Modified and thermoplastic rapeseed straw
xylan: A renewable additive in PCL
biocomposites
Antonia Svärd a , Elisabet Brännvall b , Ulrica Edlund a,*
a Fiber and Polymer Technology, KTH Royal Institute of Technology, Teknikringen 56, SE-100
44 Stockholm, Sweden
b RISE Bioeconomy, Research Institutes of Sweden, Box 5604, SE-114 86 Stockholm, Sweden
Corresponding Author: Ulrica Edlund
* E-mail: edlund@kth.se
KEYWORDS Rapeseed/canola straw; xylan; grafting; hemicellulose, biocomposite; thermoplastic
ABSTRACT Xylan extracted from rapeseed straw was chemically modified to gain hydrophobic and thermoplastic properties via macroinitiator formation followed by a free radical grafting-from polymerization with octadecyl acrylate. Biocomposites were then prepared by incorporation of 5 or 20% (w/w) rapeseed straw xylan into a poly(-caprolactone) (PCL) matrix with by melt extrusion. The grafted xylan was homogeneously distributed within the biocomposite and reinforced the PCL matrix while at the same time preserving the ability to elongate to tensile strains > 500%. Analogous biocomposites made from unmodified xylan in a PCL matrix resulted in heterogeneous mixtures and brittle tensile properties.
1. INTRODUCTION
Rapeseed straw is an agro-industrial residue from the cultivation of rapeseed. The straw has few applications except burning or re-fertilizing the soil, while the seeds are high in demand for the production of cooking oil and biofuel.
Like all other high land-based plants, the straw consists primarily of the biopolymers cellulose, hemicellulose and lignin. If the rapeseed straw biopolymers were recovered, they could be converted to valuable products, such as gels, coatings and additives in plastics (Gabrielii and Gatenholm 1998; Edlund et al. 2010; Laine et al. 2013; Farhat et al. 2017). This would mean increased revenues, useful renewable materials available to the market, and still the possibility of downgrading them to solid fuel or composting for recovery of nutrients after their service life. Biopolymers of interest are the hemicelluloses, a group of structurally diverse polysaccharides such as xylan and glucomannan. Together with sound methods for conversion, hemicelluloses offer renewable material alternatives to oil-based products.
We have shown that polymeric hemicelluloses can be extracted from rapeseed straw by hydrothermal extraction. The extraction conditions: temperature, time and pH, control the exact composition of the extracted hemicelluloses. An extract rich in xylan (a representative structure is depicted in Scheme 1) with high molecular weight (Mw ca. 30 000 g/mol) and a low content
of lignin (3 – 5%) was obtained from extraction at 1.5 M NaOH(aq) at a temperature of 80 – 140 °C for 1 h (Svärd et al., 2017).
Hemicellulose-rich fractions are however not easily converted into valuable materials. Hemicelluloses are complex, displaying a variety of sugar compositions, molecular weights and degree of branching (Ebringerová et al., 2005). Hemicelluloses are not thermoplastic, instead, they are typically very brittle and in themselves poorly processable by conventional melt processing. In addition, the inherent hydrophilicity of hemicelluloses due to the presence of hydrophilic functional groups such as hydroxyl and carboxyl groups makes them typically poorly miscible with conventional bulk plastics and attempts to mix them usually result in heterogeneous blends or composites with agglomeration due to incompatibility (Faruk et al, 2012; Habibi, 2014). Chemical modification of hemicelluloses through esterification of the abundant hydroxyl pendant groups offers a straightforward pathway to derive hybrid hemicellulose materials and tune the properties. Esterification is a useful tool to immobilize reactive moieties, such as vinyl groups, onto the hemicellulose backbone, producing a
polysaccharide macroinitiator for subsequent chain-growth grafting polymerization (Voepel et al., 2011; Voepel et al., 2010; Yaich et al., 2017). Grafting, in turn, would allow for the introduction of new polymer motifs to the hemicellulose, providing viable and new property profiles such as an enhanced compatibility with bulk plastics.
Common thermoplastic polymer matrixes in commercial composites are polyethylene and polypropylene (Faruk et al, 2012). The biodegradability of poly(-caprolactone), PCL, is a strong reason why this polymer is gaining increasing attention, although it is produced from crude oil. PCL finds its use in for example filaments for 3D printing and in polyurethane foams. By mixing PCL with biobased materials, such as xylan, the carbon footprint of the composite can be decreased. PCL is hydrophobic, semi-crystalline, and has low viscosity. PCL has a melting temperature around 60 °C and a Tg around – 60 °C, which allows for melt-processing by conventional methods in a temperature range of 100 – 130 °C (Gross and Kalra, 2002; Mohamed and Yusoh, 2015), which is favorable when combined with polysaccharides where the thermal degradation starts at around 200 °C (Shen et al. 2010; Werner et al. 2014). Blending polysaccharides with a hydrophobic PCL matrix is not straightforward and usually leads to insolubility and phase-separation resulting in heterogeneous mixtures (Faruk et al, 2012). PCL was for example blended with starch (Kim et al., 2007), brewery spent grain (Hejna et al., 2015) and grafted cellulose nanoparticles (Siqueira et al., 2008; Herzele et al., 2016). The general trend was that the PCL-biocomposite became stiffer with increasing amount of filler and the materials ability to elongated decreased (Herzele et al., 2016; Hejna et al., 2015). Chemical modification of the polysaccharide pendant groups may help overcome the incompatibility with hydrophobic bulk plastics like PCL
Our aim was to convert xylan extracted from rapeseed straw to a graft-modified polysaccharide to improve the thermoplasticity of the xylan and its compatibility with hydrophobic matrix plastics. Our next aim was to use the modified xylan as a renewable plastic additive in a matrix of a biodegradable polymer: PCL, with the aim of producing a thermoplastic biocomposite demonstrating waste straw as a functional feedstock for biobased materials.
