Article pubs.acs.org/Biomac
Fluorescent Labeling and Characterization of Cellulose Nanocrystals with Varying Charge Contents Tiffany Abitbol,*,† Anthony Palermo,‡ Jose M. Moran-Mirabal,‡ and Emily D. Cranston† Departments of †Chemical Engineering and ‡Chemistry and Chemical Biology, McMaster University, 1280 Main Street West, Hamilton, Ontario, Canada S Supporting Information *
ABSTRACT: Cotton-source cellulose nanocrystals (CNCs) with a range of surface charge densities were fluorescently labeled with 5-(4, 6-dichlorotriazinyl) aminofluorescein (DTAF) in a facile, one-pot reaction under alkaline conditions. Three CNC samples were labeled: (I) anionic CNCs prepared by sulfuric acid hydrolysis with a sulfur content of 0.47 wt %, (II) a partially desulfated, sulfuric acid-hydrolyzed CNC sample, which was less anionic with an intermediate sulfur content of 0.21 wt %, and (III) uncharged CNCs prepared by HCl hydrolysis. The DTAF-labeled CNCs were characterized by dynamic light scattering, atomic force microscopy, fluorescence spectroscopy and microscopy, and polarized light microscopy. Fluorescent CNCs exhibited similar colloidal stability to the starting CNCs, with the exception of the HCl-hydrolyzed sample, which became less agglomerated after the labeling reaction. The degree of labeling depended on the sulfur content of the CNCs, indicating that the presence of sulfate half-ester groups on the CNC surfaces hindered labeling. The labeling reaction produced CNCs that had detectable fluorescence, without compromising the overall surface chemistry or behavior of the materials, an aspect relevant to studies that require a fluorescent cellulose substrate with intact native properties. The DTAF-labeled CNCs were proposed as optical markers for the dispersion quality of CNC-loaded polymer composites. Electrospun polyvinyl alcohol fibers loaded with DTAF-labeled CNCs appeared uniformly fluorescent by fluorescence microscopy, suggesting that the nanoparticles were well dispersed within the polymer matrix.
■
half-ester groups,6 most likely due to repulsive interactions with the negatively charged residues of the enzyme’s cellulose binding or catalytic domains. The presence of non-native functional groups has also been shown to interfere with the enzymatic digestion of bulk cellulose, for example, Pan et al.7 found that the acetyl groups in acetic acid pulps inhibited enzymatic activity. In the current work, CNCs with different surface charge densities were fluorescently labeled with 5-(4, 6dichlorotriazinyl) aminofluorescein (DTAF), a fluorophore that reacts directly with the hydroxyl groups of cellulose. This particular dye has been used previously to label other cellulose substrates,8−10 but we believe this work to be the first account of its application to CNCs. Other groups have fluorescently labeled CNCs using dyes with isothiocyanate,11−13 succinimidyl ester,12 and pyrene moieties.14 The reactions typically have three steps and involve the introduction of primary amino groups that are then reacted with the isothiocyanate or succinimmidyl ester dyes, although Nielsen et al.12 achieved fluorescein-5′-isothiocyanate (FITC) and rhodamine B isothiocyanate (RBITC) labeled CNCs in a single step.
INTRODUCTION Cellulose nanocrystals are rod-shaped particles typically prepared from the acid hydrolysis of native celluloses.1 The size of the crystallites depends mostly on cellulose source and hydrolysis conditions, spanning from 4 to 7 nm in diameter by 50−300 nm in length for wood or cotton CNCs to 10−50 nm in cross-section by 0.1 to several micrometers in length for bacterial cellulose CNCs. In general, CNCs retain the cellulose I native crystal structure and are thus ideal substrates for experiments that probe the fundamental properties and interactions of natural crystalline cellulose, including enzyme degradation studies. An important aspect of CNCs prepared by sulfuric acid hydrolysis is that the treatment grafts sulfate halfesters onto the surface of the crystallites, which are negatively charged above pH 2.5.2,3 The charged groups impart colloidal stability to aqueous suspensions of CNCs, helping through electrostatic repulsive interactions to overcome the natural tendency of cellulose particles to associate. The presence of sulfate half-esters, estimated at about 1 sulfate half-ester for every 10 surface anhydroglucose units (AGUs),4,5 is attractive for water-based dispersion and processing of CNCs, but may lead to unexpected properties in terms of cellulose reactivity. For instance, a recent publication found that the enzymatic hydrolysis of cellulose was hindered by the presence of sulfate © XXXX American Chemical Society
Received: June 16, 2013 Revised: July 25, 2013
A
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
It was our goal to prepare fluorescent CNC substrates and to probe the effect of charge on reactivity using a straightforward synthesis protocol. Reliably prepared and characterized fluorescent CNCs are needed for studies that require particle localization, such as biotoxicity assays and enzymatic studies. Fluorescent CNCs may also prove useful in the fabrication of new types of optically active CNC-based sensing materials; for example, DTAF fluorescence is pH-sensitive and may therefore find use as a pH-responsive system similar to that proposed by Nielsen et al.12 Here, DTAF-labeled CNCs were shown to exhibit very similar physicochemical properties to unlabeled CNCs, and as a proof of concept, the labeled CNCs were used as an optical marker to elucidate the dispersion of nanoparticles within a polymer matrix.
