Article pubs.acs.org/Biomac
Amphiphilic Cellulose Nanocrystals from Acid-Free Oxidative Treatment: Physicochemical Characteristics and Use as an Oil−Water Stabilizer Miikka Visanko,† Henrikki Liimatainen,*,‡ Juho Antti Sirviö,‡ Juha Pentti Heiskanen,§ Jouko Niinimak̈ i,‡ and Osmo Hormi§ †
Fibre and Particle Engineering Laboratory and Thule Institute, ‡Fibre and Particle Engineering Laboratory, and §Department of Chemistry, University of Oulu, P.O. Box 4300, FI-90014, Finland S Supporting Information *
ABSTRACT: A chemical pretreatment for producing cellulose nanocrystals (CNCs) with periodate oxidation and reductive amination is reported. This new functionalization of cellulose fibers dispenses an alternative method for fabricating individual CNCs without the widely used acid hydrolysis process. CNCs can be directly modified during the pretreatment step, and no additional post-treatments are required to tune the surface properties. Three butylamine isomers were tested to fabricate CNCs with amphiphilic features. After mechanical homogenization, CNCs occurred as individual crystallinities without aggregation where high uniformity in terms of shape and size was obtained. The elemental analysis and 1H NMR measurement show that isoand n-butylamine attach the highest number of butylamino groups to the cellulose fibers. Linking the alkyl groups increases the hydrophobic nature of the CNCs, where water contact angles from self-standing films up to 110.5° are reported. Since these butylamino-functionalized CNCs have hydrophobic characteristics in addition to the hydrophilic backbone of cellulose, the stabilization impact on oil/water emulsions is demonstrated as a potential application.
■
INTRODUCTION
Regioselective oxidative techniques such as 2,2,6,6-tetramethylpiperidinyl-1-oxyl (TEMPO)5 and periodate oxidations have been used as efficient pretreatments for fabricating cellulose nanofibrils, which are elongated nanosized cellulose strands containing amorphous and crystalline parts. We have previously reported that periodate oxidation, which oxidizes the vicinal hydroxyl groups of cellulose at positions 2 and 3 to aldehyde groups, can be combined with a second-stage reaction to introduce carboxylic6 and sulfonic acid7 or quaternized ammonium groups8 to attain different properties (e.g., cationic/ anionic surface charge) of cellulose. These pretreatments significantly enhance the nanofibrillation of cellulose fibers since only mild mechanical treatments are required to separate individual nanofibrils.
Cellulose nanocrystals (CNCs), originating from elementary fibrils, are nanosized constituents of natural cellulosic structures, which have attracted much interest in recent years since these nanocrystals have unique features such as lightweight, stiffness, high mechanical reinforcing capability, e.g., in polymer nanocomposites,1 renewability, and abundant availability ranging from tunicins2 to higher plants.3 CNCs exist in different morphologies in terms of the length, width, aspect ratio (L/d about 10−100), and shape, which can vary from rodlike to spherical,4 depending on the origin and processing of the cellulosic materials. Typically, CNCs are fabricated using acid hydrolysis processes in which strong inorganic acids such as hydrochloric acid or, most often, sulfuric acid have been applied. The strong acids dissolve the more susceptible amorphous regions of cellulose while leaving the crystal structure intact. As a result, individualized nanoscale crystallinities can be isolated from the treated pulp suspension. © XXXX American Chemical Society
Received: April 29, 2014 Revised: June 13, 2014
A
dx.doi.org/10.1021/bm500628g | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Figure 1. Periodate oxidation and reductive amination of cellulose.
tion of hydrophobic alkyl groups via our acid-free preparation method, the CNCs produced in the present study were tested as potential stabilizers for o/w emulsions. Furthermore, the morphological and amphiphilic characteristics of the aminofunctionalized CNCs were measured and analyzed using elemental analysis, transmission electron microscopy (TEM), X-ray diffraction (XRD), 1H NMR, and static sessile-drop contact-angle measurement. In addition, the optical transmittance at visible light wavelengths and the viscosity of the aqueous CNC suspensions were analyzed.