2. EXPERIMENTAL 2.1 Chemicals
Glycidyl methacrylate (GMA) (Aldrich Chemistry, > 97.0%, CAS: 106-91-2), octadecyl acrylate (ODA) (Aldrich Chemistry, 97%, CAS: 4813-57-4), ammonium persulfate (ACS
reagent > 98%, CAS: 7727-54-0), DMSO (Merck, CAS: 67-68-5), toluene (Fisher Scientific, CAS: 108-88-3) and 2-propanol (VWR Chemicals, CAS: 67-63-0) were used in the chemical modification of xylan. Aqueous solutions of NaOH (1 M and 1 mM) were prepared from NaOH(s) pellets (Sigma Aldrich, CAS: 1310-73-2). Deuterium oxide (Cambridge Isotope laboratories, Inc., USA, 99.9%, CAS: 7789-20-0), DMSO-d6 (Cambridge Isotope laboratories,
Inc., USA, D, 99.9%, CAS: 2206-27-1) and toluene-d8 (Cambridge Isotope laboratories, Inc.,
USA, 99.6%, CAS: 2037-26-5) were used as solvents for NMR analysis. In the reference polymerization of an ODA homopolymer, 2,2’-azobis(2-methyl-propionitrile) (AIBN) (Acros Organics, CAS: 78-67-1, 98%) was used as the initiator. Poly(-caprolactone) (PCL), (Aldrich Chemistry, average Mn 80,000 g/mol, pellets with a diameter of ~ 3 mm, CAS nr: 2480-41-4)
was used as the polymer matrix in the prepared biocomposites.
2.2 Rapeseed straw xylan
Two different fractions of rapeseed straw xylan were obtained by extraction with 1 M NaOH at 80 °C and 110 °C, respectively, for 1 h. Both fractions, herein denoted XylA and XylB, had high molecular weight (determined using alkaline water SEC) and consisted mainly of xylan as determined by high performance liquid chromatography with a pulsed amperiometric detector (HPLC-PAD) (Table 1) (Svärd et al., 2017).
Table 1. The extraction conditions, lignin and monosaccharide composition, molecular weights (Mw and Mn), and dispersity (Ð) of the two xylan fractions used in this study. The
composition data are given as the mean values of two individual replicates.
XylA XylB
Extraction conditions 1.5 M NaOH at 80 °C for 1 h 1.5 M NaOH at 110 °C for 1 h Lignin, % 5 ± 2 3 ± 0 Arabinose, % 8 ± 1 9 ± 0 Galactose, % 9 ± 1 8 ± 0 Glucose, % 11 ± 2 7 ± 0 Xylose, % 63 ± 4 72 ± 0 Rhamnose, % 1 ± 1 0 ± 0 Mannose, % 3 ± 0 2 ± 0 MW (g/mol) 25 030 36 000 Mn (g/mol) 17 050 22 520 Ð 1.5 1.6
2.3 Chemical modification of the rapeseed straw xylan
Scheme 1. Chemical pathway for the grafting of rapeseed straw xylan: (Step 1) Pre-activation of xylan (1 → 2), (Step 2) Synthesis of a Xyl-GMA macroinitiator (4) with GMA (3), and (Step 3) Graft polymerization from the macroinitiator (4) with ODA (5) yielding Xyl-GMA-g-ODA (6).
Step 1: Pre-activation of xylan
Freeze dried xylan (1 g) was dissolved in 8 mL of 1 M NaOH(aq) and stirred for 2 h at 65 °C.
Step 2: Synthesis of the xylan macroinitiator
GMA (2 mL) was added to the reaction mixture to start the synthesis of the macroinitator and the reaction was left for 20 h at 65 °C. In individual experiments, the reaction time was varied from 2 to 26 h to analyze the conversion as a function of reaction time. The reaction mixture was allowed to cool and then precipitated in acetone. The precipitate was air-dried. Two different macroinitiators were derived, denoted XylA-GMA and XylB-GMA, respectively. As a control experiment, XylA was reacted as described above for 20 h at 65 °C but no GMA was added to the reaction mixture.
Step 3: Grafting of xylan-GMA with ODA
The macroinitiator (500 mg) was suspended in 30 mL DMSO by stirring at 50 °C for 12 h. ODA (4 g) was dissolved in 15 mL of toluene and then added dropwise to the macroinitiator suspension along with 40 mg of ammonium persulfate. The temperature was set to 65 °C and the reaction was allowed to proceed for 2 h. In a time evaluation study of the grafting reaction, five identical reaction vessels were prepared in which the grafting reactions were left for 0.5, 1, 1.5, 2, and 4 h, respectively. Each reaction liquid was precipitated in 2-propanol and the precipitate was collected through centrifugation at 3000 rpm for 5 min. The collected precipitate was dissolved in toluene and again precipitated in 2-propanol and centrifuged at 1990 rpm for 5 min. This step was repeated five times. In the final step, the precipitate was dissolved in toluene again and centrifuged at 3700 rpm for 60 min and the formed gel was collected and air- dried. Grafted xylans were prepared from both xylan fractions, denoted XylA-GMA-g-ODA and XylB-GMA-g-ODA, respectively.
ODA reference polymerization
ODA was homopolymerized as a control experiment. ODA (8 g) was dissolved in 30 mL of toluene together with 80 mg AIBN and allowed to react at 70 °C for 2 h. The ODA homopolymer was precipitated in 2-propanol and washed according to the same protocol as applied to xylan. Finally, the precipitate was, dissolved in toluene and centrifuged at 3700 rpm
for 60 min. The resulting gel was collected and air dried for 48 h. The ODA monomer and the ODA homopolymer were both analyzed by NMR (see supplementary material Figure S1).
2.4 Preparation of PCL-based biocomposites with rapeseed straw xylans
Biocomposites were prepared by mixing PCL with 5% or 20% (w/w) of either XylA-GMA-g-ODA or XylB-GMA-g-XylA-GMA-g-ODA (added as a powder to the PCL granules) using a twin-screw miniextruder (DSMXplore 5 cm3 Micro-Componder). Composite samples of PCL with 5% and
20% of non-modified XylA and pure PCL were prepared as reference samples for comparison. The temperature was set to 100 °C with a hold time of 2 min after injection and a screw speed of 100 rpm in counter-rotation mode. The dog-bone shape, length 75 mm and testing area 50 x 5 x 2 mm (mould ISO527-2-1BA), was obtained using a mini-injection molder (Thermo Scientific Haake MiniJet Pro). The material was heated in a furnace to 130 °C and injected into the mold with an injection pressure of 500 bar for 10 s. The mold had a temperature of 30 °C and the post pressure was 100 bar for 10 s.