■
reaction products were sonicated using a Branson 450 sonifier for a few minutes in order to disperse agglomerates that may have formed during the ultrafiltration process. All three CNC suspension types were DTAF DTAF labeled and are denoted as CNCDTAF sulf. , CNCdesulf. , and CNCHCl . Electrospun Polyvinyl Alcohol (PVA) Fluorescent Fibers. Fluorescently labeled sulfated CNCs were added to aqueous solutions of polyvinyl alcohol (Fluka, 72 kDa nominal molecular weight, 97.5− 99.5 mol % hydrolysis) to give an overall PVA concentration of ∼7 wt % and CNC content of ∼0.7 wt %. This was accomplished by adding 0.25 mL of a 2 wt % CNC suspension to 0.5 mL of a 10 wt % PVA solution. Ultracentrifugation was used to achieve the small volume of concentrated labeled CNCs. Electrospinning conditions to produce uniform fibers with sub-500 nm diameters were the following: 12.5 kV voltage and 10 cm working distance from tip to grounded collector, which was a glass coverslip. Conductometric Titration. Conductometric titrations were performed on the starting CNC materials in order to determine the charge content of the crystallites. For CNCs prepared from sulfuric acid hydrolysis, the technique monitors the change in conductivity that occurs as the strong acid groups associated with the anionic surface sulfate half-esters are titrated against dilute sodium hydroxide. Typically, a mixture of suspension (∼0.05 g cellulose) and dilute NaCl solution (75 g of a 1 mM solution) was titrated against 2 mM NaOH. The equivalence point from titration is related to the surface charge density of the crystallites, reported as percent sulfur content by weight,18 according to eq 1:
MATERIALS AND METHODS
Preparation of CNCs. (I) CNCs from sulfuric acid hydrolysis (CNCsulf.) were prepared using a previously published protocol, involving the reaction between cellulose (Whatman ashless filter aid) and concentrated sulfuric acid (64 wt %) at an acid to cotton ratio of 17.5 mL/g, at 45 °C for 45 min.15 The CNCs were purified by extensive dialysis until the pH of the external dialysis reservoir stabilized. The CNCs were then treated with Dowex Marathon C, a cationic exchange resin, to ensure complete protonation of sulfate halfesters. (II) Partially desulfated CNCs (CNCdesulf.) were prepared by acid catalyzed desulfation according to the method of Jiang et al.16 Briefly, the original sulfated CNC suspension (114.51 g, 4.39 wt %) was diluted to give a 1 wt % suspension, acidified to 0.025 M by the addition of HCl and reacted at 80 °C for 20 h, with stirring, after which point the reaction was quenched by immersion in an ice bath. The reaction was neutralized by dialysis against deionized water until the pH of the suspension no longer fluctuated. Finally, the resultant suspension was sonicated on ice using a Branson 450 sonifier (20 min, 60% amplitude) and filtered through a glass microfiber filter (0.45 μm pore size) to give a final concentration of 0.852 ± 0.009 wt %. (III) Uncharged CNCs (CNCHCl) were prepared from cotton (Whatman ashless filter aid) using an adaptation of the protocol published by Araki et al.17 Cotton powder (10 g) was mixed with HCl (100 mL, 2.5 N) and reacted at 100 °C for 15 min. The reaction was quenched by immersion in an ice bath. The cellulosic material was collected and neutralized by vacuum filtration through glass microfiber filters. Approximately 200 mL of deionized water was combined with the filtered solids, and the mixture was blended for 30 min in a Waringtype blender. Finally, the cellulosic slurry was centrifuged repeatedly (ca. 20 times) at 600g for 5 min cycles using a Beckman Coulter Allegra X-12R with a SX4750A rotor. The supernatant was collected to give a dilute suspension of CNCs with a final concentration of 0.12 ± 0.02 wt %. Unless otherwise noted, reported confidence intervals are calculated from the standard deviation (Sx) of replicate measurements (N) as Δx = Sx × t-value/(N)1/2, where the t-value is the Student’s tdistribution at a confidence level of 95% for N − 1 degrees of freedom. DTAF-Grafted CNCs. The CNCs were labeled with DTAF using an approach that has been previously applied to bacterial cellulose samples.8−10 DTAF (BioReagent, ≥99.0%) was purchased from Sigma-Aldrich and used without any further purification. The CNCs (500 mg) were reacted with DTAF (7.5 mg) under alkaline conditions (0.2 M NaOH) for 24 h, in the dark, with stirring. While the mass ratio of CNCs to DTAF was constant (500 mg CNCs: 7.5 mg DTAF), the volume of the reactions depended on the concentration of the initial suspension. Solid NaOH and DTAF were added directly to the appropriate volume of CNC suspension, 12.5 or 50 mL for a 4 or 1 wt % suspension, respectively. To remove the bulk of the NaOH and unreacted DTAF, a combination of centrifugation (Beckman Coulter Allegra X-12R with a SX4750A rotor, 3300g, 30 min cycles) and dialysis (volume of water approximately 100 times the volume of the sample, changed daily) were employed, followed by further purification in a Millipore solvent-resistant stirred cell fitted with a Millipore ultrafiltration disc and operated at 20−30 psi. Purified
