Cellulose Nanocrystals/Cellulose Core-in-Shell Nanocomposite Assemblies

pubs.acs.org/Langmuir © 2009 American Chemical Society

Cellulose Nanocrystals/Cellulose Core-in-Shell Nanocomposite Assemblies Washington Luiz Esteves Magalh~aes,†,‡ Xiaodong Cao,†,§ and Lucian A. Lucia*,† †

Department of Wood & Paper Science, North Carolina State University, Campus Box 8005, Raleigh, North Carolina 27695-8005, ‡Embrapa Florestas, Tecnologia da Madeira, CP319, Estrada da Ribeira km 111 Colombo, PR 834111-000, Brazil, and §Biomedical Engineering Institute, South China University of Technology, Guangzhou, Gaungdong, 510640, P. R. China Received May 29, 2009. Revised Manuscript Received July 18, 2009 We report herein for the first time how a co-electrospinning technique can be used to overcome the issue of orienting cellulose nanocrystals within a neat cellulose matrix. A home-built co-electrospinning apparatus was fabricated that was comprised of a high-voltage power supply, two concentric capillary needles, and one screw-type pump syringe. Eucalyptus-derived cellulose was dissolved in N-methylmorpholine oxide (NMMO) at 120 °C and diluted with dimethyl sulfoxide (DMSO) which was used in the external concentric capillary needle as the shell solution. A cellulose nanocrystal suspension obtained by the sulfuric acid hydrolysis of bleached sisal and cotton fibers was used as the core liquid in the internal concentric capillary needle. Three flow rate ratios between the shell and core, four flow rates for the shell dope solution, and four high voltages were tested. The resultant co-electrospun composite fibers were collected onto a grounded metal screen immersed in cold water. Micrometer and submicrometer cellulose fiber assemblies were obtained which were reinforced with cellulose nanocrystals and characterized by FESEM, FTIR, TGA, and XRD. Surprisingly, it was determined that the physical properties for the cellulose controls are superior to the composites; in addition, the crystallinity of the controls was slightly greater.

Introduction Cellulose as a renewable material is one of the most important polymeric biomaterials in the biosphere. Not only does it constitute the most abundant biomaterial, but it is a platform for the development of a suite of new chemicals, materials, and fuels. Its physical and chemical properties are only now beginning to be rationally exploited for the design of new engineered materials. Its ability to reinforce composites, provide a mechanical platform for tissue engineering, act as a hydrogel/structured assembly for drug delivery, and behave as a reaction medium are becoming more and more attractive to a number of disciplines.1-4 Notably, the mechanical properties of cellulose rank among the most superior for any natural or synthetic material. More specifically, the modulus of elasticity (MOE) of crystalline cellulosic biomaterials is extremely high and rates very favorably even when compared to any number of very strong inorganic materials. For example, crystalline cellulose type I has a calculated MOE that is as high as 138 GPa.5-7 Virgin 2.25Cr-1Mo steel, one of the strongest materials known, has an MOE that is only ∼50% greater than crystalline cellulose.8 Additionally, the more readily available crystalline cellulose types II, IIII, IIIII, and IV have been assigned MOEs of 88, 87, 58, and 75 GPa, respectively,7 although several authors have calculated somewhat different values for the MOE *To whom correspondence should be addressed: e-mail lucian.lucia@ncsu. edu; Tel þ1-919-515-7707; Fax þ1-919-515-6302.

(1) Jean, B.; Heux, L.; Dubreuil, F.; Chambat, G.; Cousin, F. Langmuir 2009, 25, 3920–3923. (2) Xie, Y.; Wang, M.; Yao, S. Langmuir 2009, DOI: 10.1021/la9014338. (3) Li, S.; Zhang, S.; Wang, X. Langmuir 2008, 24, 5585–5590. (4) Linder, A˚.; Bergman, R.; Bodin, A.; Gatenholm, P. Langmuir 2003, 19, 5072–5077. (5) Tashiro, K.; Kobayashi, M. Polym. Bull. 1985, 14, 213–218. (6) Sturcova, A.; Davies, G. R.; Eichhorn, S. J. Biomacromolecules 2005, 6, 1055–1061. (7) Nishino, T.; Takano, K.; Nakamae, K. J. Polym. Sci., Part B: Polym. Phys. 1995, 33, 1647–1651. (8) Latella, B. A.; Humphries, S. R. Scr. Mater. 2004, 51, 635–639.