2.5 Instrumental methods
Attenuated total reflectance Fourier transform infrared spectrometry (ATR - FTIR)
Samples were analyzed at room temperature using infrared spectroscopy using a Perkin-Elmer Spectrum 2000 FTIR with an attenuated total reflectance (ATR) crystal accessory (Golden Gate) with corrections made for atmospheric water and carbon dioxide. All spectra was evaluated using Spectrum Timebase software, the spectra was calculated as an average of 32 individual scans at 4 cm-1 resolution, and the baseline was corrected.
Contact angle measurement
A Contact-Angle Meter CAM-200 from KSV Instruments was used to measure the static water contact angles on films of XylA, XylA-GMA, and XylA-GMA-g-ODA. The sessile drop method was used where a droplet of 3 µL of Milli-Q water was applied on the film with an automatic dispenser. The static contact angles were recorded using a Basler A6021-2 camera and calculated using One Attention software with corrected baselines. During the experiment, the contact area between the drop and the film was not changed during analysis. Only one droplet was applied on the XylA film, while on the XylA-GMA and XylA-GMA-g-ODA films three droplets were applied at different sites, and the static contact angle was calculated as an average.
Thermograms of XylA-GMA-g-ODA and XylB-GMA-g-ODA were recorded with a Mettler Toledo DSC 820 instrument. Under a nitrogen flow of 50 mL/min, the samples were heated from 25 ⁰C to 200 ⁰C and held at 200 ⁰C for 3 min, then cooled from 200 ⁰C to 25 ⁰C, held at 25 ⁰C for 3 min, and finally heated from 25 ⁰C to 200 ⁰C. The data were processed using STAR METTLER software.
Field-emission scanning electron microscopy (FE-SEM)
The fracture surfaces of dog-bone shaped samples after tensile testing and the film surfaces prepared for contact angle measurements were observed by ultra-high-resolution field emission scanning electron microscopy (FE-SEM) using a Hitachi S-4800, operating at 5 kV. Samples were attached to stubs using double-sided adhesive carbon tape. A Cressington 20HR Au/Pd sputter coater was used to coat the samples with a 7 nm thick layer of gold/palladium under an inert atmosphere.
Nuclear magnetic resonance (NMR)
Samples of Xyl-GMA (25 mg) were dissolved in 1 mL of D2O. To remove internally bound
water, the samples were then freeze-dried. This process was repeated three times, after which 0.6 mL DMSO-d6 was added and the samples were pipetted into NMR tubes with an outer
diameter of 5 mm. 1H NMR spectra of 128 scans were recorded at 400 MHz on a Bruker
DMX-400 NMR spectrometer. Samples of XylA-GMA (25 mg) in 0.6 mL D2O were analyzed using
the same instrumentation. Samples of Xyl-GMA-g-ODA were dissolved in 0.6 mL toluene-d8
and stirred at 50 °C for a few hours and then directly pipetted into NMR tubes with an outer diameter of 5 mm and analyzed. MestReNova software was used for data processing.
Size exclusion chromatography (SEC)
The molecular weights (Mw and Mn) and dispersities (Đ) of the XylA-GMA and XylB-GMA
were estimated using a water/alkaline solution SEC. Each sample (4 mg) was dissolved in 1 mL of 10 mM NaOH(aq) and stirred on a shaking table until acceptable dissolution was achieved
(approximately 2 h). The SEC system used was a Dionex Ultimate-3000 HPLC (Dionex, Sunnyvale, CA, USA) equipped with three PS surprema columns in series (300 × 8 mm, 10 µm particle size) with pore sizes of 30 Å, 1000 Å and 1000 Ås, together with a guard column (50 × 8 mm, 10 µm particle size), LPG-3400SD gradient pump, a WPS-3000SL autosampler a Waters-410 refractive index detector (Waters, Millford, MA, USA) and a DAD-3000 UV/Vis detector (Dionex, Sunnyvale, CA, USA). Before injection, the samples were filtered through
0.45 µm PTFE filters into analysis vials. Pullulan standards with controlled molecular weights ranging from 342 to 708,000 g/mol (PSS, Germany) were used for calibration. The mobile phase was 10 mM NaOH(aq) with a flow rate of 1 mL/min and was maintained at 40 °C during
analysis. The data were processed and analyzed using Chromeleon 7.1 software.
Thermal gravimetric analysis (TGA)
XylA-GMA-g-ODA and XylB-GMA-g-ODA were analyzed using a Mettler Toledo TGA/STDA 851w instrument. The samples (approximately 10 mg each) were heated from 25 ⁰C to 800 ⁰C at a heating rate of 10 ⁰C/min under an N2 atmosphere (50 mL/min). The data
were processed using STAR software.
Tensile testing
Dog-bone shaped sample specimens were prepared by injection molding as described above and conditioned for 3 -7 days at 23 ⁰C and in 50% relative humidity. An Instron 5566 was used for tensile testing and the samples were tested at a speed of 50%/min. Bluehill software was used for test control and collection of data as well as for calculation. At least 5 specimens were tested for each material.
3. RESULTS AND DISCUSSION
Rapeseed straw biomass contains hemicelluloses, which can be recovered in appreciable yield and with high molecular weight using a purposely designed hydrothermal treatment. Two xylan-rich extracts, herein denoted XylA and XylB, respectively, were prepared and used in the present study by applying the conditions shown in our previous work to generate the highest molecular weight and yield of xylan. XylB had a somewhat higher xylan content and a higher molecular weight than XylA (Table 1) (Svärd et al., 2017). In the present study, the inherently hydrophilic (Farhat et al., 2012) xylan was turned hydrophobic by grafting polymerization in a three step process (Scheme 1). First, XylA was used as the reagent to elaborate and evaluate the synthesis conditions. Then, the most viable set of reaction conditions were applied to both XylA and XylB to prepare two grades of modified xylan that were individually mixed with PCL. For comparison, composites were also prepared from mixtures of PCL with non-modified XylA and XylB, respectively.