%S =
VNaOHC NaOHMW (S) × 100% msuspCsusp
(1)
where VNaOH is the volume of NaOH at the equivalence point, CNaOH is the concentration of NaOH, MW(S) is the molecular weight of sulfur, msusp and Csusp are the mass and concentration of the suspension, respectively. Degree of Substitution. The degree of substitution of the DTAFlabeled CNCs was assessed from fluorescence spectroscopy using a Tecan Safire microplate reader. Bottom reads were obtained for samples (200 μL) placed in the wells of a 96-well Costar black plate with transparent bottoms. The DTAF content of the labeled CNC samples was determined from calibration curves made from DTAF standards (0−500 nM range). The pH of the standards and samples were adjusted to 11 prior to measurement since the fluorescence of DTAF is pH-dependent. Fluorescence scans were obtained using the following settings: gain = 75 (optimized to 500 nM standard solution), 5 nm slit widths, a 488 nm excitation, and emission scanned from 500 to 650 nm. Three independent measurements were obtained from each labeled sample. Atomic Force Microscopy (AFM). A Nanoscope IIIa MultiMode Scanning Probe Microscope with an E scanner (Bruker AXS, Santa Barbara, CA) was used in tapping mode to image the samples, using probes from Asylum Research (AC160TS type, 42 N/m spring constant, and 300 kHz resonance frequency). Samples were prepared by depositing a 0.1 wt % solution of polyallylamine hydrochloride (Polysciences Inc., 120−200 kDa nominal molecular weight) onto freshly cleaved mica for a few seconds, rinsing with deionized water, dropping a CNC suspension (0.001 wt %) onto the substrate for a few more seconds, followed by a final rinse. The samples were left to dry at ambient temperature. Samples on silicon wafers with a native oxide layer (ca. 20 Å thick) were prepared similarly, except the deposition steps were about 1−2 h and the CNC concentration was 10-fold. Prior to deposition, the silicon wafers were cleaned by rinsing in ethanol and drying under a stream of air. Average particle sizes and standard errors were calculated from 25 to 150 isolated particles. Polarized Optical Microscopy. Color micrographs were obtained using a Nikon Eclipse LV100POL microscope with a 530 nm phase retardation plate. With the retardation plate, isotropic areas appear bright pink and oriented regions, blue or yellow, the difference being a 90° in-plane rotation of the sample alignment direction. Digital images were taken of wet samples, sandwiched between a glass slide and a glass coverslip, as they slowly evaporated. B
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Dynamic Light Scattering (DLS). Although the DLS technique is intended for spherical particles, and CNCs are rod-shaped, the measurements can be used to compare relative sizes of CNCs. The apparent DLS particle “size” of 0.025 wt % suspensions, diluted in pure water, was measured using a Malvern Zetasizer Nano S. Three independent measurements were obtained for each sample. Electrophoretic Mobility. The electrophoretic mobility of the CNCs was measured using a Zeta Potential ZetaPlus Analyzer (Brookhaven) with 5 mM NaCl added to a 0.25 wt % CNC suspension, except for the more turbid HCl-hydrolyzed suspension which was measured at 0.1 wt %. The reported electrophoretic mobility is an average of 10 measurements. Fluorescence Microscopy. Fluorescence micrographs of the labeled CNCs were obtained using an Olympus BX51 upright microscope equipped with a Retiga 2000R Camera, and excitation, dichroic, and emission filters appropriate for the observation of FITC fluorescence. Samples were prepared from 1:9 dilutions of the samples (∼1 wt %) in isopropanol by dragging a 2 μL volume across the surface of a glass slide. Fluorescence images of the CNC-doped PVA fibers were obtained using a Nikon Eclipse Ti microscope, with an Andor iXon3 camera and an Intensilight C-HGFI illuminator.