13250 DOI: 10.1021/la901928j

of types I, II, IIII, and IVI: 90, 75, 115, and 54 GPa, respectively.9 The latter discrepancies are not surprising given the tremendous complexity associated with proper characterization and identity of the various cellulosic allomorphs.9 Nevertheless, the overall strength properties of this abundant biomaterial are an attractive consideration in the design of environmentally compatible end products. However, it is not yet possible to produce a macroscopic highly crystalline cellulose type I biomaterial; hence, a more immediately attractive and successful approach is to take advantage of its intrinsic nanoscopic high strength for the manufacture of nanocellulosic composites. It is known, for example, that a composite composed of a regenerated coniferous-derived pulp acting as the matrix and ramie cellulose fiber acting as the reinforcement can be produced using LiCl/DMAc as the solvent.10 The ramie fibers can align and become impregnated with the cellulose LiCl/DMAc solution under vacuum, which after exposure to ambient conditions can undergo gelation. After subsequent immersion in methanol to extract LiCl/DMAc, the composite can be dried to provide a useful biomaterial. Remarkably, the mechanical properties are comparable or even greater than those of conventional glass fiber-reinforced composites.11 Another approach is developing self-reinforced neat cellulose composites by means of partial dissolution of microcrystalline cellulose powder in lithium chloride/N,N-dimethylacetamide and subsequent film casting. The films are isotropic, transparent to visible light, and highly crystalline and contain different amounts of undissolved cellulose I crystallites in a matrix of regenerated cellulose. The results show that by varying the cellulose I and II ratio via controlling the time of dissolution the mechanical performance of the nanocomposites can be tuned. Depending (9) Ishikawa, A.; Okano, T.; Sugiyama, J. Polymer 1997, 38, p 463 468. (10) Qin, C.; Soykeabkaew, N.; Xiuyuan, N.; Peijs, T. Carbohydr. Polym. 2008, 71, 458–467. (11) Nishino, T.; Matsuda, I.; Hirao, K. Macromolecules 2004, 37, 7683–7687.

Published on Web 09/04/2009

Langmuir 2009, 25(22), 13250–13257

Magalh~ aes et al.

Article

Figure 1. Schematic drawing of the home-built device employed in all of the electrospinning experiments.

on the composition, a tensile strength of up to 430 MPa and MOE of up to 33 GPa can be observed after soaking under wet conditions and subsequent drying.12,13 The same approach can be used for filter paper as a starting material after partially dissolving it in LiCl/DMAc, followed by compression, methanol extraction, and drying. The mechanical properties (tensile strength) of the final transparent isotropic film can go all the way up to 211 MPa at 25 °C.14 Partially dissolvinggrade beech pulp disks in LiCl/DMAc have also been made which can provide tensile strengths of 154 MPa and MOEs of 12.2 GPa after regeneration, hot pressing at a temperature of 80 °C, and a final pressure of 200 MPa.15 However, most cellulosics suffer from the random interspersion of a number of amorphous or noncrystalline regions in their cellulosic nanofibrils. These weak regions contribute to an anisotropic distribution of mechanical load that leads to an artificially low strength measurement. A very promising approach to enhancing the overall mechanical properties of cellulose is to deconstruct it to its elemental nanocrystalline domains by dissolving the less structured amorphous regions. Typically, this can be done by a strong acid hydrolysis of cellulose that provides nanocrystals. A water suspension of cellulose nanocrystals can be prepared, for example, by hydrolyzing cotton cellulose in sulfuric acid. Microcrystallite suspensions made in this manner readily form an anisotropic phase above a critical concentration of ∼4.5% (w/v) cellulose.16 Moreover, alignment of the nanocrystals is possible by mechanical shearing action and also by magnetic17 or electric fields.18 Although cellulose nanocrystals can easily be obtained, the use of this highly crystalline material as a reinforcement is limited to polymers that are soluble in water or organic solvents into which the nanocrystals can be suspended. For

