Introduction
Carbon fibre (CF) composites are increasingly important for applications which require lightweight solutions (Frank et al.
2014). CFs are typically embedded in an epoxy or other polymer matrix to create CF-reinforced composite materials (Bekyarova et al.
2007; Meier
1995; Morgan
2005). CF composites are extensively used in aeronautics, aerospace, high-end sports equipment, and racing/sport and luxury cars (Glowacz
2015; Morgan
2005). The demand for CF composites is high, however the prohibitive costs have so far limited their use in mainstream automotive applications, construction and energy sector where their implantation would result in significant improvements regarding energy efficiency and reductions in CO
2 emissions. The cost of CF composites is largely based on the high price of CFs. They are currently produced predominately from polyacrylonitrile (PAN) with small amounts made from pitches, notably mesophase (Baker and Rials
2013). PAN is an expensive petroleum derived polymer which is typically processed via solution spinning into high quality filaments. Subsequent CF manufacture requires this starting fibre (precursor) to be gently heated through a series of ovens and furnaces which is a time and energy intensive process and contributes considerably to the total cost of the CF (Buckley and Edie
1993; Fitzer
1989; Windhorst and Blount
1997). Current research efforts are thus divided between finding alternative precursor materials and developing more efficient methods for their conversion into CF (Byrne et al.
2014b; Chand
2000; Frank et al.
2014).
Renewable polymers like cellulose and lignin are both regarded as excellent candidates for CF largely due to the lower cost of the starting material (Baker and Rials
2013; Dumanlı and Windle
2012; Kadla et al.
2002; Sadeghifar et al.
2016). Indeed both cellulose and lignin based CF have been commercially produced at one point in the past (Bacon and Tang
1964; Dumanlı and Windle
2012). The major limitation preventing largescale production of either cellulose or lignin differs. Upon pyrolysis of cellulose, the maximum carbon yield of 44.4% is rarely reached. Often, substantially lower yields are observed unless the precursor fibre is impregnated with catalysts promoting the dehydration reaction (Byrne et al.
2014a; Kandola et al.
1996; Zhou et al.
2016). In combination with the limited mechanical properties that were offered by the first generation of (viscose-type) continuous cellulose filaments, cellulose-derived CFs seemed little promising and were replaced rapidly by PAN in the 1960s (Jenkins and Kawamura
1976; Savage
2012). Lignin, on the other hand, has a mass yield of 55% comparable with PAN but due to its thermoplastic nature often requires extensive stabilization times, up to 100 h (Hosseinaei et al.
2016; Oroumei et al.
2015). This long stabilization time and challenges in the melt spinning of lignin without plasticising measures, and still limited mechanical properties of the resulting CF prevented large scale production in an economical and continuous way so far (Kadla et al.
2002; Luo et al.
2011; Mainka et al.
2015a,
b; Norberg et al.
2013; Qin and Kadla
2012; Xia et al.
2016). The limited mechanical properties of both cellulose and lignin based CF, however, are not considered to be an impeding factor for certain applications such as widespread automotive use and wind energy applications which do not require the > 3GPa tensile strength offered by PAN based CF (Redelbach et al.
2012). As such, research in the development of either cellulose or lignin as precursors has reignited. Here we report on a new approach aimed at addressing both the low mass and long stabilization times of cellulose and lignin respectively by developing a composite fibre containing varying amounts of cellulose and lignin.
Using a recently developed solvent-based spinning technique, it was possible to spin continuous cellulose filaments with high molecular orientation and high mechanical properties (Hummel et al.
2015; Sixta et al.
2015). It was found, that the same solvent, 1,5-diazabicyclo[4.3.0]non-5-ene-1-ium acetate ([DBNH]OAc), is also capable of dissolving lignin of various origin (Ma et al.
2015). This opened up an entirely new approach to the utilization of lignin. Melt spinning of lignin requires a glass transition temperature below the decomposition temperature. This is often only possible through derivatization of the respective lignin (Kubo et al.
1998; Steudle et al.
2017; Uraki et al.
1995; Zhang and Ogale
2014) or by adding plasticizing agents (Kadla et al.
2002; Kubo and Kadla
2005; Saito et al.
2012). Another challenge encountered with lignin is that the polymer might differ substantially in its macromolecular properties and quality depending on the source and pulping method. Evidently, this complicates spinning of homogenous precursor filaments.
On the other hand, cellulose is available in high purity and several techniques have been established to spin continuous filaments. For further conversion into carbon fibres, non-derivatizing routes, i.e. direct dissolution of cellulose and coagulation in an anti-solvent spin bath is preferred as it produces filaments of high mechanical properties and round cross section. [DBNH]OAc is a powerful direct solvent for all wood-derived polymers. Upon co-dissolution of cellulose and lignin, the long-chain carbohydrate polymer predominately defines the visco-elastic properties of the resulting bi-polymer solution. This means, that the quality and macromolecular properties of the respective lignin is overcompensated by cellulose. The molecular weight distribution and composition of the lignin source become minor factors. Thus, a very broad set of lignin types can be used. It was possible to spin fibres with up to 50 wt.% lignin (Ma et al.
2015).
In this study, filaments with 20 and 40 wt.% organosolv lignin as model substance were investigated. Stabilization of the cellulose-lignin fibres was studied at 3 different temperatures and the fibres were characterized thoroughly in terms of thermal degradation using FTIR to determine the stabilization kinetics, evolved gases, single fibre measurements and scanning electron microscope.
