Nano-structural effects in high cellulose content thermoplastic

Jul 26, 2018 - ... or by solvent-assisted diffusion of PMMA into a CNF network (native and modified). The high content of CNF fibrils of ≈ 6 nm diam...
0 downloads 0 Views 7MB Size
Article Cite This: Biomacromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Biomac

Nanostructural Effects in High Cellulose Content Thermoplastic Nanocomposites with a Covalently Grafted Cellulose−Poly(methyl methacrylate) Interface Assya Boujemaoui,*,† Farhan Ansari,‡ and Lars A. Berglund*,†,§ †

Department of Fiber and Polymer Technology, KTH Royal Institute of Technology, SE-10044 Stockholm, Sweden Department of Materials Science and Engineering, Stanford University, Stanford, California 94305-2205, United States § Wallenberg Wood Science Center, KTH Royal Institute of Technology, SE-10044 Stockholm, Sweden Biomacromolecules Downloaded from pubs.acs.org by EASTERN KENTUCKY UNIV on 08/25/18. For personal use only.



S Supporting Information *

ABSTRACT: A critical aspect in materials design of polymer nanocomposites is the nature of the nanoparticle/polymer interface. The present study investigates the effect of manipulation of the interface between cellulose nanofibrils (CNF) and poly(methyl methacrylate) (PMMA) on the optical, thermal, and mechanical properties of the corresponding nanocomposites. The CNF/PMMA interface is altered with a minimum of changes in material composition so that interface effects can be analyzed. The hydroxyl-rich surface of CNF fibrils is exploited to modify the CNF surface via an epoxide-hydroxyl reaction. CNF/PMMA nanocomposites are then prepared with high CNF content (∼38 wt %) using an approach where a porous CNF mat is impregnated with monomer or polymer. The nanocomposite interface is controlled by either providing PMMA grafts from the modified CNF surface or by solvent-assisted diffusion of PMMA into a CNF network (native and modified). The high content of CNF fibrils of ∼6 nm diameter leads to a strong interface and polymer matrix distribution effects. Moisture uptake and mechanical properties are measured at different relative humidity conditions. The nanocomposites with PMMA molecules grafted to cellulose exhibited much higher optical transparency, thermal stability, and hygro-mechanical properties than the control samples. The present modification and preparation strategies are versatile and may be used for cellulose nanocomposites of other compositions, architectures, properties, and functionalities.



matrix.3,4 This results in poor dispersion and poor molecular scale interactions analogous to the problems with polymer blends. This has negative effects on mechanical performance, for instance, strain to failure and ultimate strength. The problem is often aggravated under high relative humidity conditions.5,6 For plant fiber/polymer composites, moisture tends to diffuse to the fiber−matrix interface, where weakening of the interface leads to reduced strength.7−9 Efforts to promote chemical compatibility between nanocellulose and hydrophobic matrices by chemical surface modification of cellulose have utilized a wide range of chemical methods such as esterification,10,11 etherification,12 silylation,13,14 and polymer grafting.15 For thermoplastic polymer matrices, the surface of nanocellulose is typically modified first, and then nanocellulose is physically blended with the polymer matrix.1 Polyethylene,16 polycaprolactone (PCL),17,18 polylac-

INTRODUCTION Increasing environmental awareness has inspired efforts on the development of materials from natural and renewable resources. Materials based on nanocellulose from wood have the potential to mitigate environmental concerns and contribute to improved material properties of cellulose/ polymer composites. Cellulose nanofibril (CNF) materials have attracted attention due to excellent mechanical properties, optical transparency, relatively low production cost, and renewable resource origin.1 CNF and similar cellulose fibrils have been extensively investigated as reinforcements for nanocomposites showing excellent mechanical and thermal properties.1,2 Still, the understanding of fundamental interfacial characteristics between the cellulose nanoreinforcement and the continuous matrix remains elusive. For instance, when mechanical properties of polymer matrix nanocomposites from CNF are lower than expected, it is difficult to separate the relative contributions from poor nanoparticle dispersion and poor interface interactions at the molecular scale. It also remains difficult to tailor cellulose− polymer interfaces for specific material combination. The challenge is often caused by the hydrophilic character of nanocellulose and hydrophobic nature of the polymer © XXXX American Chemical Society

Special Issue: The Rational Design of Multifunctional RenewableResourced Materials Received: May 1, 2018 Revised: July 23, 2018 Published: July 26, 2018 A

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules tide,19 and polypropylene20 have been blended with modified nanocelluloses. The referenced procedure uses solvent exchange of nanocellulose/water dispersion to an organic solvent, and problems include high viscosity and scalability.21 An alternative method used simultaneous modification of cellulose and covalent linking of the polymer matrix.22 Polycaprolactone (PCL) was grafted on a preformed CNF paper network or “mat” via in situ ring opening polymerization from the hydroxyl groups present on the fibril surface.22 A significant advantage is that the reinforcement content can be high by the use of a CNF network structure with favorable random-in-plane architecture, where the monomer is impregnated and polymerized. This is a facile preparation route for biocomposites inspired by resin transfer molding for thermosets, 23 and CNF contents can be as high as 50 wt %.22 Other polymers from vinyl monomers can also be grafted to the cellulose surface, where a reactive moiety is attached prior to polymerization.24 CNF network modification can also be performed via etherification by reacting an epoxide with the hydroxyls on cellulose. This altered the hydrophilic character of cellulose and improved hygromechanical properties of the resulting films.12 Poly(methyl methacrylate) (PMMA) is a transparent thermoplastic polymer often used as a glass substitute. The high transparency of PMMA is often retained in their cellulose based composites. For instance, the impregnation of delignified wood with PMMA oligomers followed by polymerization resulted in a composite with transmittance up to 85%.25 Moreover, “transparent wood” exhibited higher modulus compared to delignified wood and neat PMMA.25 In another study, PMMA nanocomposites were prepared by adding either CNF or CNC by physical blending and solvent casting. The nanocomposites with reinforcement from 0.25 and 0.5 wt % showed higher storage modulus compared to that of unfilled PMMA.26 Yano and co-workers prepared CNF/resin nanocomposites by impregnation of a dried CNF network by bulk acrylic resin of unknown composition followed by UV curing. They reported high optical transparency due to the nanoscale cellulose and matched refractive index of the acrylic resin and cellulose.27,28 In the present work, effects of CNF/PMMA interface grafting and in situ polymerization procedure on optical, thermal, moisture sorption, and mechanical properties of the resulting nanocomposites are investigated. The CNF surface is first tailored at the nanoscale so that the compatibility and reactivity with MMA is improved. The CNF surface modification is performed on a preformed nanoporous CNF network using epoxide molecules with allyl moieties. Thermoplastic CNF/PMMA nanocomposites are then prepared by either polymerization of PMMA from the modified CNF surface such that the CNF/PMMA interface is covalent or by solvent-assisted impregnation/diffusion of large PMMA molecules into the modified CNF network. An additional reference material is made by solvent-assisted impregnation/ diffusion of PMMA molecules into the unmodified CNF network. The nanostructure of the resulting nanocomposites is characterized by FE-SEM and is related to the optical, thermal, and mechanical properties under different humidity conditions. Initial reactions with epoxide-bearing molecules are technologically relevant because epoxies are widely used in the composite industry. Moreover, the present processing allows preparation of a nanostructured composite with high content (∼40 wt %) of well-dispersed CNF, where the matrix

