid Commun. 2008, 29(8), 645–650. (17) Marutani, E.; Yamamoto, S.; Ninjbadgar, T.; Tsujii, Y.; Fukuda, T.; Takano, M. Polymer 2004, 45(7), 2231–2235. (18) He, S.-x.; Gao, , B.-j. Gaofenzi Cailiao Kexue Yu Gongcheng 2007, 23 (3), 100-104, 108. (19) Ohno, K.; Morinaga, T.; Koh, K.; Tsujii, Y.; Fukuda, T. Macromolecules 2005, 38(6), 2137–2142. (20) Ramakrishnan, A.; Dhamodharan, R.; Ruehe, J. J. Polym. Sci., Part A: Polym. Chem. 2006, 44(5), 1758–1769. (21) Lindqvist, J.; Malmstroem, E. J. Appl. Polym. Sci. 2006, 100(5), 4155–4162. (22) Lindqvist, J.; Nystroem, D.; Oestmark, E.; Antoni, P.; Carlmark, A.; Johansson, M.; Hult, A.; Malmstroem, E. Biomacromolecules 2008, 9(8), 2139–2145. (23) Westlund, R.; Carlmark, A.; Hult, A.; Malmstroem, E.; Saez, I. M. Soft Matter 2007, 3(7), 866–871. (24) Carlmark, A.; Malmstrom Eva, E. Biomacromolecules 2003, 4(6), 1740–5. (25) Plackett, D.; Jankova, K.; Egsgaard, H.; Hvilsted, S. Biomacromolecules 2005, 6(5), 2474–2484. (26) Singh, N.; Chen, Z.; Tomer, N.; Wickramasinghe, S. R.; Soice, N.; Husson, S. M. J. Membr. Sci. 2008, 311(1-2), 225–234. (27) Coskun, M.; Temuez, M. M. Polym. Int. 2005, 54(2), 342–347. (28) Yi, J.; Xu, Q.; Zhang, X.; Zhang, H. Polymer 2008, 49(20), 4406–4412. (29) Habibi, Y.; Chanzy, H.; Vignon, M. R. Cellulose 2006, 13, 679–687.

Langmuir 2009, 25(14), 8280–8286

chromatography (PL-120, Polymer Laboratories) with an RI detector. The columns (30 cm PLgel Mixed-C, 2 in series) were eluted with THF and calibrated with polystyrene standards. All calibration and analyses were performed at 40 °C and a flow rate of 1 mL/min. 1H NMR spectra were recorded in CDCl3 using a Bruker DPX 300 MHz spectrometer. All spectra were referenced to CHCl3 at 7.26 ppm. FT-IR analyses were conducted on a TENSOR spectrometer from Bruker, and the data were analyzed using the OPUS software. XPS analyses were performed with a Kratos AXIS ULTRA spectrometer using a monochromated Al KR X-ray source (hν = 1486.6 eV) and a delay line detector (DLD) with a takeoff angle of 90° and an acceptance angle of 30°. The X-ray gun power was set to 150 W. The spectra were recorded using an aperture slot of 300  700 μm2 with a pass energy of 80 eV for survey scans and 10 or 20 eV for high-resolution scans. All spectra were recorded using Kratos VISION II software and processed using CASAXPS software. Elemental analyses were performed in the microanalysis service of the School of Chemistry, with a CE-440 elemental analyzer manufactured by Exeter Analytical. Contact angle measurements were performed on a CAM 200 instrument (KCV Instrument, Ltd.). An ∼2 μL volume droplet of Ultrapure water was dispensed on the sample, using a manual dispenser. Images of the droplet profile were recorded from which the CA was determined using the angle of intersection between a baseline and a circle using the Young-Laplace function fit to the drop profile. Transmission electron microscopy was performed on an FEI Tecnai 12 BioTwin electron microscope operating at an acceleration voltage of 80 kV. For unmodified cellulose nanoparticles, a carbon-coated Cu grid was treated under an 25% oxygen in argon plasma for 5 sec. A suspension of nanoparticles was deposited on the grid and left for 3 min and excess liquid removed. The grid was then stained using a 2% uranyl acetate in water solution for 5 min and dried before analysis. This procedure was also used for polystyrene-modified nanoparticles (using a DMF dispersion instead of water). In addition, polystyrene-modified nanoparticles were also deposited on a non-plasma-treated carbon-coated grid and stained using a 2.5% OsO4 in 2-methyl-2-propanol solution. Cotton Hydrolysis. Pure cotton (cotton wool) was dispersed in 64 wt % sulfuric acid in water. This suspension was held at 45 °C under mechanical stirring for 35 min to allow cotton hydrolysis. The suspension was subsequently diluted with an equal part of cold water and washed by successive centrifugation at 10 000 rpm and 10 °C (three times). Dialysis against distilled water was performed to remove free acid in the dispersion. This was verified by neutrality of the dialysis effluent. The nanocrystals were then recovered by freeze-drying after dispersion in water using a Branson Sonifier and filtration over a No. 2 fritted glass filter to remove residual aggregates. Esterification of the Nanocrystals. Cotton nanocrystals (1 g) were mixed with triethylamine (33 mL) and DMF (100 mL) under a nitrogen atmosphere. 2-Bromoisobutyryl bromide (26 mL) was then added dropwise under stirring. The mixture was stirred at 70 °C for 24 h and then filtered through an extraction thimble. Purification of the modified nanocrystals was realized by successive Soxhlet extractions with dichloromethane (24 h) and ethanol (24 h). Different amounts of DOI: 10.1021/la900452a