instance, cellulose nanocrystal reinforcement of regenerated cellulose;mainly type II or amorphous;is difficult because the solvent that is used to dissolve cellulose can easily dissolve the nanocrystals. Although many solvents can dissolve cellulose, only a few among them are used for electrospinning due to the dielectric requirements of the technique. Electrospinning is a technique that allows for the generation of nanofibers or nanoscopic bundles from a solution or suspension of a substrate. For example, a cellulose solution in an NMMO/water system can be used for electrospinning to make fibers with submicrometric diameter ranges with crystallite lattice of type II.19,20 The technique has been used to produce nanoscopic fibers of cellulose from NMMO solution to reinforce PBS (poly(butylene succinate)).21 An important criterion in the few research efforts involving cellulose electrospinning is the temperature. For example, if a cellulose/ NMMO/H2O system remains at room temperature, the solution will be solid, although recent work has shown that doped solutions can be successfully spun if the doped solution temperatures are maintained between 80 and 130 °C.19 In order to keep the temperatures of cellulose solutions in this range or higher, a heating oil circulator has been employed.21 Cellulose forms a solution in lithium chloride (LiCl)/N, N-dimethylacetamide (DMAc) that can also be used for electrospinning at room temperature; however, the cellulose fiber produced is mostly amorphous.20,22 In fact, an enzymatically treated cellulose solution in the LiCl/DMAc system can also be electrospun; yet, the bulk of the fibers formed collapse into droplets, a phenomenon that mainly occurs because the substrate is electrically conductive.23 In general, solutions of cellulose derivatives can be electrospun,19 and the cellulose regenerated forms a film or fibers.24,25 Frey and Joo used a novel

(12) Gindl, W.; Keckes, J. J. Polym. 2005, 46, 10221–10225. (13) Gindl, W.; Keckes, J. J. Appl. Polym. Sci. 2007, 103, 2703–2708. (14) Nishino, T.; Arimoto, N. Biomacromolecules 2007, 8, 2712–2716. (15) Gindl, W.; Schoberl, T.; Keckes. J. Appl. Phys. A: Mater. Sci. Process. 2006, 83, 19–22. (16) Dong, X. M.; Revol, J.-F.; Gray, D. G. Cellulose 1998, 5, 19–32. (17) Revol, J. F.; Godbout, L.; Dong, X. M.; Gray, D. G.; Chanzy, H.; Maret, G. Liq. Cryst. 1994, 16, 127–134. (18) Habibi, Y.; Heim, T.; Douillard, R. J. Polym. Sci., Part B: Polym. Phys. 2008, 46, 1430–1436.

(19) Kulpinski, P. J. Appl. Polym. Sci. 2005, 98, 1855–1859. (20) Kim, C.-W.; Kim, D. S.; Kang, S. Y.; Marquez, M.; Joo, Y. L. Polymer 2006, 47, 5097–5107. (21) Han, S. O.; Son, W. K.; Youk, J. H.; Park, W. H. J. Appl. Polym. Sci. 2008, 107, 1954–1959. (22) Kim, C.-W.; Frey, M. W.; Marquez, M.; Joo, Y. L. J. Polym. Sci., Part B: Polym. Phys. 2005, 43, 1673–1683. (23) Frenot, A.; Henriksson, M. W.; Walkenstrom, P. J. Appl. Polym. Sci. 2007, 103, 1473–1482. (24) Maa, Z.; Kotaki, M.; Ramakrishna, S. J. Membr. Sci. 2005, 265, 115–123. (25) Son, W. K.; Youk, J. H.; Park, W. H. Biomacromolecules 2004, 5, 197–201.

Langmuir 2009, 25(22), 13250–13257

DOI: 10.1021/la901928j

13251

Article

Magalh~ aes et al.

Figure 2. FESEM of electrospun cellulose samples using a 21 G needle, 20 kV, and a 10 cm distance from tip to water-immersed metal screen collector at (a) 0.267, (b) 0.937, and (c) 2.005 cm3/h flow rates. The inset magnification bars for the top three microphotographs represent 5, 20, and 50 μm, respectively. The bottom set of three microphotographs are magnifications of the top three by 10, 4, and 10, respectively.