phase is covalently linked to the reinforcement phase (CNF). The purpose is to clarify the potential of nanocomposites where molecules in the polymer phase are covalently linked to CNF. The present procedure is versatile and inspires routes for the preparation of high-performance nanocomposites with high content of CNF.



EXPERIMENTAL SECTION

Materials. CNF used in this work was prepared via enzymatic pretreatment of never-dried pulp supplied by Nordic Paper, Sweden according to a procedure reported elsewhere.29,30 The numberaverage diameter of the fibrils was reported to be 6.6 nm.31 Allyl glycidyl ether (AGE, ≥99%), triethylamine (TEA, ≥99%), methyl methacrylate (MMA, 99%), and 2,2′-Azobis(2-methylpropionitrile) (AIBN, ≥98%) were purchased from Sigma-Aldrich. Toluene (≥99%) and acetone (≥99%) were purchased from VWR International. Preparation of CNF Network. The CNF network was prepared from dilute CNF suspension (0.1 wt %) via vacuum filtration over a membrane (Millipore) with pore size of 0.65 mm. After filtration, a wet CNF/water “cake” was obtained and transferred to a bath filled with acetone. The CNF network in acetone was placed on a shaking device to facilitate solvent exchange. The acetone in the bath was changed every 8−10 h, and the process was repeated at least five times to ensure complete solvent exchange to obtain a CNF/acetone cake. The CNF network architecture is expected to be similar to the reported structures of nanoporous “paper” structures prepared by critical point drying and reported elsewhere.32 Preparation of Modified CNF Network (mCNF). The CNF network was placed in a jar containing acetone (13 mL) and TEA (0.5 g, 4.9 mmol); then, AGE (10 g, 87.6 mmol) was added, and the reaction was left to proceed for 24 h at room temperature on a shaking device. Thereafter, the modified CNF network (mCNF) was washed thoroughly with acetone at least five times. The attached amount of AGE on the CNF network was determined by gravimetric measurement, and the final composition of mCNF is reported in Table 1. The degree of substitution was 0.35 based on the moles of AGE attached per mole of cellulose hydroxyls.

Table 1. Compositions of CNF Film, AGE-Modified CNF (mCNF), PMMA-Grafted mCNF (mCNF-g-PMMA), PMMA/mCNF Blend Nanocomposite (mCNF-b-PMMA), and PMMA/CNF Blend Nanocomposite (CNF-b-PMMA) CNF content (%) CNF mCNF mCNF-g-PMMA mCNF-b-PMMA CNF-b-PMMA

100 80

AGE content (%)

0 20 Nanocomposites 38 10 38 10 30 0

PMMA content (%) 0 0 52 52 70

PMMA Grafting on Modified CNF (mCNF) Network (mCNFg-PMMA). The modified CNF network (mCNF) in acetone was solvent exchanged to toluene (as described for CNF solvent exchange from water to acetone). The CNF/toluene cake was then immersed in a reactor equipped with a magnetic stirrer and containing MMA (50 g, 0.5 mol), AIBN (164 mg, 1 mmol), and toluene (58 mL). The reaction mixture was degassed under vacuum for 45 min followed by argon bubbling for another 45 min. Thereafter, the reactor was immersed in a preheated oil bath at 75 °C, and the PMMA grafting reaction was carried for 22 h. The grafted mCNF (mCNF-g-PMMA) was washed thoroughly with toluene and then with acetone, dried, and then hot pressed using a Fontijne TP-400 NL hot-press operated at 120 °C for 10 min under a pressure of 2 MPa. Prior to hot-pressing, the nanoporous mCNF-g-PMMA network should be dominated by PMMA-coated CNF fibrils because nongrafted PMMA molecules B