Figure 2. Fluorescence spectra of DTAF-labeled CNCs obtained using an excitation wavelength of 491 nm. Labeled samples were prepared from either 1 wt % CNC suspensions or 4 wt % CNC suspensions, as indicated.
■
RESULTS AND DISCUSSION Cotton-derived CNCs with a range of surface charge densities were prepared using previously established methods as
Table 1. Degree of DTAF Labeling of CNC Samples Achieved with a Reaction Ratio of 1.5 mg DTAF per 100 mg Cellulose grafting density
CNCDTAF sulf. CNCDTAF desulf. CNCDTAF HCl a
CNC reaction concentration (wt %)
DTAF content (nmol/g cellulose)
degree of substitution (DTAF/105 AGU)
4 1 1 1a 1
24 ± 1 4.5 ± 0.1 11.1 ± 0.1 11.7 ± 0.3 47.5 ± 0.4
4 0.7 1.8 1.9 7.7
Repeat.
from 0.15 to 0.06 mequiv of charge per gram of cellulose (meq/ g) after the 20 h reaction. The surface charge achieved here corresponds well to the 0.055 meq/g reported by Jiang et al.19 after only 7 h with the same desulfation procedure. It is interesting that both Jiang et al.19 and the current work achieved the same final degree of desulfation despite differences in reaction time and the starting surface charge density of the CNCs (Jiang et al.16 had a sulfur content nearly twice our starting value). These observations may indicate a fundamental limitation to the removal of surface sulfate half-esters by acid hydrolysis, possibly due to the position of the sulfate half-ester group. The three CNC samples were labeled with DTAF according to Figure 1. The reaction is straightforward and proceeds in a one-step nucleophilic attack of the triazinyl ring by deprotonated cellulose hydroxyls, with chloride as the leaving group. Although the reaction itself proceeds virtually untended over a 24 h period, purification and isolation of the labeled CNCs was lengthy but necessary in order to gain a true, quantitative measure of the degree of labeling. The cleanup of labeled suspensions first involved extensive dialysis (minimum 2 weeks with daily water changes) in foil covered columns to prevent photobleaching. Dialysis was considered to be complete when the presence of DTAF could no longer be detected in the external dialysis reservoir by UV−vis spectroscopy and the pH of the reservoir matched pure water. Samples purified by dialysis alone gave DTAF contents in the mmol/g
Figure 1. Schematic representation of the DTAF labeling reaction of CNC samples with different surface charge densities: (I) sulfated CNCs (CNCsulf.), (II) partially desulfated CNCs (CNCdesulf.), and (III) uncharged CNCs (CNCHCl), and their labeled counterparts (Ia), (IIa), and (IIIa).
described in Materials and Methods: (I) negatively charged CNCs from sulfuric acid hydrolysis (CNCsulf.), (II) CNCs with an intermediate negative charge by acid-catalyzed desulfation of CNCs from sulfuric acid hydrolysis (CNCdesulf.), and (III) uncharged CNCs prepared by direct HCl hydrolysis (CNCHCl). The native cellulose I crystal structure is retained in samples which have undergone the hydrolysis and desulfation reactions employed in this work.16 The acid-catalyzed desulfation reduced the surface charge density (represented in terms of elemental sulfur) from 0.47 ± 0.04% S for CNCsulf. to 0.21 ± 0.03% S for CNCdesulf.. This translates into a 60% decrease in the number of charged groups C
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
DTAF Figure 3. Epifluorescence microscopy of DTAF-labeled CNC samples deposited onto glass substrates: (a) CNCDTAF sulf. , (b) CNCdesulf. , and (c) . (All labeled CNCs are from reactions with DTAF and 1 wt % CNC suspensions.) To prepare the microscopy samples, labeled CNC CNCDTAF HCl suspensions were diluted in isopropanol (9:1 v/v), and a small volume (∼2 μL) was dragged across the substrate using a micropipet tip.