Figure 3. FESEM of a co-electrospun film (under increasing magnification from 20, 5, and 1 μm inset bars from left to right) obtained from

the following conditions: 8 cm tip distance to bath collector, 10 kV applied to tip, shell flow rate=0.402 cm3/h, and a flow rate ratio of shell to core=0.9.

solvent (ethylenediamine and pure salts or mixtures of potassium thiocyanate and potassium iodide) and electrospinning to produce fibers.26 Finally, ionic liquids can also dissolve cellulose and be electrospun at room temperature.27 The main objective of the current work was to generate pure cellulose nanocomposites using regenerated cellulose (type II and amorphous) that is then reinforced with cellulose nanocrystals (rods). We therefore provide the first known report on the coelectrospinning of a cellulose-doped solution in NMMO/H2O that comprised the shell and a cellulose nanorod water suspension in the core that formed a fibrous nanocomposite material. This new material was characterized by FESEM, FTIR, XRD, and TGA.

Experimental Section Electrospinning of the Cellulose Solutions. Pure cellulose fibers may be produced from the electrospinning of a cellulosedoped solution, albeit in very low amounts, which were used (26) Frey, M. F.; Joo, Y. L. U. S. Patent 0,247,236, Nov 10, 2005. (27) Viswanathan, G.; Murugesan, S.; Pushparaj, V.; Nalamasu, O.; Ajayan, P. M.; Linhardt, R. J. Biomacromolecules 2006, 7, 415–418.

13252 DOI: 10.1021/la901928j

primarily as a baseline of comparison for electrospun cellulose composites. In all of the electrospinning experiments, a 21 G needle was used, the applied voltage was 20 kV, and the distance from tip to water-immersed screen collector was 10 cm. For the co-electrospinning experiments, a home-built apparatus was used (Figure 1).

Description and Operation of the Home-Built Electrospinning Apparatus. The electrospinning device (inset in Figure 1) was constructed of an aluminum hollowed cylinder having a 6 mm diameter and 35 mm height with a side tap (set screw) near the top end. The hollowed-out internal diameter of this screwed tap is 1 mm. Most importantly, there is an external opening at the top into which a syringe containing the “core” material can be inserted and locked by the tap screw, and the central body has an oblique side arm with another opening for a syringe that can direct material into the concentric internal (6 mm diameter) space to constitute the “shell”. Thus, by control of the screwable tap, the extension of the syringe needle end into the aluminum hollow cylinder can be adjusted. The oblique side arm also has an opening for the needle of a syringe containing the shell material that is close to the terminus of the central hollowedout cylinder. The side arm is connected to a high-voltage power supply that provides a variable high-voltage positive Langmuir 2009, 25(22), 13250–13257

Magalh~ aes et al.

Article

Figure 4. FESEM from a co-electrospun sample under the following conditions: 8 cm tip distance to bath collector, 16 kV applied to tip, shell flow rate 0.763 cm3/h, and flow rate ratio shell to core of 1.8. Two distinct regions were found in micrograph (a) (50 μm inset bar) whose details are shown in (b) (more porous area, 5 μm inset bar) and (c) (more solid area, 5 μm inset bar). bias (up to 20 kV) relative to a grounded water-immersed collector plate. Both the top and side syringes contain separate fluids whose flow rate is controlled by a single syringe pump (to ensure homogeneous flow rate). The volume ratio of the shell solution to the core suspension is dictated by the relationship between the diameter sizes of the syringes that are controlled by the same syringe pump. According to past work,28 the top internal needle should extend outside the bottom of the aluminum hollowed-out cylinder by half of the value of the hollow “core” space internal diameter. Thus, in this case, the tap should be controlled such that 0.5 mm of the internal needle extends beyond the aluminum hollow cylinder tip. This procedure will ensure that the fiber produced by the applied high voltage will extrude around the core suspension. A heat gun blowing hot air was used to heat the needles to ∼120 °C. This temperature was measured by a thermometer before applying the high voltage. The device was then connected to a high-voltage power supply (Gamma). The electrospun fibers produced were collected onto a metal screen under a cold water bath at ∼10 °C. Aluminum screens were attached to aluminum foil using adhesive tape and immersed in water having several drops of detergent to decrease the surface tension and completely wet the metal screen. The screen was submerged ∼1 mm under water. The aluminum foil was grounded. A neat cellulose-doped solution consisted of 1.5 wt % of cellulose in NMMO/H2O with DMSO as the shell solution. Typically, 2.944 g of NMMO/H2O, 0.736 g of distilled water, and 0.062 g of cellulose were heated over a hot plate and mixed until complete dissolution after evaporation of the water, and then 1.05 g of DMSO was added dropwise under mixing. The main steps in cellulose nanocrystals synthesis were (i) acid hydrolysis (64% H2SO4, 55 °C, 2 h), (ii) refining and purification of the solid residue remaining after the hydrolysis by repeated 20-40 min cycles of ultracentrifuging at 7000 rpm followed by resuspension of the solids in distilled water until a turbid supernatant was obtained, and (iii) dialysis against distilled water to pH 6-7. A cellulose nanocrystals water suspension was then used as the core material. Cellulose nanocrystals water suspensions at two concentrations, 0.42 and 3.28 wt %, were used as the core material. These concentrations were determined by weighing before and after total suspension water evaporation and subsequent nanocrystals drying at 100 °C. Several combinations of shell/core volume ratio, applied voltage, and tip distance to collector were tested. FESEM. The samples were coated by a thin 8-10 nm layer of sputtered AuPd and analyzed in a field emission scanning electron microscopy JEOL JSM-6400F scanning microscope that allows for high resolution (15-70 A˚). The digital images were captured using the 4pi Universal Spectral Engine. (28) Reznik, S N.; Yarin, A. L.; Zussman, E.; Bercovici, L. Phys. Fluids 2006, 18, 062101-1–062101-13.