Xylan (1), in Scheme 1, was first pre-activated (2) in alkali followed by an esterification reaction in which double bonds were immobilized onto xylan via coupling to GMA (3) yielding a vinylic functionalized xylan macroinitiator (4). The esterification reaction between a polysaccharide and GMA is a widely used method to attach vinyl entities to the polysaccharide backbone, and thereby create initiation sites for subsequent free radical polymerization (Edlund and Albertsson, 2014, Laine et al., 2013). An advantage is that the reaction can be performed in basic aqueous solution.
The successful immobilization of vinyl moieties to the xylan backbone was confirmed by H1
-NMR (Figure 1A) and ATR-FTIR (Figure 1B). The two peaks at 5 and 5.5 ppm in the H1
-NMR spectra (Figure 1A), are assigned to the vinyl hydrogens on the macroinitiator. These peaks were not present in the crude xylan sample but appeared early during the reaction indicating that the coupling reaction was fast and that there is no need to prolong the reaction time up to 24 h. The cluster of peaks between 3 and 4 ppm represent the anomeric protons of the polysaccharide backbone and include a peak at 3.3 ppm originating from water present in the sample. The peak at 2.5 ppm corresponds to the DMSO-d6. The small peak at 1.6 ppm is
characteristic of the acetyl pendant groups present on the hemicellulose chain (Teleman et al., 2000; Teleman et al., 2002). H1-NMR spectra of the XylA-GMA recorded from 2 to 26 h in
D2O are shown in Figure S2 in the supplementary material.
ATR-FTIR spectra (Figure 1B) confirmed the observations from NMR: the reaction reached a high conversion already after 2 h and longer reaction times did not lead to any additional changes in the FTIR absorption bands. An overlaid spectrum of Figure 1B can be viewed in supplementary material Figure S3. The ATR-FTIR spectrum of the pristine xylan resembled the typical spectrum of a polysaccharide. A broad band in the 3500 – 3200 cm-1 region
corresponds to the OH stretching vibrations and the broad peak around 1050 cm-1 indicates the
presence of C–O and C–O–C bonds. The two bands at 1600 – 1585 cm-1 and 1500 – 1400 cm -1 are characteristic of C–C stretches in aromatic ring structures, verifying the presence of a
small amount of lignin in the samples (Table 1). A small absorption band at 895 cm-1 is
characteristic of the vibration of the β-glycosidic bond between the sugar units in the polysaccharide chain (Gupta et al., 1987; Socrates, 2001). ATR-FTIR spectra of the macroinitiator show the presence of the characteristic bands of a polysaccharide and, in addition, an absorption peak at 1660 cm-1. This peak is characteristic of alkene C=C stretching
vibrations and is not present in the pristine xylan, indicating that GMA has indeed reacted with the xylan chains. In addition, a small adsorption band at 721 cm-1 assigned to –CH
and a small absorption band at 1733 cm-1 originating from the ester C=O in each GMA unit
were observed in the Xyl-GMA spectra. Small absorption bands at 3000 – 2850 cm-1 and 1470
cm-1 stem from C–H groups present in both the polysaccharide and in the immobilized GMA
(Socrates, 2001).
Figure 1. A) 1H-NMR (in DMSO-d6) spectra of the progressing reaction between rapeseed
straw xylan and GMA at different reaction times. Peak assignments: a. vinyl C=CH2 of
the GMA immobilized onto xylan, b. polysaccharide ring hydrogens overlaid with -CH2-
protons of GMA, c. DMSO-d6 solvent peak, d. –CH3 of GMA, and e. -CH3 in acetyl
substituents on xylan. B) ATR-FTIR spectra of the progressing reaction between rapeseed straw xylan and GMA at different reaction times. Peak assignments: a*. C=O in GMA, b*. primary alkene C=C stretching vibration in GMA, c*. internally bound water, d*. overlaid bands from –COO– vibrations, C-H and C-O stretching, e*. –C-H– vibration in GMA, f*. C-O and C-O-C stretching, and g*. vibration of the -glycosidic bond.
There is a potential risk of xylan degradation due to peeling under the conditions applied (65 ⁰C in 1 M NaOH) in the pre-activation step (Scheme 1, Step 1) as well as in the GMA-coupling step (Scheme 1, Step 2) (Gellerstedt, 2009). However, xylan is known to have some resistance towards degradation, since the xylan chain is substituted with arabinose and methyl glucoronic acid groups, which inhibits peeling (Gellerstedt, 2009; Testova et al., 2014).To investigate this further, SEC (Table 2) assessed the molecular weights of xylan and Xyl-GMA samples at different reaction times, 2 – 26 h. A small reduction is observed when comparing the molecular weight of pristine XylA (Mn = 17,050 g/mol, Table 1) to the molecular weight of XylA
subjected to 1 M NaOH in 65 °C for 2 h but without the addition of GMA (Mn = 14,400 g/mol,
Table 2). It can thus be expected that xylan to a small extent will degrade by the peeling reaction during the GMA coupling reaction (Knill and Kennedy, 2003). However, the molecular weights of Xyl-GMA remain constant throughout the 26 h of reaction (Table 2) and degradation is seemingly not significant. It has to be noted that there is a significant difference in molecular weights between the pristine xylans and the Xyl-GMA macroinitiators. This can be explained by the difference in affinity of Xyl and Xyl-GMA to the SEC column and the eluent solvent resulting in quite different hydrodynamic volumes.
Table 2. Molecular weights (Mw and Mn) and dispersity (Ð) of xylan, XylA-GMA after
different reaction times, and XylB-GMA.