DTAF was achieved when the concentration of CNCsulf. was quadrupled. The shift toward a higher labeling density can be explained by an increase in the collision frequency between dye molecules and CNCs, and by the increase in ionic strength, which decreases the electrostatic repulsion between CNCs and DTAF by compressing the electrostatic double layer. Results from two identical but independently prepared and purified reactions using CNCdesulf. yielded very similar results, at 11.1 and 11.7 nmol DTAF/g cellulose, confirming the reproducibility of the method. The degree of labeling achieved in this work is lower than that reported in similar studies of labeled CNCs; for example Dong and Roman reported a FITC content of 0.03 mmol/g cellulose,11 and Nielsen et al.12 achieved FITC and RBITC contents of 2.8 and 2.1 μmol/g cellulose, respectively. Direct comparisons are difficult because the dye functionalities, reaction mechanisms, and conditions were different; however, it is likely that, in some measure, the difference in results can be attributed to the high degree of purity achieved in this work by using the ultrafiltration stirred cell. The stirred cell allowed an improved measure of purity since the CNC medium was sampled directly, in comparison to attempting to detect the presence of small amounts of dye diluted within a dialysis reservoir. For single particle localization and high-resolution fluorescence microscopy studies, it is critical that the suspensions be free of unbound dye and other contaminants. The relatively low grafting density achieved here, compared to DTAF-labeled bacterial cellulose microfibrils (Degree of substitution = 0.0035−0.025 mol DTAF/mol AGU; ∼0.02−0.2 mmol DTAF/g cellulose) is mostly related to differences in the surface charge density and the crystalline structure of the cellulose substrates.8 Helbert et al.8 found that the conversion of the native cellulose I crystal structure to cellulose III led to an improvement in the grafting density of DTAF-labeled bacterial cellulose fibrils, probably a consequence of the lower packing density of chains in the cellulose III crystal structure.20,21 Importantly, the X-ray diffraction studies in their work indicated that the DTAF-labeling reaction did not alter the crystal lattice of the cellulose substrates.8 CNCs are often lauded for a large surface-area-to-volume ratio and an abundance of reactive surface hydroxyl groups; however, the highly crystalline native structure may explain why CNCs rarely seem to live up to their theorized functionalization potential, with the exception of (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) oxidized CNCs, which have a high grafting density of carboxylic acid groups. Low grafting densities may pose a challenge for modifications that aim to dramatically alter the
Figure 4. Ambient light (top) and UV-light (bottom) photographs of initial CNC suspensions and corresponding labeled materials. The concentration of all six samples is 0.5 wt % CNCs. The CNCs were from DTAF-labeling reactions conducted using 1 wt % CNCs.
cellulose range but with no clear trend with regard to the surface charge density of the CNCs. When the labeled CNC suspensions were additionally purified in an ultrafiltration stirred cell, the DTAF contents dropped significantly to the nmol/g cellulose range, reproducibility improved and the effect of surface charge on the reaction efficiency emerged from the results. In general, the dialyzed samples had to be rinsed 10−20 times in the stirred cell to ensure complete removal of unreacted materials. Purified samples were then analyzed by fluorescence spectroscopy (Figure 2) to extract the degree of labeling (Table 1). When the three different CNC samples were reacted under the same conditions, the degree of labeling increased with decreasing surface charge. The decrease in fluorophore grafting density with increasing surface charge is most likely related to electrostatic double layer repulsion between DTAF, which is negatively charged under the reaction conditions, and the sulfate half-esters. The grafting density also depended upon CNC concentration; a 5-fold increase in the grafting density of D
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Figure 5. AFM height images of unlabeled and DTAF-labeled CNC samples; 0.5 μm scale bars.