8281

Article

Morandi et al.

2-bromoisobutyryl bromide were used to study the possibility to vary the grafting density of the initiator. All other reagent quantities were kept constant. Surface Grafting Reaction: Typical ATRP. A Schlenk tube was loaded with copper(I) bromide (71 mg, 0.5 mmol) and the modified nanocrystals 2 (100 mg), capped with a rubber septum, and cycled two times between vacuum and argon to remove oxygen. In another Schlenk tube, anisole (11.5 mL), styrene (11.5 mL, 100 mmol), and ethyl 2-bromoisobutyrate (0.147 mL, 1 mmol) were introduced. The resulting solution was degassed under vacuum by three freeze-pump-thaw cycles and was added to the first Schlenk tube via a cannula. The Schlenk tube was placed in an oil bath thermostated at 100 °C. At t = 0, PMDETA (0.104 mL, 0.5 mmol) was added. The final conversion was determined by 1H NMR using anisole as an internal reference. The grafted nanoparticles were collected by filtration through a thimble and Soxhlet extracted with dichloromethane (24 h) and ethanol (24 h) to remove remaining PS homopolymer and the catalytic system compounds. The filtered solution was passed through an alumina column, and the PS homopolymer was isolated by precipitation into methanol followed by filtration.

Adsorption Experiments. 1,2,4-Trichlorobenzene absorbance was first recorded at 200 nm for a range of aqueous solution, allowing the determination of the extinction coefficient of 1,2,4-TCB according to the Beer-Lambert law: ε = 0.308 L mg-1 cm-1. This extinction coefficient was then used to determine the pollutant concentration in solution from the absorbance. Ten mg of nanocrystals were added to 1 L of a 1,2,4-TCB solution (C = 5 mg/L) under vigorous stirring. Aliquots were regularly removed from the solution, and their absorbance was measured by UV spectroscopy.

Results and Discussion Naturally occurring cellulose, composed of crystalline sections surrounded by amorphous parts, was first hydrolyzed using 64 wt % aqueous sulfuric acid at 35 °C to obtain cellulose nanocrystals (Figure 1A) according to a procedure already described.1 Because of the chemical composition of polysaccharides,

Scheme 1. Synthesis of Polysaccharide Nanocrystals-g-polystyrene by ATRP

Figure 2. FT-IR analysis of nonmodified nanocrystals 1 (gray line) and nanocrystals functionalized by ATRP initiator 2 (black line).