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Scheme 1. Schematic Illustration of Nanocomposite Preparation: Modification of CNF Network with Allyl Glycigyl Ether (mCNF), Covalent Grafting of PMMA on mCNF via Solvent-Assisted Free Radical Polymerization (mCNF-g-PMMA), Physical Blend of Modified CNF/PMMA Prepared by Solvent-Assisted Diffusion of PMMA into the Modified CNF Network (mCNF-b-PMMA), and “Blend” of CNF/PMMA Prepared by Diffusion of High Molar Mass PMMA into the Porous CNF Network Gel (CNF-b-PMMA)a

a

All polymer diffusions were carried out in PMMA/acetone solution. normalized with respect to the region 2300−1900 cm−1 corresponding to the ATR crystal absorption. Thermal gravimetric analysis (TGA) was conducted on a TGA/ DSC Mettler Toledo AG instrument to determine the thermal decomposition of the modified fibers. Approximately 5 mg of sample was heated from 30 to 600 °C at a heating rate of 10 °C min−1 under a nitrogen atmosphere with a gas flow rate of 80 mL min−1. Differential scanning calorimetry (DSC) was performed on a Mettler Toledo DSC 820 equipped with a Mettler Toledo Sample Robot TSO801RO with cooling and heating rate of 10 °C min−1 under a nitrogen atmosphere. Approximately 5 mg of the sample was heated to 150 °C, held isothermally for 4 min to erase previous thermal history, and then cooled to 0 °C. The samples were then reheated to 150 °C, and the reported data were collected from the second heating ramp. Ultraviolet−visible spectroscopy (UV−vis) was conducted on a Shimadzu UV-2550 UV−vis spectrophotometer (Kyoto, Japan) to measure direct transmittance, the software UVProbe 2.0 being used to assess the data. The results were normalized to the film thickness (100 μm). Total transmittance of the samples was measured using a similar setup as described elsewhere.25 Field-emission scanning electron microscopy (FE-SEM) was performed using a Hitachi S-4800 operated at 5 kV and 2 A. The samples were coated with an ∼5 nm layer of platinum−palladium using an Agar HR sputter coater. Tensile testing was performed using a universal testing instrument Instron 5944 equipped with a 500 N load cell. Samples were cut into rectangular specimens 3 cm long and 3−4 mm wide and stored at 50% relative humidity for 3 days prior to testing. The strain rate was 10% min−1, and strain was measured using a video extensiometer. For the tests conducted under dry conditions, the samples were dried at 105 °C overnight and then kept in a desiccator until testing. The

have been dissolved and washed away. The monomer conversion was found to be 96% (see Figure S1). The free polymer formed in solution had a number-average molecular weight (Mn) of 64650 g mol−1 and a molar-mass dispersity (Mw/Mn) of 1.7 (see Figure S2). This free polymer was used to prepare CNF/PMMA and mCNF/ PMMA blends. The grafted amount of PMMA on the mCNF network was determined by gravimetric measurement, and the final composition of mCNF-g-PMMA is reported in Table 1. Preparation of CNF and mCNF/PMMA Blends. “Blended” CNF/PMMA nanocomposites were prepared by impregnation/ diffusion of PMMA molecules into the highly porous CNF network. The CNF/acetone or mCNF/acetone network was placed in a bath containing predissolved PMMA/acetone solution (23 mg mL−1) for 2 h. Then, acetone was evaporated from the mixture, and the films were dried and then hot-pressed at 120 °C for 10 min under a pressure of 2 MPa. Film compositions are summarized in Table 1, and digital photographs are presented in Figure S3. Characterization. Proton nuclear magnetic resonance (1H NMR) was performed on a Bruker AM 400 instrument at 400 MHz to determine the conversion of MMA. Deuterated chloroform (CDCl3) was used as solvent. MestReNova 9.0 software was used for peak integration. Size exclusion chromatography (SEC) was conducted on a TOSOH EcoSEC HLC-8320GPC instrument for determination of the molecular weight and molar-mass dispersity of PMMA. The mobile phase was dimethylformamide (DMF), and an EcoSEC RI detector was used for the analysis. PSS WinGPC Unity software version 7.2 was used to process data. Fourier transform-infrared spectroscopy (FT-IR) was performed on a PerkinElmer Spectrum 2000 equipped with a single reflection ATR system. The spectra were the average of 16 scans and were C

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules water-soaked samples were soaked in water for 2 days before testing. Water uptake was calculated by gravimetric measurement after conditioning the samples as above and was based on the dry weight.



RESULTS AND DISCUSSION CNF Modification and PMMA Polymerization Approaches. A preformed, wet CNF network was obtained by filtering of a colloidal CNF/water suspension. The CNF diameter is ∼6.6 nm (see Experimental Section), and the length is typically 0.7−1.5 μm. The wet cake with ∼20% dry content was subjected to solvent exchange to acetone. Subsequent surface modification with AGE and/or polymerization of MMA was carried out by impregnation/diffusion of the monomer molecules into the “wet” CNF/acetone gel network. The CNF reinforcement network was not dried to preserve structure and the high specific surface area.32 This nanoporous network consists of physically entangled fibrils with random-in-plane architecture resulting from the filtration procedure. Individual fibrils in the CNF network could then be surface-modified with the AGE. This was followed by grafting PMMA on the AGE-modified CNF (mCNF). There are three different materials in the study (see Scheme 1): 1. CNF/PMMA nanocomposites based on grafting of PMMA to the AGE-modified CNF network (mCNF-gPMMA). 2. Reference CNF/PMMA nanocomposites where high molar mass PMMA is added to the modified mCNF network by solvent-assisted diffusion (mCNF-b-PMMA), an approach previously reported for poly(styrene-co-butadiene).33 This reference material does not have a covalently linked PMMA−CNF interface. 3. Reference CNF/PMMA nanocomposites where PMMA is added to the unmodified CNF network by solvent-assisted diffusion (CNF-b-PMMA). The CNF/PMMA nanocomposites were prepared so that one material (material 1 above) contained PMMA molecules covalently grafted to the modified CNF network, and this is the material in focus. The modified CNF has reactive moieties (double bonds) on the fibril surface. As free radical polymerization of MMA is initiated, it is expected that reactions will take place with the double bonds at the fibril surfaces. This mainly occurs by growing PMMA chains reacting with the double bonds at the modified CNF surface or occasionally by AIBN initiation of polymerization from the native CNF.34 The two references of “blended” nanocomposites (mCNF-b-PMMA or CNF-b-PMMA) are made from either the modified or unmodified CNF network with high molar mass PMMA via diffusion into the solvent-soaked CNF network. Scheme 1 illustrates the preparation routes of the different CNF-based nanocomposites. Note that the physical system is a CNF network gel in acetone. For the mCNF-g-PMMA system, excess PMMA, which was not grafted to the CNF, was removed by washing with toluene and acetone. The purpose is to increase the fraction of PMMA molecules grafted to the CNF. The free polymer was precipitated in cold methanol and used for the preparation of mCNF-b-PMMA or CNF-b-PMMA. After the preparation methods described in Scheme 1, the nanocomposites were hotpressed at 120 °C above the Tg of PMMA to ensure negligible porosity. Reactions among CNF/AGE/MMA were verified by analyzing the FT-IR spectra (Figure 1). Successful attachment of AGE on mCNF was confirmed by the appearance of a new

