Abstract
Cellulose nanocrystals (CNCs) were modified with methyl methacrylate (MMA) to improve the properties of the resulting three-dimensional (3D) stereolithography printed CNC/methacrylate (MA) resin composites. The dispersibility of the MMA-modified CNCs (MMA-CNCs) was substantially improved, as evidenced by the limited precipitation in the MA solution. Thermal gravimetry and differential scanning calorimetry measurements showed that the pyrolytic temperature of the MMA-CNC was 110 °C higher than that of the CNCs; the pyrolytic temperature and glass transition temperature of the resulting MMA-CNC/MA composites were higher than those of the CNC/MA. The tensile strength and modulus of the MMA-CNC/MA composites were improved by up to 38.3 MPa and 3.07 GPa, respectively, compared to those of the CNC/MA composites. These results demonstrated that the modification of CNC with MMA is a feasible approach to substantially improve the mechanical properties and thermal stability of the resulting MA-based composites.
Download PDF
Full Article
Reinforcing 3D Print Methacrylate Resin/Cellulose Nanocrystal Composites: Effect of Cellulose Nanocrystal Modification
Xinhao Feng,a,b Zhihui Wu,a,* Yanjun Xie,c and Siqun Wang b,*
Cellulose nanocrystals (CNCs) were modified with methyl methacrylate (MMA) to improve the properties of the resulting three-dimensional (3D) stereolithography printed CNC/methacrylate (MA) resin composites. The dispersibility of the MMA-modified CNCs (MMA-CNCs) was substantially improved, as evidenced by the limited precipitation in the MA solution. Thermal gravimetry and differential scanning calorimetry measurements showed that the pyrolytic temperature of the MMA-CNC was 110 °C higher than that of the CNCs; the pyrolytic temperature and glass transition temperature of the resulting MMA-CNC/MA composites were higher than those of the CNC/MA. The tensile strength and modulus of the MMA-CNC/MA composites were improved by up to 38.3 MPa and 3.07 GPa, respectively, compared to those of the CNC/MA composites. These results demonstrated that the modification of CNC with MMA is a feasible approach to substantially improve the mechanical properties and thermal stability of the resulting MA-based composites.
Keywords: Methyl methacrylate; Cellulose nanocrystal; Grafting; Stereolithography; Thermal performances
Contact information: a: College of Furnishings and Industrial Design, Nanjing Forestry University, Nanjing, Jiangsu 210037, P. R. China; b: Center for Renewable Carbon, University of Tennessee, Knoxville, Tennessee 37996, USA; c: College of Material Science and Engineering, Northeast Forestry University, Harbin, Heilongjiang 150040, P. R. China;
* Corresponding authors: wzh550@sina.com; swang@utk.edu
INTRODUCTION
Three-dimensional (3D) stereolithography (3D-SL) printing is used to produce layer-laminated structures through curing liquid photopolymer resin under light irradiation (Becker and Gärtner 2008; Schaedler and Carter 2016). The 3D-SL printing process has shown higher resolution and accuracy for materials than the traditional processing techniques such as extrusion and compression molding (Cooke et al. 2003; Weng et al. 2016). Therefore, 3D-SL printing has been extensively applied in the areas of electronics (Zhang et al. 1999; Lee et al. 2008), tissue engineering (Melchels et al. 2010; Suntornnond et al. 2017), and polymer engineering (Nguyen et al. 2005; Mao et al. 2018). Because of the specific requirements, only limited types of resins, such as acrylate, epoxy, and polyurethane (Melchels et al. 2010), are commercially available for the 3D-SL processing. Materials containing these resins as a base often exhibit undesirable embrittlement caused by the formation of a highly crosslinked network (Lin et al. 2015), and the preferable properties are time dependent in the epoxy- and acrylate-resin matrix after printing (Crivello and Reichmanis 2014; Yang et al. 2018). In addition, polymerization of the resin is uncontrollable and incomplete during the printing and post-curing (at low temperature) processes (Lin et al. 2007; Melchels et al. 2010; Qin et al. 2010; Yang et al. 2018), which results in insufficient mechanical strength (Moon et al. 2014; Lin et al. 2015), dimensional instability (Esposito Corcione et al. 2004; Xu and Chen 2015), and low flexibility (Vatani et al. 2015).