Langmuir 2009, 25(22), 13250–13257

FT-IR. A NEXUS 670 FT-IR (Thermo Nicolet) was used to obtain spectra of the films in transmission mode with the IR beam passing through the cellulose films. The spectra were collected in 2 cm-1 intervals over the range 4000-650 cm-1. TG and DTG. Thermograms were collected using TA equipment under a N2 flux rate of 60 cm3/min and at a heating ramp of 10 °C/min from 30 to 600 °C. XRD. All diffractograms were obtained with a Bruker D-5000 diffractometer equipped with a Highstar area detector, 40 kV, and 30 mA for the generator power. Crystallinity was evaluated in a conventional way assuming that the highest intensity at ∼21 deg is proportional to the amount of crystalline plus amorphous cellulose and the local minimum is ∼14 deg for amorphous cellulose.

Results and Discussion Figure 2 illustrates the FESEM of three electrospun cellulosedoped solution samples. As evident from the micrographs, a low flow rate obtains fibers as thin as 70 nm in diameter; however, they severely adhere to each other easily forming a film. This occurs because the generated thin fibers cannot easily overcome the water surface tension of the collector bath. Nonetheless, increasing the flow rate does allow the fibers to become more individualized and display diameters in the range 300-800 nm; yet, films are still formed. On the other hand, if a metal collector is employed, the electrospun fibers coagulate into droplets. Figure 3 illustrates FESEM images from a film obtained from core and shell co-electrospinning while applying a voltage of 10 kV and fixing the tip distance to bath collector at 8 cm, a flow rate ratio shell-to-core of 0.9 (syringe volume ratio of shell to core=3/2), and a shell flow rate of 0.402 cm3/h. Many fibers have condensed to form a film; however, the micrograph demonstrates that the final sample remains a porous structure with fibers having diameters of ∼100 nm. Figure 4 illustrates FESEM images of a film obtained from core and shell co-electrospinning after applying a voltage of 16 kV and fixing the tip distance to the bath collector at 8 cm, the flow rate ratio of shell to core=1.8, and a shell flow rate at 0.763 cm3/h. Under these conditions, the obtained film displayed two distinct morphologies: one being porous with more individualized fibers and the other being a filmlike structure due to the adhesion of the fibers. Figure 5 shows the FESEM of films obtained by co-electrospinning under three different conditions of voltage and flow rate to ascertain the influence of these parameters on morphology. The distance from tip to bath collector and the flow rate ratio of shell to core were kept constant at 10 cm and 9.4, respectively. Figures 5a,b show a sample obtained using 16 kV and a shell flow rate of 0.949 cm3/h; Figures 5c-e show films obtained from 16 kV DOI: 10.1021/la901928j