Figure 1. FT-IR spectra of unmodified CNF, modified CNF (mCNF), and modified CNF grafted with PMMA (mCNF-gPMMA).

peak at 1650 cm−1 (C=C unsaturation from AGE).35 PMMA grafting to mCNF in the network was indicated by the sharp peak at 1730 cm−1 attributed to the C=O bond.36 After PMMA grafting, the intensity of the hydroxyl stretching peak between 3600 and 3000 cm−1 decreased due to the “shell” coating of CNF surfaces by PMMA chains. The peak observed at ∼2890 cm−1 for CNF and mCNF is ascribed to the C−H stretching band from the cellulose and hemicellulose backbone and AGE.37 For grafted mCNF (mCNF-g-PMMA), two additional peaks were observed in the same region as the peak at 2890 cm−1. They are attributed to the C−H stretching of CH2 and CH3 from PMMA.38 In this grafted nanocomposite, the interface between CNF and the matrix is expected to be dominated by covalent bonds and the CNF content was as high as 38 wt % (Table 1). Nanocomposites with “Noncovalent” Interface. Further, physically blended nanocomposites were prepared by impregnating either the native CNF network (designated as CNF-b-PMMA) or the mCNF network (designated mCNF-bPMMA) with PMMA. Note that polymerized PMMA was impregnated into the nanocellulose network, as opposed to MMA impregnation followed by polymerization as done for the mCNF-g-PMMA nanocomposite. This was done to prepare nanocomposite structures with similar CNF/PMMA composition but substantially different interfaces. To allow meaningful comparison of material properties, we aimed to obtain polymers with similar physical characteristics for all cases. Therefore, the PMMA used for impregnation was formed using the same reaction conditions as that for grafting to form mCNF-g-PMMA. However, the molecular weight (Mn) of the PMMA polymer grafted on the mCNF surface may differ from that of the free polymer formed in solution. This is because the grafting-from approach used for mCNF-gPMMA could lead to more complex polymer structures, such that loops and bridges are formed (because of the termination of a growing PMMA chain with an adjacent molecule on the same or an adjacent fibril) instead of linear PMMA chains. Furthermore, growing polymers in solution may also attach to cellulose via grafting-to approach. Previous attempts to cleave the grafted polymer from cellulose surface by enzymatic treatment had limited success due to the high CNF surface coverage and thus inaccessibility to the surface.15 Cellulose D

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Figure 2. (a, d, g) Schematic representation of the relative CNF/PMMA distribution and SEM images of cryo-fractured surface of (b, c) CNF blended with PMMA (CNF-b-PMMA), (e, f) modified CNF blended with PMMA (mCNF-b-PMMA), and (h, i) modified CNF grafted with PMMA (mCNF-g-PMMA).

dissolution by hydrolysis cleaves the ester groups in the PMMA polymer,15,39 thus rendering the analysis inconclusive. Morphology and Optical Properties. Morphological differences between the nanocomposites with similar CNF/ PMMA composition, but different interfaces were analyzed by SEM imaging of their cryo-fractured cross sections (Figure 2 and Figure S4). At low magnification, all of the nanocomposites show a layered morphology reminiscent of the pure nanopaper structure,40 and nanocellulose appears to be homogeneously dispersed within the polymer matrix even though the nanocellulose content was high (38 wt %). This was due to the preparation method adopted, where a preformed CNF network (filtered from water) was used as the starting material, thus minimizing the nanocellulose aggregation.5 It is interesting to note that the mCNF-gPMMA nanocomposite has relatively fewer dark regions, indicating the lack of voids and denser CNF/PMMA packing relative to the other two nanocomposites. Moreover, the layered structure is notably more defined for mCNF-g-PMMA due to the impregnation of the CNF network with MMA monomers followed by polymerization. For both of the blended nanocomposites, long polymeric chains of PMMA

were impregnated into the network, which could limit coating of the individual nanofibrils. In particular, for CNF-b-PMMA, the high magnification images show a loosely connected CNF network (Figure 2c), indicating that they were poorly bonded with the matrix. This was a direct consequence of the incompatibility between the native CNF and PMMA, such that the impregnation was not homogeneous, and the nanocomposites had voids at different scales. However, for mCNF-g-PMMA, individual nanofibrils at the scale of 10 nm were clearly observed (Figure 2i). The subtle difference in morphology resulting from the different interfaces had a noticeable effect on the optical properties of the nanocomposites. The total transmittance of the PMMA and the nanocomposites was ∼90% with PMMA showing the highest value followed by mCNF-g-PMMA and then the blended nanocomposites. For gaining further insight into the nanocomposite structure, “direct transmittance” normal to the sample plane was measured (which is based on the fraction of light transmitted in the same direction as the emitted light). mCNF-g-PMMA exhibited the highest visible light transmittance, ∼40% higher than mCNF-b-PMMA with the same PMMA content (Figure 3). Light transmittance E

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

Figure 3. (a) Total and (b) direct transmittance in the different materials: neat PMMA, CNF blended with PMMA (CNF-b-PMMA), modifiedCNF blended with PMMA (mCNF-b-PMMA), and modified-CNF grafted with PMMA (mCNF-g-PMMA).

Figure 4. (a) TGA and (b) DSC thermograms of unmodified CNF, neat PMMA, CNF blended with PMMA (CNF-b-PMMA), modified CNF blended with PMMA (mCNF-b-PMMA), and modified CNF grafted with PMMA (mCNF-g-PMMA).