Reinforcements, such as cellulose nanocrystal (CNC), have been used to overcome the problems of 3D-SL-printed materials because of its high specific strength, large aspect ratio, and renewability (Kedzior et al. 2016; Yang et al. 2018). Various types of cellulose, such as fibers, CNC, and bacterial nanocellulose, have exhibited potential to transfer stresses between deformed matrices and improve the properties of 3D-printed composites (Kramer et al. 2006; Farahani et al. 2012; Chiappone et al. 2014; Kim et al. 2014). They have been applied to improve the mechanical properties of methacrylate-based photoreactive resins (Shanmuganathan et al. 2009; Singha and Thakur 2009; Feng et al. 2017). However, the mixture of CNC and resin must be continuously stirred before sample fabrication due to the lack of adequate compatibility of hydrophilic cellulose in the hydrophobic resin matrix. As a result, aggregation of cellulose in the matrix and phase separation were found both in the mixture of CNC and resin before printing and in the printed nanocomposites, resulting in insufficient interfacial interaction and weak reinforcement efficacy of CNC (Singha and Thakur 2009; Feng et al. 2017; Feng et al. 2018; Yang et al. 2018).
To solve this issue, a feasible strategy is to obtain a hydrophobic characteristic of CNC through surface modification. The best modifying agent should be both reactive with CNC and compatible to the methacrylate (MA) resin matrix. The modification of cellulose with hydrophobic monomers, such as methyl methacrylate (MMA), has been shown to improve the compatibility between cellulose and the photopolymer resin (Kedzior et al. 2016). Cotton cellulose, bamboo cellulose, and nanofibrillated cellulose (NFC) were modified by the acrylic monomers in an aqueous solution via free radical reaction, and the resulting cellulosic materials were found to be more hydrophobic and thermally stable compared to the unmodified ones, and grafting with acrylic monomers resulted in no changes to the microstructure of cellulose materials (Littunen et al. 2011; Wan et al. 2011; Routray and Tosh 2012). The modified NFC and CNC were used as a reinforcement particles within a poly(methyl methacrylate) (PMMA) matrix to fabricate nanocomposites, and the interfacial adhesion between NFC/CNC and PMMA was improved (Littunen et al. 2013; Kedzior et al. 2016). However, the mechanical properties and thermal stability of the resulting composites decreased due to the aggregation of the MMA grafted NFC/CNC in the PMMA matrix (Littunen et al. 2013).
The authors have previously used CNCs to reinforce 3D-printed methacrylate composites and found that the addition of CNCs below 0.5% did not influence the tensile properties; incorporating higher CNC amounts resulted in a decrease in the tensile strength (8%) and elongation (55%) (Feng et al. 2017). The decrease is mainly attributable to the uneven dispersion of CNCs in the MA matrix and the failure of interfacial adhesion between the CNC and MA matrix due to the different polarity of the CNC and MA molecule. Therefore, in this study, the chemical modification of CNC via MMA grafting was performed to improve its compatibility with the MA molecule. The dispersibility of CNC both before and after modification was investigated by dispersing CNCs in the MA matrix, and the thermal stability of CNC was analyzed by the thermal analyzer. The reinforcement effect of MMA-modified CNC (MMA-CNC) in the printed composite was substantially exploited, as evidenced by its mechanical and thermal properties, compared with those of CNC/MA composites.
EXPERIMENTAL
Materials
Aqueous CNC suspensions with a dry CNC content of 6.3 wt% were obtained from the University of Maine (Orono, ME, USA). They were mainly produced by using sulfuric acid to hydrolyze the amorphous regions of the cellulose polymer, leaving the acid-resistant crystals as a product. Subsequently the crystals were purified by diluting and neutralizing the acid and then separating the soluble chemicals from the insoluble CNC. The diameter and length of the CNC particles were from 20 nm to 40 nm and from 150 nm to 250 nm, respectively.
The MA, a commercial product, was purchased from a 3D printing manufacturer (Formlabs Inc., Somerville, MA, USA). It mainly consisted of an oligomer and photoinitiator. Acetone, MMA, cerium (IV) ammonium nitrate (CAN), and nitric acid were supplied by Fisher Scientific Company LLC (Pittsburgh, PA, USA) and used as received. All solvents were analytical reagent grade.