13253

Article

Figure 5. FESEM of films obtained at 10 cm distance from tip to bath collector and a flow rate ratio shell to core of 9.4: (a) 16 kV, 0.949 cm3/h, (b) details of (a); (c) 16 kV, 0.397 cm3/h, (d) and (e) different regions of (d); (f) 20 kV, 0.397 cm3/h.

and flow rate of 0.397 cm3/h, and Figure 5f shows a film obtained by using 20 kV and flow rate of 0.397 cm3/h. It was found that as the volume ratio of shell to core increased, it became more difficult to obtain individualized fibers. This can be explained by the water inside the core that can diffuse into the cellulose-doped solution in the shell. A lower shell-to-core ratio indicates a higher relative concentration of water/fiber (due to the core) which can coagulate the cellulose shell and thus prevent or impede adhesion between fibers. From the FESEM images, it was concluded that individualized nanofibers may be produced at the lower flow rates, flow rate ratios of shell to core, and applied voltages. Figure 6 shows typical XRD diffractograms of cellulose type II illustrating the evolution of the characteristic doublet at the main peak (clear in sample 12). The diffractograms correspond to several nanocomposite films made by electrospinning whose parameters for obtaining the samples are listed in Table 1. Analyses of the diffractograms in Figure 6 indicate that sample 7 has the highest crystallinity index value (67%), while samples 12 and 5 are 64% and 34%, respectively. All samples were made by co-electrospinning at different conditions. Samples 7 and 5 were composed of diluted nanocrystals (NC, 0.42 wt %) core suspension, while sample 12 had a NCs core suspension that was nearly 10-fold more concentrated (3.28 wt %). Relative to the shell/core flow rates (sample 7 is formed at a shell/core flow rate that is nearly 5-fold more that 5 and 12), although sample 7 contains a lower relative concentration of NCs inside the core compared to samples 5 and 12, it surprisingly displays a higher crystallinity. In parallel with the XRD diffractograms, FT-IR spectra of the films obtained by electrospinning and co-electrospinning were 13254 DOI: 10.1021/la901928j

Magalh~ aes et al.

obtained to provide crystallinity indices comparisons. The obtained spectra presented typical signatures for cellulose.29 The peak appearing at 1111 cm-1 is quite strong and indicative of cellulose crystallinity type I appearing only as a shoulder in cellulose II and amorphous cellulose due to the development of a strong, broad band near 1090 cm-1.30,31 The strong 1111 cm-1 peak can be observed in the spectra of samples made of cellulose reinforced with cellulose nanocrystals. However, several of the samples did not display this peak due to the lack of nanofiber formation during co-electrospinning of the NCs suspension. The FT-IR spectra can be used to calculate the crystallinity index, which is generally in good agreement with XRD measurements.30-32 The absorption ratio with the best fit to XRD measurements corresponds to the FT-IR bands at 1372 and 2900 cm-1 and can be used to predict crystallinity even for a mixture of cellulose types I and II. The relative crystallinity indices calculated from the FT-IR spectra are shown in Table 1. Many of the calculated FT-IR crystallinity indices were quite high relative to the published data, indicating that the electrospinning technique does indeed align the cellulose molecules. The crystallinity slightly increases with flow rate, although not linearly for samples produced without a nanocrystals suspension inside the core. The XRD results corroborate this FT-IR analysis. Increasing the amount of NC as reinforcement causes an increase in the FT-IR crystallinity indices, as would be expected because these indices correlate well with a crystallinity of a mixture of cellulose types I and II.32 However, as shown with the XRD, the samples with the highest FT-IR crystallinity indices were those without NCs. In general, FT-IR crystallinity indices increase with flow rate ratio shell-to-core and applied voltage. Past work indicates that cellulose type I (the NCs) does not induce increases in cellulose type II crystallization unless the amorphous region is greater than 75% of the total cellulose available which is not the case here.32 The flow rate of the shell does not affect the FT-IR crystallinity indices for co-electrospinning experiments, despite the effect it has on spinnability. The only two samples (samples 5 and 7) obtained by coelectrospinning that do not show cellulose type I in the FT-IR spectra displayed the highest values for FT-IR crystallinity indices. These indices are greater than those evaluated for electrospun films without a core suspension. Thus, it can be inferred that the absence of core NCs allowed enhanced cellulose alignment in an electrospun fiber. The thermograms (TGA and DTG) shown in Figure 7 correspond to oven-dried NCs from 3.28 wt % water suspensions (A), a film without a core (B), and electrospun films with shells and cores with increasing amounts of NCs inside the cores (C-E). The TGA and DTG of oven-dried NCs are distinctly different from the thermograms of freeze-dried NCs (not shown here). Two main events occur at approximately 180 and 381 °C for ovendried NCs: the porous NCs obtained by freeze-drying show a thermal event at higher temperatures, at approximately 270 and 380 °C. Films with a higher content of NCs in the core clearly display a thermal event at ∼189 °C (see Figures 7D,E) that can be attributed to the presence of NCs in the fiber cores. This peak does not appear for the films made without a core; moreover, it cannot be observed in the films with low concentrations (0.42 wt %) of (29) Fengel, D. Holzforschung 1992, 46, 283–288. (30) Nelson, M. L.; O’Connor, R. T. J. Appl. Polym. Sci. 1964, 8, 1311–1324. (31) Carrillo, F.; Colom, X.; Sunol, J. J.; Saurina, J. Eur. Polym. J. 2004, 40, 2229–2234. (32) El-Wakil, N. A.; Hassan, M. L. J. Appl. Polym. Sci. 2008, 109, 2862–2871.