Table 2. Mechanical and Thermal Property Data at Different Relative Humidities for Neat PMMA, CNF Blended with PMMA (CNF-b-PMMA), Modified CNF Blended with PMMA (mCNF-b-PMMA), and Modified CNF Grafted with PMMA (mCNF-gPMMA)a 5% RH

PMMA CNF-bPMMA mCNF-bPMMA mCNF-gPMMA

modulus (GPa)

strength (MPa)

2.3 (0.1) 4.6 (0.6)

50% RH

soaked in water strain (%)

modulus (GPa)

strength (MPa)

strain (%)

Tg (°C)

T50% (°C)

38 (1) 47 (1)

3.7 (0.2) 2.4 (0.1)

1.7 (0.1) 2.6 (0.3)

37 (1) 43 (2)

7.0 (1.1) 4.0 (0.5)

94 93

350 385

4.1 (0.9)

36 (4)

1.2 (0.4)

2.9 (0.4)

21 (2)

1.0 (0.2)

114

380

5.1 (0.4)

88 (6)

2.2 (0.2)

4.9 (0.1)

72 (7)

2.3 (0.3)

99

400

strain (%)

modulus (GPa)

strength (MPa)

48 (4) 52 (5)

5.7 (1.7) 1.7 (0.1)

2.0 (0.1) 2.5 (0.1)

7.2 (0.3)

72 (3)

1.1 (0.1)

6.8 (0.4)

98 (6)

2.1 (0.1)

a

Tg is the glass transition temperature from DSC analysis, and T50% is the temperature at which 50% weight loss occurs (TGA analysis).

through nanocomposites is sensitive to the interface between the components because optical losses in the form of diffraction may occur at the interface, especially in the presence of nanoscale air voids. Thus, the higher transmittance of mCNF-g-PMMA was due to the better polymer distribution resulting from the strong covalent interactions at the interface between CNF and PMMA. Because the polymers were grafted from the CNF surface, the CNF/PMMA interface in this case is likely to be favorable with minimal voids. Moreover, CNF-bPMMA had the lowest transmittance even though the PMMA content in CNF-b-PMMA is slightly higher than the other two nanocomposites (see Table 1).

The large difference in transmittance of the two blended nanocomposites (mCNF-b-PMMA vs CNF-b-PMMA) is rather surprising. This could be due to the fact that the allyl groups on the CNF surface may allow for a more uniform impregnation of the PMMA than the hydroxyls on the native CNF surface. Thus, the polymer distribution in CNF-b-PMMA will be far from ideal and is likely to be accompanied by smallscale voids. A recent study reported rather low transmittance of ∼29% for PMMA/CNF nanocomposites with only 0.25 wt % of CNF, although the thickness of the films was not reported.26 Thermal Properties. Thermal stability of CNF/PMMAbased nanocomposites was investigated using thermogravimetric analysis (Figure 4a). Surface modification of nanoF

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

CNF with PMMA (mCNF-b-PMMA) had a Tg increase of 20 °C over that of pure PMMA. Note that the polymer matrix in these nanocomposites is exactly the same, and therefore, the Tg difference must be related to the stronger interactions between PMMA and the modified CNF network containing allyl groups on the surface such that the chain mobility of PMMA is reduced.26,44 Better polymer distribution in mCNFb-PMMA (as indicated by optical transmittance data) would also cause a larger fraction of the PMMA matrix to be affected by the CNF surface, thus driving the Tg higher. Further, mCNF-g-PMMA had a Tg of 99 °C, which is higher than neat PMMA but lower than mCNF-b-PMMA even though the interfacial interactions are likely to be strongest for the mCNFg-PMMA. This anomaly could be due to minor differences in the polymer structure formed during the grafting-from approach or physical aging effects. Hygromechanical Properties. The moisture uptake of CNF/PMMA nanocomposites was monitored over time at 50% RH and after soaking in water. The equilibrium water uptake under both of these conditions is reported in Figure 5. PMMA, being a hydrophobic polymer, had low water uptake. Among the nanocomposites, mCNF-g-PMMA showed the lowest water uptake under both conditions, indicating that the covalent bonds at the CNF/PMMA interface indeed had a strong effect on the moisture sorption of the nanocomposite. Moreover, mCNF-b-PMMA had lower equilibrium moisture content among the two blended nanocomposites due to the relatively hydrophobic nature of the modified CNF. It is interesting to note that the water uptake of mCNF-g-PMMA after soaking in water was less than a quarter of that of CNF-bPMMA. This was because the CNF network swelling was rather limited in the case of mCNF-g-PMMA due to the covalently bonded CNF/PMMA interface. The effect of equilibrium moisture content due to the different interface

Figure 5. Water uptake of neat PMMA, CNF blended with PMMA (CNF-b-PMMA), modified CNF blended with PMMA (mCNF-bPMMA), and modified CNF grafted with PMMA (mCNF-g-PMMA) at 50% RH and after soaking in water.

cellulose affects its degradation temperature, as observed for oxidized CNF.41 The attachment of AGE and the subsequent grafting of PMMA on the CNF network increased their thermal stability (Figure 4 and Table 2). Furthermore, the blended nanocomposites (CNF-b-PMMA and mCNF-bPMMA) also exhibited higher degradation temperature compared to that of PMMA. A similar trend was observed for PMMA/CNF or CNC nanocomposites with a loading of 0.25 and 0.5 wt %.26 This observation was related to hydrogen interactions between carbonyl groups of the PMMA matrix and cellulose hydroxyl groups.42,43 High thermal stability has a strong technological relevance because it is of great importance for melt processing or shaping processes at elevated temperatures. Pure PMMA and CNF-b-PMMA exhibited relatively similar Tg of 94 and 93 °C (Figure 4b), respectively. However, the nanocomposite prepared by physical blending of modified