Reaction Time (h) Mn (g/mol) Mw (g/mol) Ð
XylA-GMA 2 9 200 9 700 1.1 4 9 000 9 400 1.1 6 9 900 11 700 1.2 8 10 250 12 200 1.2 10 9 500 10 100 1.1 12 9 700 10 600 1.1 14 9 800 10 700 1.1 16 9 600 10 650 1.1 18 9 600 10 400 1.1 20 9 500 10 600 1.1 22 9 400 9 800 1.0 24 9 500 10 000 1.0 26 9 200 9 600 1.0 XylB-GMA 20 10 000 11 200 1.1 XylA in 1 M NaOH, 65 °C, for 2 h 14 400 23 300 1.6
3.2 Grafting-from polymerization of xylan macroinitiator with ODA
Xylan is inherently not thermoplastic and cannot be processed with conventional melt-processing techniques as it is thermally degraded above 200 °C (Belmokadder et al. 2011; Shen et al. 2010). To obtain a more thermoplastic and hydrophobic xylan, and hence potentially better compatibility with composite bulk plastics, a viable strategy may be to decorate the polysaccharide backbone with polymeric graft chains using a hydrophobic monomer. ODA with its long carbon chain stand out as such a monomer. Interfacial polymerization with toluene and DMSO was utilized, where the Xyl-GMA macroinitiators were suspended in DMSO while ODA was dissolved in toluene. This interfacial polymerization system has previously been successfully utilized to mediate grafting from nanofibrillated cellulose, overcoming parent insolubility problems (Navarro and Edlund, 2017). A phase separation occurred as the propagation of the grafted chains proceeded with a precipitated phase containing the poly(ODA)-grafted xylan: XylA-GMA-g-ODA (Figure S4 in supplementary material). In addition, the same phenomenon was observed by Haddleton et al (2013; 2014) and by Navarro and Edlund (2017), and the phase separation does not seem to disturb the propagation of the grafts. The kinetic progression of the grafting polymerization of ODA onto Xyl-GMA was monitored with H1-NMR (Figure 2) and ATR-FTIR (Figure S5, in supplementary material).
The peak at 1.4 ppm is attributed the –CH2– protons in the ODA main chain and an increase in
peak intensity over time indicates that the ODA graft chains propagated. The peak at 6.1 ppm was assigned to the vinyl groups of the ODA monomer and this peak decreases in intensity as the reaction progresses. The conversion seems to be high after 2 h of reaction. The peak at 2.1 ppm is characteristic of water and the cluster of peaks at 7.0 ppm corresponds to toluene. An increasing intensity of the absorption band at 3000 – 2850 cm-1 representing the C-H bonds in
Figure 1.1H-NMR spectra (in toluene-d8) of the progressive grafting polymerization of
XylA-GMA with ODA at different reaction times. Peak assignments: a. toluene-d8 solvent peaks, b. vinyl C=CH2 of the GMA immobilized onto xylan, c.–CH3 adjacent to the vinyl
bond of GMA, d.-CH2 and -CH of the grafted poly(ODA) main chain, e. -CH2 of the
grafted poly(ODA) side chain and –CH3 in GMA that reacted with ODA.
For comparison, and to verify the peak assignments, a homopolymer of ODA was polymerized and analyzed with 1H-NMR (Figure S1, in supplementary material). Based on the NMR data
in Figures 1 and 2, reaction times of 20 h for the GMA immobilization and 2 h for the grafting were chosen as sufficient. Hence, both XylA-GMA-g-ODA and XylB-GMA-g-ODA were prepared using these conditions for subsequent biocomposite preparation and analyzed with
1H-NMR and ATR-FTIR (Figure S6, in supplementary material).
The effect of grafting of ODA onto rapeseed xylan was evident when analyzing the grafted xylan with DSC (Figure 3). The analysis showed that XylA did not undergo any transition in the entire temperature range, 0 – 200 ⁰C, during neither heating nor cooling. This was expected considering the restricted segmental mobility of the polysaccharide chains. The same was true
for the macroinitiator XylA-GMA. As the grafting of ODA onto Xyl-GMA proceeded, the modified xylan gained a more thermoplastic behavior. The grafted xylan, XylA-GMA-g-ODA, underwent a clear melting transition during heating and crystallization during cooling (35 – 60 ⁰C). These transition temperatures are very similar to the melting and crystallization temperatures of the ODA homopolymer (35 – 50 ⁰C) and another clear indication of a successful grafting reaction. XylB-GMA-g-ODA was similarly analyzed with DSC and grafted xylan exhibited a melting endotherm (Figure S7, in supplementary material)
Figure 2. DSC thermograms of heating (right) and cooling (left) of (from top to bottom): ODA homopolymer, XylA grafted with ODA at different times, the macroinitator XylA-GMA20h and XylA.
The NMR data in Figure 1 indicate that reaction times of 2-6 h could be sufficient to synthesize a macroinitiator by GMA coupling to the rapeseed xylan (Scheme 1, Step 2). However, when comparing the DSC traces of XylA-GMA6h-g-ODA and XylA-GMA20h-g-ODA (Figure S8, in supplementary material) it was evident that XylA-GMA20h-g-ODA underwent a much stronger melting endotherm. Ultimately, a reaction time of 20 h was chosen for Step 2.
The grafted samples were in addition analyzed with TGA in a temperature range of 25 – 800 ⁰C (Figure S9, supplementary material). Initial weight loss is observed at 25 – 150 ⁰C for XylA and XylA-GMA due to the evaporation of bound water. Grafted xylan did not undergo any weight loss in this temperature range, indicating that the grafting rendered the modified xylan less hygroscopic. Xylan starts to degrade around 200 ⁰C (Shen et al., 2010), which is evident in the thermogram of XylA and XylA-GMA. The onset of thermal degradation was shifted towards 300 ⁰C for the grafted xylan samples, indicating increasing thermal stability, in accordance with other esterification modifications such as acetylation (Belmokadder et al. 2012; Fundador et al. 2012).