Graft copolymerization
The aqueous CNC suspension was adjusted to a pH of 1 with diluted nitric acid. The CAN was added as an initiator, and then the obtained mixture was magnetically stirred for 30 min at 45 ℃. The MMA was gradually added over 30 min and then stirred for 60 min to allow a complete grafting reaction between CNC and MMA (Fig. 1). The initiator concentration was 0.004 mmol/cm3, and the amount of MMA was 40 mmol per g of dry CNC.
The reaction mixture was stirred and then centrifuged (4000 rpm, 20 min) with acetone three times to remove the homopolymer, initiator, and nitric acid (Littunen et al. 2011). The MMA-modified CNC (abbreviated as MMA-CNC in this context) was dispersed in acetone solution and then sealed for further experiments.
Fig. 1. Schematic of MMA grafting onto a CNC and polymerization
Preparation of MA-based composites by 3D printing
The MMA-CNCs or CNCs were directly mixed with MA at a content of 0 wt%, 0.5 wt%, 1 wt%, 2 wt%, and 4 wt%. Subsequently, the mixture was magnetically stirred for 10 min and then ultrasonically treated at 300 W for 6 min. The resulting mixture was 3D printed to prepare the samples through the process previously described (Feng et al. 2017). A total of 50% of the printed samples were treated using a post-curing process in an oven at 130 °C for 40 min.
Methods
Fourier transform infrared (FTIR) analysis
The FTIR spectra of unmodified and modified CNCs were analyzed using a FTIR-attenuated total reflectance (ATR) spectrometer (Spectrum One; PerkinElmer Corporation, Waltham, MA, USA) at 25 ℃. The samples were placed onto the diamond crystal of an ATR accessory. Data were collected over the spectral range of 400 cm-1 to 4000 cm-1 with a resolution of 4 cm-1. A total of 20 scans were applied for each sample.
Scanning transmission electron microscopy (STEM) analysis
The images of the MMA-CNCs and CNCs were analyzed by STEM, using a Zeiss Auriga dual beam SEM/Focus ion beam (FIB) (crossbeam 550; Carl Zeiss AG, Jena, Germany) with an accelerating voltage of 30 kV. Each sample for the STEM analysis was prepared by depositing one drop of MMA-CNC or CNC sample (0.05 wt%) onto a standard holey-carbon-film-covered copper grid (CF200-CU; Electron Microscopy Sciences, Hatfield, PA, USA), followed by the removal of excess liquid from the bottom of the grid and drying at ambient temperature before loading onto the microscope. The diameter and length of at least 100 particles were measured from five randomly selected STEM images. To investigate the dispersing durability of MMA-CNCs and CNCs in the MA matrix, the images of mixtures of MMA-CNC/MA and CNC/MA were taken at intervals of 0, 5, 10, 15, 20, and 60 days.
Tensile testing
The tensile strength, tensile modulus, and elongation of the printed MMA-CNC/MA and CNC/MA composites were tested at a speed of 1 mm/min according to ASTM D638 (2014) using a universal testing machine (Model 5567; Instron®, Norwood, MA, USA). The dumbbell-shaped samples with dimensions of 63.5 mm × 9.53 mm × 3.2 mm (L × W × T) were placed at 23 ± 2 ℃ and 50 ± 5% relative humidity for 24 h. A total of five duplicates were tested for each sample.
Thermal stability and performances
The thermal stability of MMA-CNCs, CNCs, and 3D-printed MMA-CNC/MA and CNC/MA composites were determined using the Perkin-Elmer 7 series thermal analyzer (PerkinElmer, Inc., Shelton, CT, USA). Each sample (approximately 8 mg) was heated at a heating rate of 10 ℃/min from 25 ℃ to 600 ℃ in a dry air atmosphere. Three duplicates were measured for each sample.
The thermal performances of the composites were measured using a differential scanning calorimeter (DSC; Diamond, PerkinElmer Corporation, Waltham, MA, USA). First, a sample of approximately 5 mg was heated from 25 ℃ to 210 ℃ (10 ℃/min) to remove the residual moisture and erase any thermal history. The temperature was held at 210 ℃ for 5 min and then cooled to 25 ℃ (10 ℃/min). Next, the sample was reheated from 25 ℃ to 210 ℃ (10 ℃/min). A nitrogen flow of 20 mL/min was applied throughout the DSC process to prevent oxidation. The glass transition temperature (Tg) was determined from the endothermic transition in the second scan, and the thermal polymerization temperature (Tp) and enthalpy (ΔH) were determined by the corresponding exothermic transition and peak area in the first scan, respectively.