Figure 3. Final bromine content vs volume of 2-bromoisobutyryl bromide initially introduced. Table 1. Conditions and Results for the Nanocrystal Esterification with BiB elemental analysis

entry

VBra (mL)

time (h)

temperature (°C)

%C

% Br

% modified glucose unitsb

% modified surface glucose unitsc

% conversion surface hydroxyl groupsc

2a 2b 2c 2d 2e

26 24 70 40.8 5.0 10 53 35 26 24 30 41.4 2.1 4 21 14 26 16 70 39.4 3.4 7 37 25 13 24 70 40.7 3.0 6 32 21 52 24 70 40.8 9.5 20 105 70 a Volume of 2-bromoisobutyryl bromide for 1 g of nanocrystals. b Obtained through comparison between elemental analysis results and theoretical grafting density calculations (see Supporting Information for details). c Calculated based on crystal structure and dimension calculations which show that 19% of cellulose chains are found at the surface for cotton nanocrystals29 (see Supporting Information for details).

8282 DOI: 10.1021/la900452a

Langmuir 2009, 25(14), 8280–8286

Morandi et al.

Article

Figure 4. (A) Kinetic study of ATRP of styrene initiated by EBiB in presence of 2 and (B) dependence of number-average molecular weights and PDI with monomer conversion.

these nanocrystals bear numerous hydroxyl functions on their surface (Figure 1B). The nanocrystals 1 (Scheme 1) were then chemically modified to introduce initiating sites for ATRP. As the nanocrystals present a high specific surface area (∼300 m2/g)1, the amount of hydroxyl functions available for modification is high. Thus, only a partial esterification of the nanocrystal surface is expected. The esterification was conducted by the reaction of 2-bromoisobutyryl bromide (BiB) in DMF, an organic solvent able to disperse the nanocrystals well.1 An initial reaction, conducted with 26 mL of BiB for 1 g of nanocrystals, led to modified cellulose nanoparticles 2 (Scheme 1) with a bromine content of 5 wt % (as determined by elemental analysis). This bromine content corresponds to approximately one initiator site every ten glucose units as obtained through comparison between elemental analysis results and theoretical grafting density calculations. Relating this to the amount of available surface hydroxyl groups at the nanowhisker surface, this amounts to a conversion of 35% (see Supporting Information for details of calculations). The nanocrystal functionalization was confirmed by FT-IR analysis; the modified nanocrystals 2 spectra showing the appearance of an ester vibration band at 1724 cm-1 (Figure 2). No significant change in OH signal at 3350 cm-1 is noticed. This can be expected as only a limited amount of surface hydroxyl groups is modified, while the internal hydroxyl groups remain untouched. The influence of several factors (amount of BiB, temperature, and duration) on the final Br content was then investigated to allow us to tailor the grafting density of the polymer brushes through control of the initiator site density on the surface (Table 1). Grafting of the nanoparticles was confirmed by elemental analysis and FT-IR spectroscopy. As expected, a decrease of the BiB amount (2d), the reaction temperature (2b), or the reaction time (2c) resulted in a lower Br content in the final nanoparticles. In contrast, a higher amount of BiB leads to a final bromine content of 9.5 wt % (2e), which amounts to 70% conversion of the surface hydroxyl groups that are accessible for modification (assuming only one of the secondary hydroxyl groups is reactive per glucose unit since the second one is pointed slightly inward and expected to form hydrogen bonds with chains inside the crystal). Details of the calculation can be found in the Supporting Information. For identical time and temperature conditions (2a, 2d, and 2e), a linear dependence appears between the amount of BiB initially introduced in the media and the final nanocrystals bromine content (Figure 3). The range of bromine content following this linear dependence relates to an extent of surface hydroxyl group modification varying from 21% to 70%. Thus, the density of initiating sites on the modified nanocrystals surface can easily be controlled through the amount of reactant initially introduced, allowing for control of the final grafting density of nanoparticles after ATRP. Langmuir 2009, 25(14), 8280–8286

Table 2. Synthesis of NC-g-PS PS homopolymer entry

[S]0:[EBiB]0

a

conv (%)

-1 b

M n (g mol )

graft b

DPn,theoc

PDI

3a 3b 3c 3d

100:1 24 2800 1.10 27 100:1 43 4800 1.09 46 200:1 49 9600 1.10 92 400:1 40 17 800 1.09 171 a Determined by 1H NMR. b Determined by SEC in THF. c DPn,theo = M n(PS homopolymer)/Mw(styrene).