Visual comparison of the turbidity of the samples before and after the labeling reaction suggested that the stability of the CNCs either remained the same or was improved by the labeling reaction. Photographs taken in ambient light of the three initial CNC samples and their labeled counterparts, all at 0.5 wt %, are presented in the top panel of Figure 4. The turbidity of the sulfated CNCs, which is a reflection of their dispersion in solution, appears unchanged by the labeling reaction. On the other hand, the desulfated and HClhydrolyzed samples appear more stable after labeling, as evidenced by a decrease in opacity, albeit slight in the case of the HCl-hydrolyzed CNCs. The milky aspect of the desulfated CNC suspension (CNCdesulf.) is due to agglomeration caused by removal of about 60% of the stabilizing charged groups, and similarly, the high opacity of the HCl-hydrolyzed material (CNCHCl) is the result of the reassociation of CNCs due to an absence of electrostatic repulsion in the uncharged suspension. Visual inspection of samples at the same concentration suggested that the labeling did not have a negative impact upon the dispersion of CNCs in suspension. AFM, DLS, and electrophoretic mobility measurements were conducted to further explore the stability of the suspensions before and after labeling (Supporting Information, Table S1). Results indicate that stability depends upon the surface chemistry of the starting CNCs and the grafting density of DTAF. AFM allowed for 2D visualization of particle dispersion
surface chemistry of the CNCs, but in the current situation, where it is desired that the CNCs remain as “native-like” as possible, the apparently low chemical reactivity of CNCs is attractive. Despite the low degree of substitution, we could easily detect the fluorescence by spectroscopy (Figure 2), microscopy (Figure 3), and by eye under UV illumination (Figure 4, bottom panel). The low degree of substitution suggests that the labeled CNCs retained the properties and surface character of the unmodified particles. This is particularly relevant for applications where the fluorescent CNCs need to closely mimic unlabeled, native substrates; for example, Helbert et al.8 found that enzymatic activity of H. insolens was inhibited by highly labeled bacterial cellulose substrates. DTAF-labeled bacterial cellulose films have been proposed as useful substrates for highthroughput screening of cellulases because monitoring the fluorescence released as DTAF-labeled cellulose is degraded by cellulolytic enzymes would provide a sensitive measure of the total released sugars.8 We have not explored the use of fluorescent CNC films for this purpose, however, the use of DTAF-labeled CNCs for the study of biochemical saccharification may assist in understanding the enzyme interactions with an insoluble substrate that more closely resembles the recalcitrant fraction of the plant biomass used for biofuel production (i.e., highly crystalline cellulose I). E
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
was significantly less aggregated than CNCHCl. The electrophoretic mobility values (Supporting Information, Table S1) of the sulfated and partially desulfated samples were not significantly altered by the DTAF-labeling reaction, remaining constant at about 3 × 108 m2/(V s). The small decrease in electrophoretic mobility observed in the 1 wt % CNCDTAF sulf. sample may be related to the slight particle agglomeration seen by DLS (Supporting Information, Table S1). The CNCDTAF HCl sample was more negatively charged and had smaller particle sizes compared to the unlabeled CNCHCl. The increase in stability observed in CNCDTAF HCl may be due to the grafting of negatively charged DTAF moieties onto the uncharged substrate, although the degree of functionalization is quite low at 4.75 × 10−5 meq/g, offering a negligible change in the overall charge density and ionic strength of the dispersion. The self-assembly behavior of CNCs also sheds light on the stability of CNCs in suspension, since it seems that good dispersion is a requirement for particle alignment and liquid crystal organization. Polarized optical micrographs (Figure 6) of CNCDTAF and CNCDTAF sulf. desulf. showed the fingerprint texture characteristic of chiral nematic liquid crystals from stable dispersions of CNCs (details in Figure 6 are more easily visualized in the electronic version). CNCDTAF HCl did not form a chiral nematic structure but appeared more birefringent than the corresponding unlabeled sample and presented a crosshatch pattern that was not apparent in the unlabeled sample. A similar texture was observed by Araki et al.22 in postsulfated HClhydrolyzed CNC suspensions from Whatman CF11 which had a charge content of 0.038 meq/g compared to 4.75 × 10−5 DTAF meq/g for our CNCHCl . Araki et al.22 point out the resemblance of the crosshatch pattern to the “frozen in shear” structures observed in boehmite systems,23 which are thought to arise from particle immobilization due to long-range repulsive interactions. This explanation is not applicable to CNCDTAF HCl but perhaps the high viscosity of the samples (i.e., flow is hindered at concentrations greater than 1 wt %) coupled with the natural tendency of CNCs to align gives the crosshatch pattern. Another explanation or contributing factor may be that the grafted DTAF plays a role in the organization of CNCDTAF HCl , for example, the particle orientation may be guided by π−π stacking interactions from the relatively bulky substituent, although this is entirely speculative. In general, polarized optical microscopy showed that the CNCs prepared here self-assemble into ordered structures and that alignment in suspension for DTAF labeled CNCs was either similar or enhanced compared to the unlabeled samples. Finally, as proof of concept, DTAF-labeled CNCs were incorporated into electrospun PVA fibers to assess the quality of the dispersion of CNCs within a polymer composite. Several examples of CNC-loaded PVA films, 24−27 electrospun fibers,28,29 and gels30 have been described previously in the literature. In general, the combination of PVA and CNCs for the formation of composites exhibits good compatibility and uniformity and is thus useful as a model system to test the feasibility of employing fluorescently labeled CNCs as a marker for dispersion. The electrospun fibers were prepared using CNCDTAF suspended at 0.7 wt % in a 7 wt % PVA solution. sulf. These conditions were selected because they produced fibers which appeared mostly smooth and defect free by optical microscopy. Fluorescence microscopy images of PVA fibers loaded with fluorescent CNCs are presented in Figure 7 (top panel). The fluorescence of the fibers was integrated along the full length of the fibers indicating a good dispersion of CNCs
Figure 6. Polarized optical micrographs of unlabeled and DTAFlabeled CNC suspensions, where blue and yellow colors indicate the self-assembly of CNCs into ordered structures.