RESULTS AND DISCUSSION
Characteristics of the MMA-modified CNC
The CNCs displayed typical absorption bands such as OH stretching at 3336 cm-1, CH stretching at 2897 cm-1, and bound water at 1646 cm-1 (Fig. 2) (Routray and Tosh 2012; Tran et al. 2016). After modification with MMA, the MMA-CNCs exhibited new bands, such as asymmetrical CH3 stretching at 2951 cm-1, C=O stretching at 1704 cm-1, asymmetric COO- stretching at 1538 cm-1, and CH2 rocking at 774 cm-1, which were assigned to the grafting of MMA onto CNCs and the polymerized MMA (Dai Prè et al. 2013; Anžlovar et al. 2016; Tran et al. 2016). This result confirmed that MMA was successfully grafted onto the CNC surface. The band intensity of MMA-CNCs located at 3336 cm-1 was considerably lower than that of the CNCs, and the band at 1646 cm-1 disappeared after modification. These variations were generally related to the modification of the H-bonding between the C=O group and the free -OH group in the MMA-CNC network, resulting in the hydrophobic characteristics of the MMA-CNCs (Sain et al. 2015; Tran et al. 2016).
Fig. 2. FTIR spectrums of CNC unmodified and modified with MMA
After MMA modification, the dimensions of MMA-CNC were smaller than that of CNC; the length and diameter was 155.2 nm ± 24.3 nm and 27.7 nm ± 4.4 nm, respectively, for MMA-CNC, and 222.7 nm ± 46.4 nm and 29.9 nm ± 8.5 nm, respectively, for CNC (Fig. 3). The CNC aggregation was clearly observed from the longitudinal direction (Figs. 3a and 3b), and the diameter of the CNCs was narrowly dispersed in the range of 20 nm to 35 nm. When MMA was grafted onto the CNCs, the length of MMA-CNCs were shorter than CNCs, and the diameter also decreased (Figs. 3c and 3d), indicating that the hydrogen bonds and van der Waals forces between the CNCs was substituted by chemical bonds formed between the CNCs and MMA. Moreover, the rupture of CNC molecular chains occurred through incorporation of the initiator and acid medium during grafting copolymerization (Canche-Escamilla et al. 2002). More importantly, the MMA-CNC was evenly dispersed in the solution without obvious aggregation, resulting in the homogenous distribution of MMA-CNC in the MA resin, as was previously discussed.
Fig. 3. STEM morphology of CNCs (a, b) and MMA-CNCs (c, d); inset graphs are the length (a, c) and diameter (b, d) of CNCs and MMA-CNCs
The CNCs exhibited a minor weight loss at approximately 80 ℃ caused by the evaporation of moisture (Fig. 4) (Sain et al. 2015). The CNCs displayed a one-step degradation that started at 225 ℃ (T5%, temperature at 5% weight loss) and obtained peak degradation at 315 ℃ (Tmax1, peak degradation temperature). However, MMA-CNCs showed a two-step degradation, with an initial degradation at 189 ℃ and two degradation peaks at 203 ℃ (Tmax1) and 420 ℃ (Tmax2).
The initial temperature of MMA-CNC, which was attributed to the elimination of coordinated water molecules (possibly through hydrogen bonds) and the decomposition of weak carbon-carbon bonds between units linked head-to-head (Sain et al. 2012; Tran et al. 2016), was lower than that of CNC, and the Tmax1 shifted to a lower temperature in the MMA-CNC relative to that observed for the default CNC. This result was explained by the relatively low thermal stability of MMA compared to unmodified CNC and the degradation of CNC caused by both the oxidation of CNC and the rupture of CNC molecular chains during grafting in the initiator and acid medium (Canche-Escamilla et al. 2002; Tran et al. 2016).
However, a new degradation peak (Tmax2) was observed in the MMA-CNCs that was higher than the Tmax1 of CNCs, and the weight of MMA-CNCs (60%) did not change from 250 ℃ to 400 ℃. The CNCs exhibited peak degradation and maximum weight loss (down to 20%), which indicated that the thermal stability of the CNCs were noticeably improved by graft copolymerization with MMA.