Table 3. Characterization of the Final Nanoparticles elemental analysis entry

%C

%H

XPS %C

PS wt % %O

IRa

EAb

2a 3a 3b 3c 3d

40.8 6.0 63.5 32.3 46.4 6.0 76.1 21.1 8 6 47.6 6.0 80.5 17.2 11 9 48.3 6.0 86.3 11.8 13 10 52.7 6.3 90.6 8.7 22 20 a Determined from the relative heights between the 700 and 670 cm-1 peaks after calibration. b Estimated from the EA results by comparison with the theoretical evolution of the carbon content vs the polystyrene percentage.

The modified nanocrystals 2 were then used as macroinitiators in ATRP (Scheme 1). Styrene was chosen as the monomer to obtain a final material with a high compatibility with aromatic compounds (toward persistent organic pollutant removal applications) or a good miscibility with a polystyrene matrix (toward nanocomposite formation). The polymerization was conducted with a CuBr/PMDETA (N,N,N0 ,N0 ,N00 -pentamethyldiethylenetriamine) catalytic system30 in anisole (which allows a good dispersion of the nanocrystals) at 100 °C. Ethyl 2-bromoisobutyrate (EBiB) was added to the reaction media as a sacrificial initiator. When working with surfaces containing a low amount of initiating sites, the use of a sacrificial initiator allows to tailor the grafts length by changing the initial ratio [monomer]/[EBiB] and the final conversion21 (the amount of initiating sites on the nanocrystals being negligible compared to the amount of sacrificial initiator). Furthermore, analysis of the homopolymer grown from EBiB gives an idea of the molecular weight and the polydispersity of the grafted polymer, even though the polymerization kinetics on the surface may differ. Sacrificial initiators have consequently been extensively used in surface-initiated ATRP to exhibit better control.18-20,24,31 The controlled character of the polymerization was investigated through a kinetic study and by following the evolution of the homopolymer number-average molecular weight vs (30) Xia, J.; Matyjaszewski, K. Macromolecules 1997, 30(25), 7697–7700. (31) Feng, W.; Zhu, S.; Ishihara, K.; Brash, J. L. Biointerphases 2006, 1(1), 50–60.

DOI: 10.1021/la900452a

8283

Article

Morandi et al.

Figure 5. (A) Full FT-IR analysis, (B) zoom-in FT-IR analysis in the 700 cm-1 area, and (C) XPS analysis for NC-g-PS 3 with various DPn,theo grafts .

conversion. The ln([M]0/[M]) vs time plot (Figure 4A) is linear, showing a first-order kinetic compatible with a constant concentration of active species for low conversion. The homopolymer number-average molecular weight vs conversion plot (Figure 4B) provides good agreement between experimental results and theoretical ones, and polydispersity values (PDI) are low (1.091.13). The polymerization of styrene in presence of modified nanocrystals 2 is thus well controlled up to moderate conversions. To obtain a range of NC-g-PS 3 with different graft lengths, several polymerization reactions were conducted with various [S]0:[EBiB]0 initial ratios and stopped at different conversions (Table 2). All reactions were stopped at relatively low conversion in order to keep a good control of the polymerization. The polystyrene (PS) homopolymers, issued from the sacrificial initiator, display number-average molecular weights between 2800 and 17 800 g mol-1 and low polydispersity indexes (PDI e 1.10). According to the homopolymer molecular weights and assuming a similar behavior on the surface, the theoretical graft length of the corresponding nanocrystals-g-polystyrene 3 (NC-g-PS) is ranging from 27 to 171 units. The final NC-g-PS 3 can easily be purified and separated from PS homopolymer via filtration through an extraction thimble followed by two successive Soxhlet extractions (dichloromethane followed by ethanol). The effective surface modification of NC-g-PS 3a-3d was investigated by elemental analysis (EA) and XPS (Table 3). As expected, the weight and number 8284 DOI: 10.1021/la900452a