Here, instead of fabricating cellulose nanofibrils, an acid-free chemical pretreatment based on subsequent regioselective oxidative and reductive amination to produce amphiphilic CNCs is presented. This methodology can also potentially be used to introduce various other surface functionalities to CNCs, as already demonstrated with cellulose nanofibrils. Thus far, only a few other acid-free treatments of cellulosic material have been studied to fabricate CNCs. Highly charged dicarboxylic acid cellulose (DCC) has been produced by excessively long oxidation times where individual CNCs and cellulose nanofibrils were found from the oxidized cellulose pulp when different fractions were separated with centrifugation.9 Due to the highly charged DCC chains, the fabricated CNCs are referred to as electrosterically stabilized nanocrystalline cellulose (ENCC). Ultrasonic-assisted TEMPO oxidation10 and ammonium persulfate (APS) oxidation11,12 are the only other reported studies of acid-free methodologies. Thus, based on our best knowledge, consequent periodate oxidation and reductive amination or another method has not been used to prepare amphiphilic CNCs directly. The regioselective oxidative chemical pretreatment of cellulose material has many useful advantages compared to the acid hydrolysis processes commonly used to individualize CNCs.13,14 Because the targeted surface functionality can be introduced before the crystallinities are liberated, no additional post-treatments such as grafting,15 surfactant adsorption,16 cationization,17 or hydrophobization18 are required for CNC suspensions, which are difficult to treat afterward. Consequently, the washing of the reactive chemicals after the pretreatment is straightforward, and complicated separation processes such as dialysis against distilled water or centrifugation are not required.19 Aggregation of the CNCs, a common problem when hydrolysis with hydrochloric acid20 is used, can be prevented since the surface charge of the CNCs associated, for example, with the amine side chain stabilize the CNC suspension. Solid colloidal particles with affinity for both phases of an emulsion can be applied to stabilize the so-called Pickering emulsions.21,22 Thus, CNCs with amphiphilic nature have been used as effective stabilizers with oil/water (o/w) emulsions23−25 in which the CNCs adsorb irreversibly to the o/w interface. Therefore, CNCs offer an auspicious alternative for common synthetic surfactants since CNCs can be classed as nontoxic, sustainable, green products.26 Due to the introduc-
■
EXPERIMENTAL SECTION
Materials. Bleached kraft hardwood (Betula pendula) pulp was used as a cellulose raw material for the preparation of CNCs. NaIO4 (India, purity ≥ 99.0%) and LiCl (Germany, ≥98.0%) were procured from Sigma-Aldrich to fabricate dialdehyde cellulose. For the reductive amination of dialdehyde cellulose, 2-picoline borane (Sigma-Aldrich, U.S.A. (95%)), and three butylamine isomers, isobutylamine hydrochloride (Tokyo Chemical Industry, Belgium, >99%), n-butylamine hydrochloride (Tokyo Chemical Industry, Belgium, >98%), and tertbutylamine hydrochloride (Sigma-Aldrich, Switzerland, ≥98.0%) were purchased. These chemicals were used as received without any further purification. Ethanol (96%) to wash oxidized pulp was bought from VWR (Finland). Methods. Preparation of Amphiphilic CNCs. Bleached chemical wood pulp was first converted to DAC with lithium chloride (LiCl)assisted sodium metaperiodate (NaIO4) oxidation, as previously reported.27 Oxidation was conducted at 75 °C for 3 h to attain DAC with 3.79 mmol g−1 of aldehydes. Three amphiphilic celluloses were manufactured from DAC using reductive amination with butylamine isomers (Figure 1).28 First, a 10-fold excess of iso-, n-, or tert-butylamine hydrochloride in relation to the aldehyde groups of DAC was mixed with deionized water (900 mL), and the pH of the solution was set to 4.5 with a dilute HCl. DAC fibers (9 g) and a 2-fold excess of 2-picoline borane, based on the assessed amount of the aldehyde groups, were added to the suspension, and the reaction was continued for 72 h under magnetic stirring in a closed container at room temperature. The reaction was stopped by removing the reactive chemicals from the solution with vacuum filtration. The product was washed first with ethanol (500 mL) and then with water (1000 mL). Washed pulp was diluted (0.3% w w−1) and dispersed with an Ultra-Turrax mixer (IKA T25, Germany) at 10000 rpm rotational speed for 1 min. The CNCs were individualized from the amphiphilic cellulose fibers using a twochamber high-pressure homogenizer (APV-2000, Denmark). Three passes, with pressures of 220, 480, and 600 bar, respectively, resulted in clear nonviscous CNC suspensions. The amino group content of B