The monomer was purposely chosen to add a hydrophobic element to the otherwise mostly hydrophilic xylan (Farhat et al., 2017). The hydrophilicity of films of XylA-GMA20h-g-ODA2h were analyzed by measuring the static contact angles and compared to films of XylA and XylA-GMA (Figure 4). The static contact angle was significantly higher for the grafted xylan film than the XylA film and XylA-GMA macroinitiator film. After 5 seconds, some of the water was adsorbed into the grafted xylan film and the static contact angle decreased. On the xylan film, the drop was adsorbed fully after 3 seconds, while on the film of the macroinitiator, the water was not fully adsorbed and the contact angle stabilizes around angle of 15⁰. The surface of the film also influences the static contact angle. The true wettability of a film will not be obtained from the static contact angle measurement unless the surfaces are smooth and flat (Jackson et al., 2004). However, if major changes in wettability are measured it will still provide a good indication of the hygroscopy of the material surface. Therefore, the surfaces of the films were analyzed with FE-SEM. The grafted xylan film surface was even and smooth, while the xylan film surface was uneven and the XylA-GMA was a little more even.
Figure 3. A) Contact angle on films of XylA, XylA-GMA and XylA-GMA-g-ODA as it varies over time. B) Photos of the droplets behavior on the XylA, GMA and XylA-GMA-g-ODA film surfaces and FE-SEM pictures of the film surfaces at 500x magnification.
3.3 Composites of PCL and rapeseed straw xylan
Biocomposites were prepared using a twin-screw miniextruder and a mini-injection molder was then used to obtain dog-bone shaped specimens (Figure S10, in supplementary material). PCL has a melting point around 60 ⁰C and Tg around – 60 ⁰C and can thereby be melt processed relatively low temperatures, around 100 ⁰C (Gross and Kalra, 2002; Mohamed et al., 2015). It is important to keep the working temperature low when hemicelluloses such as xylan are being processed, since they contain glycosidic linkages which are known to degrade around 200 ⁰C (Shen et al., 2010). Six biocomposites were prepared in this work: PCL with 5% and 20% XylA-GMA-g-ODA, PCL with 5% and 20% XylB-GMA-g-ODA, and finally PCL with 5% and 20% unmodified XylA as reference samples. In addition, specimens of pure PCL were produced for comparison, using the same melt processing parameters as for the biocomposites. The biocomposites became yellowish with the addition of the grafted xylan and brownish when unmodified xylan was added. Pure PCL samples were white (Figure S9, in supplementary material).
The tensile properties of the biocomposites and references are listed in Table 3 and FE-SEM images of the fracture areas are shown in Figure 5. PCL exhibited a typical viscoelastic behaviour during tensile testing with necking around a strain of 10% and a strain-at-break > 500%. A stiffening and reinforcing effect was observed when xylan was introduced into the PCL matrix and the effect was more pronounced with higher amounts of xylan. Unmodified xylan showed a stronger reinforcing effect than grafted xylan but unmodified xylan severely impaired the ductility. The tensile strain at maximum tensile stress decreased from >500% for pure PCL to around 6% for PCL/XylA. Previously reported blends of micro-fibrillated cellulose and PCL (Herzele et al., 2016) and brewery spent grain blended with PCL (Hejna et al., 2015) exhibited as similar mechanical behaviour as the PCL/XylA ungrafted blend, where an increase in stiffness and a concomitant decrease in ductility was observed. In both these previously reported studies, the stiffness of the composite increased with increasing filler content while the elongation decreased.
Interestingly, mixtures of PCL with grafted xylan, XylA-GMA20h-g-ODA2h or XylB-GMA20h-g-ODA2h, were homogenous and highly ductile. With a 20% addition of grafted xylan, the tensile strain at maximum tensile stress was above 500%. Out of 5 specimens tested, three failed by slipping the grips as the volume of the specimen changed during tensile testing of the dog-bone specimens and the recorded maximum strains were hence lower than it would
have been if the samples would have continued to elongate to a break. The true strain-at-break may actually be higher than the values given in Table 3 for the PCL/grafted xylan biocomposites. PCL/unmodified xylan biocomposites were brittle and heterogeneous with a clear phase separation (Figure 5). The formation of heterogeneous blends is a common problem when combining PCL with polysaccharides or fibers (Faruk et al., 2012; Mohamed et al., 2015; Herzele et al., 2016; Hejna et al., 2015). It was clear that the ODA grafting polymerization strategy ODA chain was vital to achieve a homogenous mixture between PCL and rapeseed straw xylan, as can be clearly viewed when comparing the fracture surfaces of the biocomposites.
There were only minor differences between XylA and XylB with respect to the outcome of the grafting reaction, PCL compatibility, and the tensile performance of the biocomposites. XylA and XylB seems equally viable reagents for the production of biobased plastic additives to PCL biocomposites. XylA was extracted at lower temperature than XylB and from an energy consumption perspective, XylA would be preferred.
Table 3. Young's modulus, maximum tensile stress and tensile strain of PCL and PCL/rapeseed straw xylan biocomposites
Sample Young’s modulus, MPa Maximum tensile stress, MPa Tensile strain at maximum tensile stress, % PCL 300 ± 30 26 ± 2 555 ± 65 PCL + 5 % XylA-GMA20h-g-ODA2h 340 ± 20 23 ± 2 455 ± 46 PCL + 5 % XylA 510 ± 40 6 ± 0.6 17 ± 0.5 PCL + 20 % XylA-GMA20h-g-ODA2h 440 ± 20 26 ± 0.7 540 ± 50 PCL + 20 % XylA 570 ± 30 7 ± 0.8 2 ± 0.2 PCL + 5 % XylB-GMA20h-g-ODA2h 350 ± 10 24 ± 1 515 ± 46 PCL + 20 % XylB-GMA20h-g-ODA2h 420 ± 50 27 ± 2 600 ± 100
Figure 5. FE-SEM images at 500x magnification of fracture surfaces after tensile testing of PCL/rapeseed straw xylan biocomposites.