Langmuir 2009, 25(22), 13250–13257

Magalh~ aes et al.

Article

Figure 6. XRD diffractograms of cellulose nanocomposites obtained by co-electrospinning. Sample descriptions can be found in Table 1. Table 1. Crystallinity of Cellulose and Nanocrystals Cellulose-Reinforced Cellulose Fibers Using FT-IR Absorption Relationships and XRD Diffractograms samples with NMMO.H2O/DMSO as shell solvent 1 2 3 4 5 (no detection of NCs by FT-IR) 6 7 (no detection of NCs by FT-IR) 12 13 14 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

wt % NCs in core suspension

syringe size ratio (shell/ core mL/mL)

shell/core flow rate ratio (cm3 h-1/cm3 h-1)

shell flow rate (cm3/h)

total flow rate (cm3/h)

voltage (kV)

distance (cm)

20 20 20 10 16

10 10 10 8 8

0.94 0.95 1.06 0.85 1.13 (XRD 35%) 1.16 1.35 (no detection of NCs by FT-IR) (XRD 67%) FT-IR crystallinity was not possible to evaluate (XRD 64%) 0.87 0.60 0.55 0.52 0.60 0.64 0.46 0.52 0.59 0.54 0.56 0.53 0.51 0.72 0.56 0.63 0.56 0.49

0.42 0.42

3/2 5/2

0.9 1.8

0.305 1.068 2.286 0.402 0.763

0.42 0.42

30/3 10/1

6.3 9.4

0.949 0.397

0.439

16 16

10 10

3.28

10/5

0.793/0.508 (1.56)

0.793

1.301

18

8

3.28 3.28 0.42 3.28 3.28 3.28 3.28 0.42 0.42 0.42 3.28 3.28 0.42 0.42 0.42 0.42 0.42 0.42

10/1 3/3 3/3 10/5 3/3 3/3 3/3 3/3 3/3 3/3 10/1 10/1 10/5 10/5 10/5 10/5 10/1 10/1

0.793/0.084 (9.4) 0.668/0.668 (1) 0.267/0.267 (1) 0.397/0.254 (1.56) 1 0.267/0.267 (1) 0.267/0.267 (1) 1 0.668/.668 (1) 1 0.397/0.042 (9.4) 0.397/0.042 (9.4) 1.56 0.793/0.508 (1.56) 1.56 0.397/0.254 (1.56) 9.4 9.4

0.793 0.668 0.267 0.397 0.267 0.267 0.267 0.668 0.668 0.267 0.397 0.397 0.793 0.793 0.397 0.397 1.19 0.793

0.877 1.336 0.534 0.651 0.534 0.534 0.534 1.336 1.336 0.534 0.439 0.439 1.301 1.301 0.651 0.651 1.488 0.877

18 18 18 18 14 18 18 14 18 14 14 18 14 18 14 18 12 12

8 8 8 10 8 8 8 8 8 8 10 10 8 8 8 8 8 8

Langmuir 2009, 25(22), 13250–13257

crystallinity indices by FT-IR A1372/A2900

DOI: 10.1021/la901928j

13255

Article

Magalh~ aes et al.