percentages of carbon increase when the theoretical grafts length increases. The increase in carbon content is even more stressed by XPS which only investigates the sample surface (∼10 nm deep). The result for 3d indicates that the sample surface in almost entirely covered by a polystyrene layer. FT-IR analysis of the final grafted substrates (Figure 5A) shows the appearance of several signals related to the polystyrene structure at 3025 cm-1 (C-H stretching), 1494 cm-1 (CdC stretching), and 700 cm-1 (C-H bending). The superimposition of the signals at 700 cm-1 (Figure 5B) shows that the intensity of this peak increases with the graft length, as observed by others.21 The height of the 700 cm-1 peak relative to the 670 cm-1 peak (corresponding to the polysaccharide structure) was used to estimate the weight percentage of polystyrene in the final nanoparticles (using a calibration curve obtained from physical mixing of polystyrene and cellulose nanoparticles). The final PS content was also estimated through calculation of the theoretical carbon content in function of the PS weight percentage and by comparison with the EA results. FT-IR and EA give consistent results (Table 3); the final nanoparticles exhibit PS content ranging from approximately 5% to 20%. High-resolution XPS spectrograms of the C1s signal (Figure 5C) also show a distinct increase of the number of C-C bond on the nanoparticles surface with increasing DPn,theo. The characterization results obtained from EA, XPS, and FTIR are consistent with the theoretical evolution of the graft Langmuir 2009, 25(14), 8280–8286

Morandi et al.

length and show the possibility to change the grafts length by varying the [S]0:[EBiB]0 initial ratio and the final conversion. After the PS grafting, the morphology of the cellulose nanocrystals does not appear to have been affected as shown by TEM (Figure 6). This is in agreement with X-ray diffraction studies of the modified nanoparticles, which showed no change in crystallinity between unmodified nanoparticles and NC-g-PS (see Supporting Information). There is however a small change in crystalline morphology, the reason and extent of which are currently under investigation. NC-g-PS mainly appear as individualized nanoparticles, and only a small number of specimens were observed probably due to their bad adhesion to the grid. Staining the final nanoparticles with osmium tetroxide allows observing the PS shell surrounding the nanocrystals (Figure 6C). No clusters of OsO4-stained nanoparticles were seen on the grids investigated. Several attempts to hydrolyze the NC-g-PS were conducted in order to separately analyze the PS grafts and obtain a direct characterization of the polymer brush. The ester function located between the nanocrystals and the PS chains is the most obvious site to hydrolyze. However, the gem-dimethyl group on the R-position gives a high stability to that functionality. Therefore,

Article

the conditions required to hydrolyze this ester (KOH, THF, reflux) also resulted in nanocrystal hydrolysis, and no separated PS chains could be recovered from the final thick mixture. A recently reported procedure28 for a similar structure using HCl over a period of 3 weeks was also attempted, but the hydrolysis was unsuccessful. This might be due to a higher grafting density obtained for our nanoparticles, making the ester linkage less accessible. The surface modification of the NC-g-PS 3 was further investigated using contact angle (CA) measurements. Unmodified nanocrystals 1 and NC-g-PS 3c were compacted into a small pellet, and a drop of distilled water was deposited onto the surface. A clear and significant increase of the contact angle is observed between the unmodified nanocrystals 1 (Figure 7A, CA = 43°) and the grafted nanocrystals 3c (Figure 7B, CA = 94°). In addition, the contact angle decreases with time for the unmodified nanocrystals, whereas it remains constant for the grafted substrate, as expected for a hydrophobic surface (Figure 7C). While ungrafted nanocrystals are hydrophilic, they become hydrophobic after grafting, which further proves the coverage of the final nanoparticles with a PS layer.