Figure 7. Top panel shows epifluorescence microscopy images of electrospun PVA fibers loaded with DTAF-labeled CNCs (CNCDTAF sulf. ). Bottom panel shows bright field image (left) and corresponding epifluorescence image (right) of PVA fibers loaded with unlabeled CNCs (CNCsulf.).
in the dry form, DLS provided an apparent particle “size” in suspension which was useful for comparison between samples, and electrophoretic mobility gave an indication of the sign and magnitude of the surface charge. By AFM (Figure 5), DTAF CNCDTAF sulf. , and CNCdesulf. presented morphologies similar to their unreacted counterparts, in contrast to CNCDTAF HCl , which F
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
■
within the PVA matrix, with no large bright spots present. Minor intensity variations may be due to fluctuations in the dimensions of the electrospun fibers which were not investigated in detail. As a control, PVA fibers were loaded with nonfluorescent CNCs and visualized by bright field and epifluorescence microscopy (Figure 7, bottom panel). The bright field image (Figure 7c) shows a high density of electrospun fibers, whereas the corresponding epifluorescence image (Figure 7d) shows only dim fluorescence from the background. Epifluorescence images of neat PVA fibers were similarly unremarkable. Thus, the fluorescence of the DTAFlabeled CNC loaded PVA fibers is from the CNCs contained within the fibers, not from autofluorescence. Addressing the dispersion of CNCs in a polymer matrix is nontrivial and often requires careful consideration of bulk properties and indirect electron microscopy techniques. The method described here provides a straightforward visual marker to assess the compatibility of CNCs with a bulk polymeric matrix, which can be performed without altering the natural state of the composite.
CONCLUSIONS DTAF was grafted onto cotton-derived CNCs with different surface charge densities in a one-step, water-based labeling reaction. The labeling efficiency was in the nmol/g range, and was found to depend upon the charge content of the starting CNCs, with the amount of bound DTAF increasing with decreasing surface charge density. The properties of the fluorescently labeled CNCs were nearly identical to the unlabeled materials, an attractive feature for studies that require fluorescent CNC substrates that retain as many of the native characteristics as possible. Finally, the DTAF-labeled CNCs were used to study the uniformity of CNC dispersion in electrospun PVA fibers. The uniformity of the fluorescence along the lengths of the fibers was taken as an indication of good dispersion of CNCs within the PVA matrix. The DTAFlabeled CNCs prepared in this work possess characteristics that make them suitable as visual markers for biotoxicity, cellulase− cellulose interaction and dispersion studies. ASSOCIATED CONTENT
S Supporting Information *
Conductometric titration results, electrophoretic mobility, dynamic light scattering, and particle size analysis by atomic force microscopy. This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