dx.doi.org/10.1021/bm500628g | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Table 1. Observed 1H NMR Signals and Integrals of Butylamino-Functionalized CNCs
a
Based on observed signal integrals. bElemental analysis. Water Contact Angles of Self-Standing CNC Films. The hydrophobicity of the fabricated CNCs was evaluated by applying a static sessile-drop contact-angle measurement where Milli-Q water was used as a probe liquid at room temperature. For the measurement, selfstanding films were prepared from diluted CNC suspensions. The suspensions were dispersed using an Ultra-Turrax mixer (IKA T25, Germany) at 10000 rpm rotational speed for 3 min after which the suspensions were vacuum filtrated on a glass filter funnel (7.2 cm in diameter) using a filter membrane (polyvinylidene fluoride, 0.65 μm, Millipore Durapore, France). The measurements were carried out with a Krüss DSA100 (Germany) system. The instrument was equipped with a high-speed camera (360 fps) and analysis software. The contact angle was determined immediately after the drop formed on the film surface. A drop volume of less than 2 μm was used to avoid distorting effects on the droplet due to gravity.29 Contact angles were extracted with the height-width method, in which a rectangle enclosed by a contour line was regarded as the segment of a circle. As a result, contact angles were calculated from the height-width relationship of the enclosing rectangle. For each sample, three droplets on different locations were studied, the results were averaged, and the standard deviations were calculated. Optical Transmittance of the CNC Suspensions. Transmittance of dilute CNC suspensions (0.1% w w−1) was measured with a Hach DR 2800 spectrophotometer (U.S.A.). Measurements were conducted as duplicates at wavelengths ranging from 350 to 800 nm, and the results were averaged. Viscosity of the CNC Suspensions. The rheological demeanor of the CNCs was studied in varying consistencies using a low-shear viscometer. Depending on the sample, viscosities were measured in various consistencies (0.6−2.9% w w−1). The measurements were conducted with a Brookfield DV-II+ Pro EXTRA viscometer (U.S.A.) at a temperature of 20 °C using a vane-shaped spindle (V-73) and rotational speeds of 10, 20, 50, and 100 rpm. Butylamino-Functionalized CNCs as Stabilizers in o/w Emulsion. The performance of amphiphilic CNCs as stabilizers in o/w emulsions was investigated using a soybean oil−water emulsion. Iso-, n-, or tertbutylamino-functionalized CNCs (50 g, 0.2% w w−1) was mixed with water (40 g), and soybean oil (10 g) was added by mixing with an Ultra-Turrax mixer at 10000 rpm rotational speed for 1 h. The reference o/w emulsion was fabricated by mixing deionized water (90 g) and soybean oil (10 g). The stability of the CNC treated o/w emulsions was evaluated with a laser diffraction particle size analyzer (LS 13 320, Beckman Coulter, U.S.A.) by measuring the average particle size of the oil droplets. Three replicates were measured from each sample. The long-term stability of the o/w emulsions was measured against physical changes
the CNCs was determined using an elemental analyzer (Thermo Scientific FLASH 2000 Series CHNS/O Analyzer, U.S.A.) by measuring the nitrogen content of the dried sample. All the samples were analyzed without different fractions separated from the suspensions. 1 H NMR. Samples were dissolved in DMSO-d6 and placed in NMR tubes (5 mm). The 1H NMR spectra of butylamino-functionalized CNCs were measured by using the Bruker DPX 200 MHz NMR instrument (U.S.A.). Integration value 2 was given to the area from 4.50 to 4.90 ppm that remained static during functionalization. The saturated alkyl region around 0.90−1.40 ppm was analyzed and integrated as shown in the Supporting Information (Figures S1−3). Transmission Electron Microscope (TEM). The morphological features of the fabricated CNCs were analyzed with a Tecnai G2 Spirit transmission electron microscope (FEI Europe, Eindhoven, The Netherlands). Samples were prepared by diluting each CNC sample with Milli-Q water. A small