Figure 6. Transmission electron micrographs of (A) nonmodified cellulose nanocrystals stained with uranyl acetate, (B) NC-g-PS 3c stained with uranyl acetate, and (C) NC-g-PS 3c stained with OsO4.

Figure 7. Contact angle pictures for (A) unmodified nanocrystals 1 and (B) NC-g-PS 3c and (C) time-dependent plot for nonmodified NC (0) and NC-g-PS ([). Langmuir 2009, 25(14), 8280–8286

DOI: 10.1021/la900452a

8285

Article

Morandi et al.

Figure 8. (A) Percentage of pollutant in solution vs time and (B) amount of pollutant adsorbed by the nanoparticles vs time for nonmodified NC (0) and NC-g-PS ([).

The presence of polymeric chains on which aromatic solutes can be adsorbed, or a polymer shell in which these solutes can be absorbed, combined with the large surface area of the final NC-g-PS should lead to interesting substrates for pollutant removal applications. 1,2,4-Trichlorobenzene (1,2,4-TCB), which is classified as a priority persistent organic pollutant, was used to investigate the nanoparticles capacity to absorb organic pollutant from water. NC-g-PS containing 8 wt % of PS was introduced in an aqueous solution of 1,2,4-TCB, and the evolution of the amount of pollutant in solution with time was followed by UV analysis. The same experiment was repeated with nonmodified nanocrystals as a reference. The percentage of pollutant in solution vs time plot (Figure 8A) shows a faster decrease of the amount of pollutant in the water when using NC-g-PS. After 80 min, only 9% of the initial pollutant remains in solution whereas more than 30% of the pollutant is still in solution with nonmodified nanocrystals in similar conditions. It is believed that the polystyrene shell around the nanoparticles act as reservoirs for absorption of the organic pollutant due to a preferential partitioning of the pollutant in polystyrene over water. This behavior mimics the driving force for bioaccumulation. The quantity of pollutant absorbed vs time plot (Figure 8B) shows 0.5 mg of 1,2,4-TCB can be absorbed by 1 mg of NC-g-PS nanoparticles, which represents 50% of their weight, when in the same time the adsorption of nonmodified NC is limited to 30 wt %. Thus, the PS chains greatly enhance the absorption/ adsorption capacity of the nanoparticles and lead to interesting candidates for organic pollutants removal applications. We are currently investigating the effect of different grafting densities and polymer chain lengths on the absorption behavior of these NC-g-PS nanoparticles as well as the effect of other grafted polymers (giving rise to a different partition coefficients) and other pollutants.

8286 DOI: 10.1021/la900452a

Conclusions The synthesis of cellulose nanocrystals bearing initiating sites for ATRP was successfully conducted. The possibility to control the final content of initiating sites was demonstrated, allowing us to tailor the grafting density of the polymer surface modification. This implies the possibility to tailor the final graft density. The grafting of polystyrene from these nanocrystals was then conducted under controlled conditions leading to a range of NC-g-PS with various graft lengths. XPS, IR, and EA results prove the possibility to control the grafts length by changing the monomer/ initiator initial ratio and the final conversion. A good coverage of the nanocrystals surface by a PS layer was further demonstrated by contact angle measurements, and transmission electron micrographs show that the shape of the initial nanocrystals remained largely unchanged. XRD data show no noticeable change in crystallinity but suggest a partial change in cellulose crystalline morphology, which is currently under investigation. The final NC-g-PS exhibited the capacity to absorb the equivalent of 50% of their weight of 1,2,4-TCB, proving their potential for pollutant removal applications. Further investigations of their ability to remove pollutants from water are currently in progress. Acknowledgment. The authors thank the EPSRC for funding through the DICE (Driving Innovation in Chemistry and Chemical Engineering) Science and Innovation Award (Grant EP/D501229/1). Supporting Information Available: X-ray diffraction results are presented and discussed as well as a detailed explanation of the theoretical grafting calculations for the interested reader. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2009, 25(14), 8280–8286