Experimental
Materials
Birch (
Betula pendula) prehydrolysis kraft pulp ([η] = 476 ml/g, DP = 1133, M
n = 65.9 kDa, M
w = 269.3 kDa, polydispersity 4.1, Enocell Speciality Cellulose, Finland) was delivered in sheet form and cut to a powder by means of a Willey mill. Beechwood organosolv lignin was received from Fraunhofer Institutes, Germany. 1,5-diazabicyclo[4.3.0]non-5-ene (DBN, 99%, Fluorochem, UK) and acetic acid (glacial, 100%, Merck, Germany) were used to synthesize [DBNH]OAc. The lignin-cellulose fibres where dry jet spun from a solution in [DBNH]OAc as 60-filament tow with a draw ratio of 5. Pure cellulose filaments were spun from a 13 wt.% solution. Filaments with 20 wt.% lignin (relative to cellulose) were spun from a solution with 15 wt.% polymer concentration, 40 wt.% lignin fibers from 18 wt.% polymer solution, respectively. Details regarding the synthesis of the ionic liquid, solution preparation and spinning equipment have been described previously (Ma et al.
2015). The composition of the fibres used in this study with their sample codes are summarized in Table
1.
Table 1
Cellulose-lignin precursor fibres used in this study
100C | 100% cellulose |
2080LC | 20% Lignin, 80% Cellulose |
4060LC | 40% Lignin, 60% Cellulose |
Stabilization
Samples were heated in air using a Thermotec 2000 laboratory type oven at different temperatures ranging from 200 to 280 °C and times varied from 30 to 300 min. A fixed heating rate of 8 °C/min was used to raise the oven temperature from ambient temperature to the desired set temperature. The pre-treatment times were recorded after the temperature reached the set value. The effect of applying tension on stabilization was investigated by mounting the fibres on a stainless steel rig hanging with known weights varying from 17 to 34 N/m per fibre.
Thermogravimetric analyses
Thermal stability of the lignin-cellulose fibres were measured using thermogravimetric analyses (TGA) on a TA Q50 TGA thermogravimetric analyser. Measurements were performed using 5–8 mg of the samples. The fibres were heated from 30 to 600 °C at a heating rate of 10 °C/min under N
2 atmosphere. The thermal degradation temperature at which the weight loss begins (T
d) was calculated as the onset. The activation energy of pyrolysis (E
k) for fibres were calculated by the Kissinger method using the below equation (Blaine and Kissinger
2012). According to the equation, where
β
i
is the heating rate,
T
pi
is the endothermic peak temperature,
A
k
is the Arrhenius pre-exponential factor,
R is the gas constant (8.314 J/mol K) and
E
k
is the activation energy. The sample was heated at different heating rates (
β) of 5, 10, 15 and 20 °C/min and the endothermic peak temperature was recorded. The data plot of ln (
β/T
m
2
) versus 1/
T
m
was fitted with a linear trend line where the
E
k
was calculated from the slope of the line.
$$ \ln \left( {\frac{\beta }{{T_{m}^{2} }}} \right) = \ln \left( {\frac{{A_{k} R}}{{E_{k} }} } \right) - \frac{{E_{k} }}{R} \frac{1}{{T_{m} }} $$
FTIR and STA-FTIR
FTIR spectra of the fibres were measured on Bruker LUMOS FTIR microscope. A thin layer of fibres was mounted on a glass slide by means of double-sided tape. The fibre layer was scanned in the frequency range of 600–4000 cm−1 at a scan resolution of 4 cm−1 with a background and sample scan time of 64 scans. FTIR measurements were repeated 6 times per sample.
The evolved gas analysis during pyrolysis of fibres was carried out using hyphenated simultaneous thermal analysis-Fourier transform infrared spectroscopy (STA-FTIR). Here, the simultaneous thermal analyser (STA 8000, Perkin Elmer) was coupled with a FT-IR spectrometer (Frontier, Perkin Elmer) via transfer line hyphenation using a TL9000 interface. Measurements were performed using approximately 3.00 mg of the samples. The specimens were heated from 30 to 600 °C at a heating rate of 10 °C/min under N2 atmosphere. STA data analysis was performed using the Pyris software (version 11.1.1.0492). The gasses evolved were immediately transferred to the FTIR detector through the transfer lines (balanced flow evolved gas analyser, TL9000) to continuously monitor the evolved gasses during pyrolysis. The FT-IR data were collected in the range of 4000–600 cm−1 at a resolution of 4 cm−1. Data collection and analysis was performed using Spectrum TimeBaseTM (version 3.1.3.0042) and Spectrum (version 10.4.4.449) for the time resolved IR data.
Single fibre measurements
Tensile tests of the fibre specimens were measured using an Instron Tensile Tester fitted with a 5.00 N load cell. All tests were conducted at a fixed gauge length of 20.00 mm and at a controlled extension rate of 1.5 mm/min. The instrument was programmed to apply a pre-load of 0.10 cN before recording load-extension data. Tensile test measurements were repeated 5 times for each sample. The tensile stress of the fibres was calculated from dividing the load by the fibre cross section area, units in MPa. The cross sectional area was determine by means of an optical microscope. The Young’s modulus was obtained from measuring the gradient of the elastic region of the stress–strain curve, units in GPa. All samples were conditioned at 20 ± 2 °C and 65 ± 2% RH for 24 h prior to testing.
Scanning electron microscope
The morphologies of the fibre cross sections were visualized with a Zeiss Supra 55VP scanning electron microscope (SEM) at an accelerating voltage of 5.00 kV. For cross section images, fibre samples were submerged in liquid nitrogen for nearly 5 min to ensure they were completely frozen. They were then removed from the liquid nitrogen bath and immediately snapped using a scalpel blade. The surface and the cleaved edge of the fibres were gold coated before observation.