droplet of the dilution was dosed on top of a carbon-coated and glow-discharged copper grid. Excess sample was removed from the grid by touching the droplet with the corner of a filter paper. The samples were negatively stained by placing a droplet of uranyl acetate (2% w v−1) on top of each specimen. The excess amount of the uranyl acetate was removed with filter paper as described. Grids were dried in room temperature and analyzed at 100 kV under standard conditions. Images were captured with a Quemesa CCD camera, and iTEM image analysis software (Olympus Soft Imaging Solutions GMBH, Munster, Germany) was used to measure individual CNC width and length. A total of 50 crystals from each CNC grade were measured. The final results were averaged, and the standard deviations were calculated. X-ray Diffraction. The crystallinity of the butylamino-functionalized CNCs was analyzed using wide-angle X-ray diffraction (WAXD). Measurements were conducted on a Siemens D5000 diffractometer (Germany) equipped with a Cu Kα radiation source (λ = 0.1542 nm). Specimens were prepared from the homogenized CNC suspensions by pressing tablets with a thickness of 1 mm after freeze-drying the samples. Scans were taken over a 2θ (Bragg angle) range from 5 to 50° at a scanning speed of 0.1° s−1 using a step time of 5 s. The degree of the peak intensity of the main crystalline plane (002) diffraction (I002) was located at 22.5°. The peak intensity associated with the amorphous fraction of cellulose (Iam) was observed at 18.0°. The degree of crystallinity in terms of the crystallinity index (CrI) was calculated according to the following eq 1.
CrI =
I002 − Iam × 100% I002
(1) C
dx.doi.org/10.1021/bm500628g | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
with an analytical centrifuge (LUMiFuge, L.U.M. GmbH, Germany). The centrifuge consists of a light source, a rotor above which the sample cells containing the suspension were horizontally positioned, and a CCD linesensor below the rotor. The centrifuge simultaneously measures light transmission at 800 nm over the plastic sample cells as a function of time and position. Local alterations in concentration and formation of oil−water interface during centrifugation are detected by changes in light transmission.30 Emulsion stability is indicated by lower residual transmission as the oil does not separate to its own phase from water. From each CNC-treated o/w sample and from the reference emulsion, two parallel samples were prepared and centrifuged for 1 h at 1200 rpm rotational speed (equals a relative centrifugal force of approximately 205 G in the bottom of the sample holder) at 20 °C.
■
RESULTS AND DISCUSSION
Amino Group Content of Amphiphilic CNCs. Amphiphilic CNCs were fabricated from dialdehyde cellulose (DAC) fibers with reductive amination where the reactive aldehyde groups were converted to amino groups with iso-, n-, or tertbutylamine hydrochloride (Figure 1). From the 1H NMR spectra characteristic proton signals of butylamino groups were resolved and integrated (see Supporting Information, Figures S1−3). 1H NMR analysis shows that the contents of the nbutylamino and isobutylamino groups are almost equal as they are approximately eight times higher than the content of the tert-butylamino group (Table 1). Based on the elemental analysis, the substituent content of the CNC samples after reacting with butylamine isomers was 0.567, 0.562, and 0.072 mmol g−1 for iso-, n-, and tertbutylamine, respectively. These values correspond to a conversion percentage of 15.0, 14.8, and 1.9% since the aldehyde content of the DAC fibers was 3.79 mmol g−1. This observation is consistent with the results of the 1H NMR analysis. Overall, the reactivity with the aldehyde groups was moderate with iso- and n-butylamine whereas tert-butylamine reached a very low conversion rate. The lower reactivity with tert-butylamine might be linked to sterical effects since the isoand n-butylamino groups are sterically less demanding (Figure 1). The reaction yield (relative to original cellulose) after reductive amination was 50% for iso- and n-butylamino and 45% for tert-butylamino-functionalized CNCs. This value is clearly higher compared to acid hydrolysis where yields around 30% have been reported.31−33 Crystallinity. X-ray diffraction patterns measured from the butylamino-functionalized CNCs are presented in Figure 2. The diffraction patterns show the characteristic peaks