Figure 6. (a) Stress strain curves, (b) modulus, and (c) strength at different humidities of neat PMMA, CNF blended with PMMA (CNF-bPMMA), modified CNF blended with PMMA (mCNF-b-PMMA), and modified CNF grafted with PMMA (mCNF-g-PMMA). G

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

PMMA with graft links at the interface. Furthermore, thermal stability was better for grafted CNF-PMMA nanocomposites. The grafted CNF-PMMA nanocomposites prepared by in situ polymerization also consistently showed the best mechanical properties of the three materials in moist state. The combined nanostructural effects from in situ polymerization (more favorable matrix distribution and grafted PMMA molecules at the interface) resulted in roughly twice as high modulus and strength as for the two other nanocomposites under water soaking conditions. The results have implications for nanocomposite design at the nano- and molecular scale and for processing strategies for thermoplastic nanocomposites of high cellulose content. Fibril−fibril agglomerates may scatter light, and they will also show increased moisture sorption at cellulose−cellulose interfaces. This is detrimental to quasi-static tensile properties under moist conditions. Instead, the present results show that hydrophobic cellulose surface modification in combination with in situ polymerization and grafted polymer matrix chains at the interface provide favorable mechanical properties, and the approach is particularly advantageous under moist conditions. The in situ polymerization approach to processing, or reactive processing, is therefore a promising strategy for thermoplastic cellulose nanocomposites. It can not only lead to better polymer matrix distribution than in physical CNF− polymer blends but also be designed so that the polymer can be covalently grafted to the CNF.

was further investigated by mechanical characterization of the nanocomposites under different humidity conditions. The tensile test data under different humidity conditions are summarized in Table 2, and stress−strain curves are presented in Figure 6. Under dry conditions (∼5% RH) and at room temperature, which is below Tg (∼94 °C), PMMA showed a modulus and strength of approximately 2.3 GPa and 48 MPa, respectively. The blend of CNF and PMMA, i.e., CNF-bPMMA, showed a higher modulus of 4.6 GPa under dry conditions, which is almost 2× that of neat PMMA, while the strength was at the same level (Figure 6a, Table2). The relatively negligible improvements in strength allude to poor polymer distribution caused by an unfavorable interface between CNF and PMMA. Both nanocomposites prepared with modified CNF, mCNFb-PMMA and mCNF-g-PMMA, exhibited similar modulus, which was more than 3× that of neat PMMA. Interestingly, the strength and ductility of mCNF-g-PMMA was the highest among all the nanocomposites. This was due to the much improved polymer distribution such that defects were minimized. Under moist conditions, the mechanical performance of the blended nanocomposites CNF-b-PMMA and mCNF-b-PMMA decreased noticeably, and their modulus and strength were reduced to be in the same range as that of neat PMMA. When exposed to a humid environment, water molecules are preferentially located on the CNF surface,45,46 thus causing considerable weakening of the CNF/PMMA interface, such that the stress transfer is severely compromised. However, the mechanical properties of the nanocomposite with covalent bonds between the CNF and PMMA (mCNF-g-PMMA) were relatively well-preserved at high humidity and even after soaking in water. Moreover, the strength and strain-at-break of mCNF-g-PMMA were more than twice as high as those of mCNF-b-PMMA at different relative humidity conditions. In earlier works,47,48 it was shown that the effect of interface is stronger under high humidity conditions, but testing on samples soaked in water was not reported. In a study with low CNC content (0.5−3 wt %), PCL/CNC nanocomposites reinforced with grafted CNC maintained their mechanical performance even at high relative humidity conditions compared to their counterparts reinforced with either unmodified CNC or CNC modified by physisorption.49



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biomac.8b00701. NMR spectra and SEC traces of free PMMA, digital photographs, and low magnification cross section images of the composites (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected].



ORCID

Farhan Ansari: 0000-0001-7870-6327 Lars A. Berglund: 0000-0001-5818-2378

CONCLUSIONS Thermoplastic cellulose nanocomposites, with 30−38 wt % cellulose nanofibrils (CNF) oriented random-in-plane, were prepared by three different routes. The first two materials were physical blends between CNF and PMMA, with native CNF in one case, and surface-modified, hydrophobic CNF in the other material (modified by allyl glycidyl ether (AGE)). In the third material, the AGE-modified CNF was used to form grafted CNF-PMMA links at the interface during PMMA in situ polymerization. Optical transmittance is expected to scale with the content of light-scattering CNF aggregates, and indeed, grafted CNFPMMA showed the highest transmittance followed by the AGE-modified CNF-PMMA blend and the native CNFPMMA blend. The same order was obtained for moisture sorption under soaking conditions with the lowest moisture sorption in the grafted CNF-PMMA nanocomposite. The data show lower hygroscopicity and indicate better polymer matrix distribution at the nanoscale for the in situ-polymerized CNF-

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The Swedish Foundation for Strategic Research (SSF, project number 63634) and the Wallenberg Wood Science Centre (WWSC) are gratefully acknowledged for financial support. F.A. acknowledges funding from the Knut and Alice Wallenberg Foundation under the Wallenberg Research Link (WRL) at Stanford University.



REFERENCES

(1) Dufresne, A. Nanocellulose: From Nature to High Performance Tailored Materials; De Gruyter, 2012. (2) Lee, K.-Y.; Aitomäki, Y.; Berglund, L. A.; Oksman, K.; Bismarck, A. On the use of nanocellulose as reinforcement in polymer matrix composites. Compos. Sci. Technol. 2014, 105, 15−27. H