Figure 7. TGA and DTG of (A) nanocrystals oven-dried from a 3.28 wt % water suspension; (B) film without core from 1.56 wt % dope solution in NMMO/DMSO; (C) film from flow rate ratio 9.4, 1.56 wt % dope solution in NMMO/DMSO, 0.42 wt % NCs water suspension; (D) film from flow rate ratio 1, 1.56 wt % dope solution in NMMO/DMSO, 3.28 wt % NCs water suspension; (E) film from flow rate ratio 1.56, 2.0 wt % dope solution in NMMO, 3.28 wt % NCs water suspension.

NCs (Figure 7C). Despite the fact that cellulose of crystallinity type I should be much more easily observed in the FT-IR spectra, it was not (at the low 0.42 wt %), thus providing corroborating evidence for the thermal data in Figure 7. The nanocomposites show a thermal event with a shift of ∼9 deg toward higher temperature compared to the pure NCs thermogram. This phenomenon is due to interactions between the solvents (NMMO and DMSO) and the NCs. It is postulated that after collection in the water bath collector, at the very least, the NCs surface may become less ordered, becoming amorphous or type II cellulose. 13256 DOI: 10.1021/la901928j

From the TG/DTG thermograms and also FT-IR spectral analysis, it can be inferred which parameters during the co-electrospinning experiment are important for all-cellulose nanocomposites reinforced with NCs. The two most important parameters are the cellulose dope solution formulation and NCs suspension concentration. For example, when NMMO 3 H2O instead of NMMO 3 H2O/ DMSO was used as the solvent, a 2 wt % cellulose dope solution was easy to generate that should potentially display a higher reinforcement with NCs. However, higher reinforcement was not what was found; this finding likely can be explained by the fact that Langmuir 2009, 25(22), 13250–13257

Magalh~ aes et al.

Article

the viscosity (and surface tension) of the shell solution was likely high enough to force orientation/placement of the core NCs suspension but not quite high enough to impede nanofiber formation. Although the dope solution was able to form nanofibers with NCs inside the core, this suspension was not concentrated enough to allow the NCs to be detected by TG and XRD, although in general they were by FT-IR, which is an altogether more sensitive technique. Samples prepared with NMMO or NMMO/DMSO as the dope solution for the shell material, using a flow rate ratio close to 1, and high flow rate, are unstable for electrospinning mainly at low voltages. In this case, the submicrometer fibers formed are not spun fast enough to simultaneously remove mass at the flow rate provided by the syringe pump for the fluids. Thus, big drops are formed on the tip that must be removed; otherwise, these will drop into the water bath collector. This instability during the electrospinning process may therefore result in a decreasing amount of NCs in the core of the spun fibers.

Conclusions We provide the first evidence to demonstrate that it is possible to co-electrospin a core-in-shell nanomaterial consisting of

Langmuir 2009, 25(22), 13250–13257

a cellulose solution in the shell and a cellulose nanocrystals water suspension in the core. By decreasing the shell-to-core volume ratio, it was found that precise control of the voltage and flow rate promoted individual fiber formation. NCs suspensions were found to be easily electrospun when using solely NMMO.H2O as the carrier likely because of its higher viscosity relative to a NMMO/DMSO dual solvent carrier. Finally, during the electrospinning of a neat cellulose solution, any increases in the flow rate slightly increased the cellulose crystallinity, but no effect was observed during co-electrospinning. Acknowledgment. We acknowledge the Brazilian agency CNPq for the generous research fellowship awarded to WLEM that allowed this work to be realized. We also acknowledge Profs. O. Rojas (Wood & Paper Science) and E. Loboa (Biomedical Engineering) for their advice and insight and the NC State Metal Workshops for their expert fabrication of the equipment described herein. Finally, portions of this work were made possible under the auspices of USDA Grant 2006-3841117035.

DOI: 10.1021/la901928j

13257