for cellulose I indicating no rearrangement occurred to other allomorphs of cellulose. The crystallinity indexes calculated according to eq 1 for iso-, n-, and tert-butylamino-functionalized CNCs were 52.7, 56.4, and 57.0%, respectively. Partial reduction in the crystallinity is suspected to be caused by the periodate oxidation where opening of the glucopyranose rings and disruption of their ordered packing occur.34 Destruction of the ordered packaging and conversion to aldehyde groups facilitate the oxidation of neighboring groups next to the opened glucopyranose rings causing the oxidation to proceed unevenly alongside the crystals.34 Morphology and Visual Appearance of the Aminated CNCs. The morphology of the aminated CNCs was portrayed with TEM, and the captured images were analyzed with iTEM image analysis software to evaluate the average width, length, and aspect ratio. According to the TEM images the CNCs appear as rod-like individualized crystallinities (Figure 3) instead of larger clusters having a more uniform shape and
Figure 2. X-ray diffraction patterns of the original cellulose pulp, DAC fibers, and butylamino-functionalized CNCs.
Figure 3. TEM-image of n-butylamino-functionalized CNCs.
size than those from acid hydrolysis. 35,36 Butylamino functionalization of CNCs is assumed to be linked to the prevention of crystal coalescence observed in acid hydrolysis.20 The width and length (Table 2) of a single CNC correspond Table 2. Average Dimensions for the ButylaminoFunctionalized CNCs Based on the iTEM Image Analysis sample
width (nm)
length (nm)
L/d
isobutylamine n-butylamine tert-butylamine
3.65 ± 0.82 3.58 ± 0.66 3.44 ± 0.56
135.42 ± 31.96 115.11 ± 25.94 128.69 ± 31.35
37.15 32.17 37.42
the dimension reported for wood cellulose microfibrils in the crystal region37 and corroborates the scheme proposed in an earlier study.34 The crystal dimensions were not dependent on the reductive amination treatment since all the samples had similar morphology. Optical transmittance also confirms the presence of nanosized particles as high values were achieved from dilute (0.1% w w−1) CNC suspensions at visible light wavelengths ranging from 350 to 800 nm (Figure 4). Based on the visual appearance D
dx.doi.org/10.1021/bm500628g | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
In addition to the reductive amination, the mechanical stress produced by magnetic stirring and high-impact homogenizer assist the disintegration of the cellulose fibers and the individualization of the CNCs. The continuous magnetic stirring during the reductive amination likely slowly starts to degrade and dissolve the cellulose chains especially from the weakened amorphous regions. In a previous study, up to 97% nanofibril conversion was achieved for TEMPO-oxidized fibers when the pulp suspension was magnetically stirred for 10 days.40 The mechanical nanofibrillation of the pulp suspension is known to separate individual nanofibrils preferentially in the longitudinal direction and cause some cutting in the crosssection. With the butylamino-functionalized samples, the majority of the remaining amorphous regions may be torn away leaving only strong CNCs intact. However, additional testing is required to understand how the CNCs are actually liberated. Hydrophobicity of Self-Standing CNC Films. Static sessile-drop contact angle measurement was used to measure the contact angles from the self-standing films to indicate the surface hydrophobicity of the CNCs caused by the amination. Based on three measurements from different film locations, the following results were obtained: 110.5 ± 2.2°, 72.2 ± 0.9°, and 66.9 ± 4.5° for iso-, n-, and tert-butylamine, respectively. Since the number of attached alkyl groups is assumed to influence the hydrophobicity of the CNC surface, the results support the elemental analysis because isobutylamino-functionalized CNCs with the highest substituent content had the largest contact angle. In contrast, tert-butylamino-functionalized CNCs, which had the lowest substituent content, had the least hydrophobic surface probably because of the lack of the attached alkyl groups. Rheological Behavior of CNC Suspension in Elevated Consistencies. Rheological properties of the CNCs were measured with a rotational viscometer. At a consistency of 0.6% (w w−1), typical shear thinning behavior was observed, as Figure 5a shows. The impact of the increasing dry matter content in the