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules

reinforced polymer nanocomposites: A review. Polymer 2017, 132, 368−393. (22) Boujemaoui, A.; Carlsson, L.; Malmström, E.; Lahcini, M.; Berglund, L.; Sehaqui, H.; Carlmark, A. Facile Preparation Route for Nanostructured Composites: Surface-Initiated Ring-Opening Polymerization of ε-Caprolactone from High-Surface-Area Nanopaper. ACS Appl. Mater. Interfaces 2012, 4, 3191−3198. (23) Rudd, C. D.; Long, A. C.; Kendall, K. N.; Mangin, C. Liquid Moulding Technologies: Resin Transfer Moulding, Structural Reaction Injection Moulding and Related Processing Techniques; Elsevier Science, 1997. (24) Malmstrom, E.; Carlmark, A. Controlled grafting of cellulose fibres - an outlook beyond paper and cardboard. Polym. Chem. 2012, 3 (7), 1702−1713. (25) Li, Y.; Fu, Q.; Yu, S.; Yan, M.; Berglund, L. Optically Transparent Wood from a Nanoporous Cellulosic Template: Combining Functional and Structural Performance. Biomacromolecules 2016, 17, 1358−1364. (26) Erbas Kiziltas, E.; Kiziltas, A.; Bollin, S. C.; Gardner, D. J. Preparation and characterization of transparent PMMA−cellulosebased nanocomposites. Carbohydr. Polym. 2015, 127, 381−389. (27) Iwamoto, S.; Nakagaito, A. N.; Yano, H.; Nogi, M. Optically transparent composites reinforced with plant fiber-based nanofibers. Appl. Phys. A: Mater. Sci. Process. 2005, 81, 1109−1112. (28) Yano, H.; Sugiyama, J.; Nakagaito, A. N.; Nogi, M.; Matsuura, T.; Hikita, M.; Handa, K. Optically Transparent Composites Reinforced with Networks of Bacterial Nanofibers. Adv. Mater. 2005, 17, 153−155. (29) Henriksson, M.; Henriksson, G.; Berglund, L. A.; Lindström, T. An environmentally friendly method for enzyme-assisted preparation of microfibrillated cellulose (MFC) nanofibers. Eur. Polym. J. 2007, 43, 3434−3441. (30) Henriksson, G.; Nutt, A.; Henriksson, H.; Pettersson, B.; Ståhlberg, J.; Johansson, G.; Pettersson, G. Endoglucanase 28 (Cel12A), a new Phanerochaete chrysosporium cellulase. Eur. J. Biochem. 1999, 259, 88−95. (31) Prakobna, K.; Galland, S.; Berglund, L. A. High-Performance and Moisture-Stable Cellulose−Starch Nanocomposites Based on Bioinspired Core−Shell Nanofibers. Biomacromolecules 2015, 16, 904−912. (32) Sehaqui, H.; Zhou, Q.; Ikkala, O.; Berglund, L. A. Strong and Tough Cellulose Nanopaper with High Specific Surface Area and Porosity. Biomacromolecules 2011, 12, 3638−3644. (33) Dagnon, K. L.; Shanmuganathan, K.; Weder, C.; Rowan, S. J. Water-Triggered Modulus Changes of Cellulose Nanofiber Nanocomposites with Hydrophobic Polymer Matrices. Macromolecules 2012, 45, 4707−4715. (34) Roy, D.; Semsarilar, M.; Guthrie, J. T.; Perrier, S. Cellulose modification by polymer grafting: a review. Chem. Soc. Rev. 2009, 38, 2046−2064. (35) Duanmu, J.; Gamstedt, E. K.; Rosling, A. Synthesis and Preparation of Crosslinked Allylglycidyl Ether-Modified Starch-Wood Fibre Composites. Starch - Stärke 2007, 59, 523−532. (36) Boujemaoui, A.; Mongkhontreerat, S.; Malmströ m, E.; Carlmark, A. Preparation and characterization of functionalized cellulose nanocrystals. Carbohydr. Polym. 2015, 115, 457−464. (37) Pandey, K. K. A study of chemical structure of soft and hardwood and wood polymers by FTIR spectroscopy. J. Appl. Polym. Sci. 1999, 71, 1969−1975. (38) Worthley, C. H.; Constantopoulos, K. T.; Ginic-Markovic, M.; Pillar, R. J.; Matisons, J. G.; Clarke, S. Surface modification of commercial cellulose acetate membranes using surface-initiated polymerization of 2-hydroxyethyl methacrylate to improve membrane surface biofouling resistance. J. Membr. Sci. 2011, 385−386, 30−39. (39) Heinze, T.; Petzold-Welcke, K. Recent Advances in Cellulose Chemistry. In Polysaccharide Building Blocks; John Wiley & Sons, Inc., 2012; pp 1−50.