Xylan and other hemicelluloses are potentially valuable biobased resources. Their applications, and hence their competitiveness with concurrent fossil-based materials, are limited by the inherent hydrophilicity. Property control and usage at higher relative humidities is very challenging. The potential applications of these bio-based resources are extended by
hydrophobization and similar modifications. (Ibn Yaich et al., 2017) Hydrophobized xylan may find use as plastics, resins and films. (Gordobil et al., 2014; Belmokaddem et a., 2011). In composites, the poor interfacial compatibility of pristine hydrophilic biopolymers with conventional composite matrices have negative effects on the composite homogeneity na mechanical performance. In this work, we mixed PCL with xylan to combine the thermoplastic properties of fossil-derived PCL with a bio-based matrix. Our results clearly show that non-modified rapeseed xylan do not form viable composites with PCL while the grafting non-modified xylan blends very well with the matrix. The resulting biocomposite could be used in products where pure PCL is used. PCL finds its use in for example filaments for 3D printing and in polyurethane foams. By mixing PCL with biobased materials, such as xylan, the carbon footprint of the composite can be decreased low value and few applications.
4. CONCLUSIONS
Rapeseed straw xylan was converted to a thermoplastic and hydrophobic material through chemical modification. Xylan was first pre-activated by an esterification reaction with GMA, successfully introducing vinyl moieties as pendant groups along the xylan backbone. The resulting xylan macroinitiator was then decorated with octadecyl acrylate (ODA) polymer graft chains by interfacial free radical grafting-from polymerization. The grafting was confirmed by NMR and ATR-FTIR where the grafted xylan exhibited a melting endotherm in the interval (35 – 60 °C). Grafted and ungrafted xylan was blended (5% or 20%) with PCL through melt processing mini-extrusion. The biocomposites were molded into dog-bone shaped specimens for tensile testing by injection molding. Xylan had a stiffening effect on PCL. PCL blended with unmodified xylan resulted in heterogeneous and very brittle mixtures. The grafted xylan on the other hand showed good compatibility with the PCL matrix resulting in a homogeneous mixture. The Young’s modulus increased from 300 MPa (pure PCL) to 440 MPa (PCL with 20% of grafted xylan) and the ability to elongate to a strain-at-break > 500% was preserved.
ASSOCIATED CONTENT
Supplementary material: 1H-NMR spectra of ODA monomer and ODA polymer, 1H-NMR
spectra of the reaction between xylan and GMA in D2O, ATR-FTIR overlaid spectra of the
after grafting reaction, ATR-FTIR of the progressing reaction between Xyl-GMA and ODA,
1H-NMR and ATR-FTIR of XylB, DSC thermograms of XylB, DSC thermograms of the
melting endoterm of Xyl A-GMA reacted for 6 h or 20 h, TGA graph, and picture of the dog-bone biocomposites. The following files are available free of charge. Thermoplastic xylan_Supplementary_material.pdf
AUTHOR INFORMATION
Author Contributions
The manuscript was written through contributions of all authors. All authors jointly planned the work. AS performed all experimental work and analyses, except for SEM analyses performed by UE. All authors have approved the final version of the manuscript.
Funding Sources
The Swedish research council Formas, project number 2013-844.
Notes
The authors declare no competing financial interest.
5. ACKNOWLEDGMENTS
The authors thank The Swedish Research Council Formas (Project no. 2013-844) for their financial support. We thank Julien Navarro for valuable advice regarding polymerization. Xi Yang, Karin Adolfsson, Nejla Benyahia Erdal, and Giada Lo Re are thanked for their valuable advice regarding extrusion and injection molding. In addition, Martin Sterner is thanked for valuable advice on tensile testing.
6. REFERENCES
Anastasaki, A., Waldron, C., Nikolaou, V., Wilson, P., McHale, R., Smith, T., Haddleton, D. M., 2013. Polymerization of long chain [meth] acrylates by Cu (0)-mediated and catalytic chain transfer polymerisation (CCTP): high fidelity end group incorporation and modification. Polym. Chem. 4 (15), 4113-4119.
Belmokaddem, F., Pinel, C., Huber, P., Petit-Conil, M., Perez, D., 2011. Green synthesis of xylan hemicellulose esters. Carbohyd. Res., 346(18), 2896-2904
Ebringerová, A., Hromádková, Z., Heinze, T., 2005. Hemicellulose, in: Heinze, T. (Ed.), Polysaccharides I. Springer, Berlin Heidelberg, pp. 1-67.
Edlund, U., Ryberg, Y. Z., Albertsson, A. C., 2010. Barrier films from renewable forestry waste. Biomacromolecules, 11(9), 2532-2538.
Edlund, U., Albertsson, A.-C., 2014. A controlled radical polymerization route to polyepoxidated grafted hemicellulose materials. Polimery. 59 (1), 60-65.
Farhat, W., Venditti, R. A., Hubbe, M., Taha, M., Becquart, F., Ayoub, A., 2017. A review of water-resistant hemicellulose-based materials: processing and applications. ChemSusChem. 10 (2), 305-323.
Faruk, O., Bledzki, A. K., Fink, H. –P., Sain, M., 2012. Biocomposites reinforced with natrual fibers: 2000-2010. Prog. Polym. Sci. 37(11), 1552-1596.
Fundador, N., Enomoto-Rogers, Y., Takemura, A., Iwata, T., 2012. Acetylation and characterization of xylan from hardwood kraft pulp. Carbohydr. Polym., 87(1), 170-176.
Gabrielii, I., Gatenholm, P., 1998. Preparation and properties of hydrogels based on hemicellulose. J. Appl. Polym. Sci., 69(8), 1661-1667.
Gordobil, O., Egüés, I., Urruzola, I., Labidi, J., 2014. Xylan-cellulose films: improvement of hydrophobicity, thermal and mechanical properties. Carbohydr. Polym. 112, 56-62.
Gross, R. A., Kalra, B., 2002. Biodegradable polymers for the environment. Science. 297 (5582), 803-807.
Gupta, S., Madan, R.N., Bansal, M.C., 1987. Chemical composition of Pinus caribaea hemicellulose. Tappi J. 70(8), 113-116.
Habibi, Y., 2014. Key advances in the chemical modification of nanocelluloses. Chem. Soc. Rev. 43 (5), 1519-1542.