(1) Habibi, Y.; Lucia, L. A.; Rojas, O. J. Chem. Rev. 2010, 110, 3479− 3500. (2) Cranston, E. D.; Gray, D. G.; Rutland, M. W. Langmuir 2010, 26, 17190−17197. (3) Wang, H.; Qian, C.; Roman, M. Biomacromolecules 2011, 12, 3708−3714. (4) Gray, D. G. Estimation of the surface sulfur content of cellulose nanocrystals prepared by sulfuric acid hydrolysis. 245th ACS National Meeting and Exposition, New Orleans, LA, April 7−11, 2013, American Chemical Society: Washington, DC, 2013. (5) Kan, K. H. M.; Li, J.; Wijesekera, K.; Cranston, E. D. Biomacromolecules 2013, DOI: 10.1021/bm400752k. (6) Jiang, F.; Kittle, J. D.; Tan, X.; Esker, A. R.; Roman, M. Langmuir 2013, 29, 3280−3291. (7) Pan, X.; Gilkes, N.; Saddler, J. N. Holzforschung 2006, 60, 398− 401. (8) Helbert, W.; Chanzy, H.; Husum, T. L.; Schülein, M.; Ernst, S. Biomacromolecules 2003, 4, 481−487. (9) Moran-Mirabal, J. M.; Santhanam, N.; Corgie, S. C.; Craighead, H. G.; Walker, L. P. Biotechnol. Bioeng. 2008, 101, 1129−1141. (10) Luterbacher, J. S.; Walker, L. P.; Moran-Mirabal, J. M. Biotechnol. Bioeng. 2013, 110, 108−117. (11) Dong, S.; Roman, M. J. Am. Chem. Soc. 2007, 129, 13810− 13811. (12) Nielsen, L. J.; Eyley, S.; Thielemans, W.; Aylott, J. W. Chem. Commun. 2010, 46, 8929−8931. (13) Mahmoud, K. A.; Mena, J. A.; Male, K. B.; Hrapovic, S.; Kamen, A.; Luong, J. H. T. ACS Appl. Mater. Interfaces 2010, 2, 2924−2932. (14) Zhang, L.; Li, Q.; Zhou, J.; Zhang, L. Macromol. Chem. Phys. 2012, 213, 1612−1617. (15) Beck-Candanedo, S.; Roman, M.; Gray, D. G. Biomacromolecules 2005, 6, 1048−1054. (16) Jiang, F.; Esker, A. R.; Roman, M. Langmuir 2010, 26, 17919− 17925. (17) Araki, J.; Wada, M.; Kuga, S. Langmuir 2000, 17, 21−27. (18) Dong, X. M.; Revol, J. F.; Gray, D. G. Cellulose 1998, 5, 19−32. (19) Jiang, F.; Esker, A. R.; Roman, M. Langmuir 2010, 26, 17919− 17925. (20) Wada, M.; Chanzy, H.; Nishiyama, Y.; Langan, P. Macromolecules 2004, 37, 8548−8555. (21) Igarashi, K.; Wada, M.; Samejima, M. FEBS J. 2007, 274, 1785− 1792. (22) Araki, J.; Wada, M.; Kuga, S.; Okano, T. Langmuir 2000, 16, 2413−2415. (23) Buining, P. A.; Philipse, A. P.; Lekkerkerker, H. N. W. Langmuir 1994, 10, 2106−2114. (24) Roohani, M.; Habibi, Y.; Belgacem, N. M.; Ebrahim, G.; Karimi, A. N.; Dufresne, A. Eur. Polym. J. 2008, 44, 2489−2498. (25) Paralikar, S. A.; Simonsen, J.; Lombardi, J. J. Membr. Sci. 2008, 320, 248−258. (26) Fortunati, E.; Puglia, D.; Monti, M.; Santulli, C.; Maniruzzaman, M.; Kenny, J. M. J. Appl. Polym. Sci. 2013, 128, 3220−3230. (27) Pakzad, A.; Simonsen, J.; Yassar, R. S. Nanotechnology 2012, 23, 085706. (28) Peresin, M. S.; Habibi, Y.; Zoppe, J. O.; Pawlak, J. J.; Rojas, O. J. Biomacromolecules 2010, 11, 674−681. (29) Ago, M.; Okajima, K.; Jakes, J. E.; Park, S.; Rojas, O. J. Biomacromolecules 2012, 13, 918−926. (30) Abitbol, T.; Johnstone, T.; Quinn, T. M.; Gray, D. G. Soft Matter 2011, 7, 2373−2379.
■
■
Article
AUTHOR INFORMATION
Corresponding Author
*E-mail: tiff
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors gratefully acknowledge Professors Gray, Guarne, Pelton, Stöver, and Brennan for access to instrumentation and X. Yang for preparing the batch of H2SO4-hydrolyzed CNCs used in this work. E. Luckham and E. Forsberg (McMaster University’s Biointerfaces Institute) are thanked for microscope and plate reader training, respectively, F. Naeem is appreciated for providing the tip used in the electrospinning experiments, and K. Kan is recognized for useful discussions. G
dx.doi.org/10.1021/bm400879x | Biomacromolecules XXXX, XXX, XXX−XXX