(3) Habibi, Y. Key advances in the chemical modification of nanocelluloses. Chem. Soc. Rev. 2014, 43, 1519−1542. (4) Klemm, D.; Kramer, F.; Moritz, S.; Lindström, T.; Ankerfors, M.; Gray, D.; Dorris, A. Nanocelluloses: A New Family of Nature-Based Materials. Angew. Chem., Int. Ed. 2011, 50, 5438−5466. (5) Ansari, F.; Berglund, L. A. Toward Semistructural Cellulose Nanocomposites: The Need for Scalable Processing and Interface Tailoring. Biomacromolecules 2018, 19, 2341−2350. (6) Benítez, A. J.; Torres-Rendon, J.; Poutanen, M.; Walther, A. Humidity and Multiscale Structure Govern Mechanical Properties and Deformation Modes in Films of Native Cellulose Nanofibrils. Biomacromolecules 2013, 14, 4497−4506. (7) Dhakal, H. N.; Zhang, Z. Y.; Richardson, M. O. W. Effect of water absorption on the mechanical properties of hemp fibre reinforced unsaturated polyester composites. Compos. Sci. Technol. 2007, 67, 1674−1683. (8) Abdelmouleh, M.; Boufi, S.; Belgacem, M. N.; Dufresne, A.; Gandini, A. Modification of cellulose fibers with functionalized silanes: Effect of the fiber treatment on the mechanical performances of cellulose−thermoset composites. J. Appl. Polym. Sci. 2005, 98, 974−984. (9) Ansari, F.; Sjöstedt, A.; Larsson, P. T.; Berglund, L. A.; Wågberg, L. Hierarchical wood cellulose fiber/epoxy biocomposites − Materials design of fiber porosity and nanostructure. Composites, Part A 2015, 74, 60−68. (10) Cunha, A. G.; Zhou, Q.; Larsson, P. T.; Berglund, L. A. Topochemical acetylation of cellulose nanopaper structures for biocomposites: mechanisms for reduced water vapour sorption. Cellulose 2014, 21, 2773−2787. (11) Sehaqui, H.; Zimmermann, T.; Tingaut, P. Hydrophobic cellulose nanopaper through a mild esterification procedure. Cellulose 2014, 21, 367−382. (12) Ansari, F.; Lindh, E. L.; Furo, I.; Johansson, M. K. G.; Berglund, L. A. Interface tailoring through covalent hydroxyl-epoxy bonds improves hygromechanical stability in nanocellulose materials. Compos. Sci. Technol. 2016, 134, 175−183. (13) Goussé, C.; Chanzy, H.; Excoffier, G.; Soubeyrand, L.; Fleury, E. Stable suspensions of partially silylated cellulose whiskers dispersed in organic solvents. Polymer 2002, 43, 2645−2651. (14) Andresen, M.; Stenstad, P.; Møretrø, T.; Langsrud, S.; Syverud, K.; Johansson, L.-S.; Stenius, P. Nonleaching Antimicrobial Films Prepared from Surface-Modified Microfibrillated Cellulose. Biomacromolecules 2007, 8, 2149−2155. (15) Malmström, E.; Carlmark, A. Controlled grafting of cellulose fibres - an outlook beyond paper and cardboard. Polym. Chem. 2012, 3, 1702−1713. (16) Junior de Menezes, A.; Siqueira, G.; Curvelo, A. A. S.; Dufresne, A. Extrusion and characterization of functionalized cellulose whiskers reinforced polyethylene nanocomposites. Polymer 2009, 50, 4552− 4563. (17) Habibi, Y.; Dufresne, A. Highly Filled Bionanocomposites from Functionalized Polysaccharide Nanocrystals. Biomacromolecules 2008, 9, 1974−1980. (18) Habibi, Y.; Goffin, A.-L.; Schiltz, N.; Duquesne, E.; Dubois, P.; Dufresne, A. Bionanocomposites based on poly(e-caprolactone)grafted cellulose nanocrystals by ring-opening polymerization. J. Mater. Chem. 2008, 18, 5002−5010. (19) Goffin, A.-L.; Raquez, J.-M.; Duquesne, E.; Siqueira, G.; Habibi, Y.; Dufresne, A.; Dubois, P. From Interfacial Ring-Opening Polymerization to Melt Processing of Cellulose Nanowhisker-Filled Polylactide-Based Nanocomposites. Biomacromolecules 2011, 12, 2456−2465. (20) Ljungberg, N.; Bonini, C.; Bortolussi, F.; Boisson, C.; Heux, L. Cavaillé, New Nanocomposite Materials Reinforced with Cellulose Whiskers in Atactic Polypropylene: Effect of Surface and Dispersion Characteristics. Biomacromolecules 2005, 6, 2732−2739. (21) Kargarzadeh, H.; Mariano, M.; Huang, J.; Lin, N.; Ahmad, I.; Dufresne, A.; Thomas, S. Recent developments on nanocellulose I

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX

Article

Biomacromolecules (40) Henriksson, M.; Berglund, L. A.; Isaksson, P.; Lindström, T.; Nishino, T. Cellulose Nanopaper Structures of High Toughness. Biomacromolecules 2008, 9, 1579−1585. (41) Fukuzumi, H.; Saito, T.; Okita, Y.; Isogai, A. Thermal stabilization of TEMPO-oxidized cellulose. Polym. Degrad. Stab. 2010, 95, 1502−1508. (42) Kuo, S.-W. Hydrogen-bonding in polymer blends. J. Polym. Res. 2008, 15, 459−486. (43) Dong, H.; Strawhecker, K. E.; Snyder, J. F.; Orlicki, J. A.; Reiner, R. S.; Rudie, A. W. Cellulose nanocrystals as a reinforcing material for electrospun poly(methyl methacrylate) fibers: Formation, properties and nanomechanical characterization. Carbohydr. Polym. 2012, 87, 2488−2495. (44) Han, G.; Huan, S.; Han, J.; Zhang, Z.; Wu, Q. Effect of Acid Hydrolysis Conditions on the Properties of Cellulose NanoparticleReinforced Polymethylmethacrylate Composites. Materials 2014, 7, 7. (45) Terenzi, C.; Prakobna, K.; Berglund, L. A.; Furó, I. Nanostructural Effects on Polymer and Water Dynamics in Cellulose Biocomposites: 2H and 13C NMR Relaxometry. Biomacromolecules 2015, 16, 1506−1515. (46) Bergenstråhle, M.; Wohlert, J.; Larsson, P. T.; Mazeau, K.; Berglund, L. A. Dynamics of Cellulose−Water Interfaces: NMR Spin−Lattice Relaxation Times Calculated from Atomistic Computer Simulations. J. Phys. Chem. B 2008, 112, 2590−2595. (47) Ansari, F.; Skrifvars, M.; Berglund, L. Nanostructured biocomposites based on unsaturated polyester resin and a cellulose nanofiber network. Compos. Sci. Technol. 2015, 117, 298−306. (48) Ansari, F.; Galland, S.; Johansson, M.; Plummer, C. J. G.; Berglund, L. A. Cellulose nanofiber network for moisture stable, strong and ductile biocomposites and increased epoxy curing rate. Composites, Part A 2014, 63, 35−44. (49) Boujemaoui, A.; Cobo Sanchez, C.; Engström, J.; Bruce, C.; Fogelström, L.; Carlmark, A.; Malmström, E. Polycaprolactone Nanocomposites Reinforced with Cellulose Nanocrystals SurfaceModified via Covalent Grafting or Physisorption: A Comparative Study. ACS Appl. Mater. Interfaces 2017, 9, 35305−35318.

J

DOI: 10.1021/acs.biomac.8b00701 Biomacromolecules XXXX, XXX, XXX−XXX