viscosity behavior of the CNCs is presented in Figure 5b,c with low rotational speeds of 10 and 20 rpm. Tertbutylamino-functionalized CNCs had the lowest viscosity throughout the entire measurement range. The results agree with the transmittance measurements since the n-butylaminofunctionalized product was assumed to have the largest fraction of formed CNCs. The uppermost viscosity values with nbutylamino-functionalized CNCs support this finding since the tert-butylamino-functionalized product was assumed have the smallest fraction of formed CNCs. Butylamino-Functionalized CNCs as Stabilizers in o/w Emulsion. Introducing amino groups onto the CNC surface gives them potential affinity for two different phases as the butylamino groups possess hydrophobic characteristics as demonstrated with the contact angle measurement, while the cellulosic backbone remains hydrophilic. Therefore, the butylamino-functionalized CNCs were tentatively tested as potential stabilizers in o/w emulsions. The stabilizing effect of the CNCs was evaluated with a laser diffraction particle size analyzer after 1 h of mixing with Ultra-Turrax. The averaged results are shown in Figure 6, and more detailed statistics are provided in Table 3. As shown in Figure 6, the iso- and n-butylamino-functionalized CNCs stabilize the o/w emulsions most efficiently by decreasing the average particle size from 187.60 μm to around 12−14 μm. The results suggest that the oil droplets remain as
Figure 4. Optical transmittance of (0.1% w w−1) amphiphilic CNC suspensions.
and the mild substituent content of tert-butylamino groups, it reached the lowest transmission values as expected. Iso- and nbutylamino-functionalized CNCs had clearly higher transmission in the studied range; however, the results suggest still larger fractions were present since the transmission was not yet above 90%. The mechanism behind individualization of CNCs is assumed to be a result of multiple factors since the periodate oxidation begins to shorten the aspect ratio of the cellulose fibers.38 The conversion of the hydroxyl groups to the aldehyde groups in the vicinal positions of 2 and 3 is suspected to occur preferentially in the disordered and more reactive amorphous parts and less profusely in the crystalline regions.39 The amorphous regions can be hypothesized to be exposed to more strenuous chemical attack causing it to initially dissolve from the surface of the cellulose fibrils. While the surface dissolves, the chemicals react further toward the core of the amorphous region making it frailer. The impact of periodate oxidation on the degradation of cellulose fibers was demonstrated by preparing pure DAC fibers without the reductive amination reaction. Once the periodate oxidation was completed, the DAC fibers were mixed for 72 h in water solution (1% w w−1) to include the mechanical stress from the magnetic stirring. After stirring, the DAC suspension was filtrated and diluted (0.3% w w−1). A mechanical homogenizer was used to nanofibrillate the DAC fibers with the same bypass pressures as the butylamino-functionalized specimen. The periodate oxidation has an influence on the morphology of the cellulose since the DAC fibers were successfully passed through the homogenizer without any clogs. The homogenized suspension remained nonviscous and turbid, and shortly after the homogenization, settling was observed when the majority of the fibers were seen to separate to the bottom of the vessel instead of remaining as a steady suspension. This clearly indicates that the periodate oxidation was not sufficient enough to liberate the CNCs without the derivatization of the aldehyde groups with amines to reduce cohesion within the cellulose structure. Consequently, the reductive amination is the key element for finalizing the chemical modification of the cellulose fibers and to eventually liberate individual CNCs by predisposing the fibers under mechanical stress. E
dx.doi.org/10.1021/bm500628g | Biomacromolecules XXXX, XXX, XXX−XXX
Biomacromolecules
Article
Figure 6. Average particle size vs total volume of the particle in the sample for butylamino-functionalized CNC treated o/w emulsions and the reference sample.
Table 3. Volume Statistics Based on the Results from the Laser Diffraction Particle Size Analyzer Measurements sample
mean (μm)
SD (μm)
reference isobutylamine n-butylamine tert-butylamine
187.60 14.21 12.67 26.46
±113.50 ±8.46 ±7.82 ±17.33