Hejna, A., Formela, K., Saeb, M. R., 2015. Processing, mechanical and thermal behavior assessments of polycaprolactone/agricultural wastes biocomposites. Ind. Crops Prod. 76, 725-733.
Herzele, S., Veigel, S., Liebner, F., Zimmermann, T., Gindl-Altmutter, W., 2016. Reinforcement of polycaprolactone with microfibrillated lignocellulose. Ind. Crops Prod. 93, 302-308.
Ibn Yaich, A., Edlund, U., Albertsson, A.-C., 2017. Transfer of Biomatrix/Wood Cell Interactions to Hemicellulose-Based Materials to Control Water Interaction. Chem. Rev. 12 (117), 8177-8207.
Jackson, P., Hunt, J., Doherty, P., Cannon, A., Gilson, P., 2004. Hydrophilicity of 3-D biomaterials: the Washburn equation. J. Mater. Sci.: Mater. Med. 15 (4), 507-511.
Kim, E. G., Kim, B. S., Kim, D. S., 2007. Physical properties and morphology of polycaprolactone/starch/pine‐leaf composites. J. Appl. Polym. Sci. 103 (2), 928-934.
Knill, C. J., Kennedy, J. F., 2003. Degradation of cellulose under alkaline conditions. Carbohydr. Polym. 51 (3), 281-300.
Laine, C., Harlin, A., Hartman, J., Hyvärinen, S., Kammiovirta, K., Krogerus, B., Pajari, H., Rautkosoki, H., Setälä, H., Sievänen, J., Uotila, J., Vähä-Nissi, M., 2013. Hydroxyalkylated xylans – Their synthesis and application in coatings for packaging and paper. Ind. Crops Prod. 44, 692-704.
Mohamed, R. M., Yusoh, K., 2015. A review on the recent research of polycaprolactone (PCL). Adv. Mater. Res. 1134, 249-255.
Shen, D., Gu, S., Bridgwater, A., 2010. The thermal performance of the polysaccharides extracted from hardwood: cellulose and hemicellulose. Carbohydr. Polym., 82(1), 39-45.
Siqueira, G., Bras, J., Dufresne, A., 2008. Cellulose whiskers versus microfibrils: influence of the nature of the nanoparticle and its surface functionalization on the thermal and mechanical properties of nanocomposites. Biomacromolecules. 10 (2), 425-432.
Socrates, G., 2001. Infrared and Raman characteristic group frequencies: tables and charts, third ed. John Wiley & Sons, England.
Svärd, A., Brännvall, E., Edlund, U., 2017. Rapeseed straw polymeric hemicelluloses obtained by extraction methods based on severity factor. Ind. Crops Prod. 95, 305-315.
Teleman, A., Lundqvist, J., Tjerneld, F., Stålbrand, H., Dahlman, O., 2000. Characterization of acetylated 4-O-methylglucuronoxylan isolated from aspen employing 1 H and 13 C NMR spectroscopy. Carbohydr. Res. 329(4), 807-815.
Teleman, A., Tenkanen, M., Jacobs, A., Dahlman, O., 2002. Characterization of O-acetyl-(4-O-methylglucurono) xylan isolated from birch and beech. Carbohydr. Res. 337(4), 373-377.
Testova, L., Roselli, A., Costabel, L., Kovasin, K., Tenkanen, M., Sixta, H., 2014. Combined production of polymeric birch xylan and paper pulp by alkaline pre-extraction followed by alkaline cooking. Ind. Eng. Chem. Res. 53(19), 8302-8310.
Voepel, J., Edlund, U., Albertsson, A. C., 2011. A versatile single‐electron‐transfer mediated living radical polymerization route to galactoglucomannan graft‐copolymers with tunable hydrophilicity. J. Polym. Sci., Part A: Polym. Chem. 49 (11), 2366-2372.
Voepel, J.; Edlund, U.; Albertsson, A.-C.; Percec, V., 2010. Hemicellulose-based multifunctional macroinitiator for single-electron-transfer mediated living radical polymerization. Biomacromolecules. 12 (1), 253-259.
Waldron, C., Anastasaki, A., McHale, R., Wilson, P., Li, Z.; Smith, T., Haddleton, D. M., 2014. Copper-mediated living radical polymerization (SET-LRP) of lipophilic monomers from
multi-functional initiators: reducing star–star coupling at high molecular weights and high monomer conversions. Polym. Chem. 5 (3), 892-898.
Werner, K., Pommer, L., Broström, M. 2014. Thermal decomposition of hemicelluloses. J. Anal. Appl. Pyrolysis, 110, 130-137.
Supplementary material for publication
Modified and thermoplastic rapeseed straw
xylan: A renewable additive in PCL
biocomposites
Antonia Svärd a , Elisabet Brännvall b , Ulrica Edlund a,*
a Fiber and Polymer Technology, KTH Royal Institute of Technology, Teknikringen 56,
SE-100 44 Stockholm, Sweden
b RISE Bioeconomy, Research Institutes of Sweden, Box 5604, SE-114 86 Stockholm, Sweden
Figure S1. 1H-NMR (in toluene-d
Figure S2. 1H-NMR spectra (in D
2O) of the progressing reaction between xylan and GMA at different reaction times.
Figure S3. ATR-FTIR overlaid spectra of the progressing reaction between xylan and GMA at different reaction times.
Figure S5. ATR-FTIR of the progressing grafting reaction between Xyl-GMA and ODA at different reaction times.
Figure S6. A) 1H-NMR (in DMSO-d
6 for XylB and XylB-GMA20h and toluene-d8) and B) ATR-FTIR spectra of Xyl B, XylB-GMA and XylB-GMA20h-ODA2h.
Figure S7. DSC thermograms of heating (right) and cooling (left) of (from top to bottom): XylB-GMA20h-g-ODA2h XylB-GMA20h and XylB.
Figure S8. DSC thermograms collected during heating of Xyl A-GMA reacted for 6 h or 20 h, and then grafted with ODA for 2 h.
Figure S9. TGA thermograms of XylA, XylA-GMA and XylA-GMA-g-ODA at different grafting polymerization times.
