Department of Physics, Chemistry and Biology
Master’s Thesis
Evaluation of the Carbonization of Thermo-Stabilized
Lignin Fibers into Carbon Fibers
Henrik Kleinhans
5th July 2015
LITH-IFM-A-EX--15/3112--SE
Linköping University Department of Physics, Chemistry and Biology 581 83 Linköping
Department of Physics, Chemistry and Biology
Evaluation of the Carbonization of Thermo-Stabilized
Lignin Fibers into Carbon Fibers
Henrik Kleinhans
Thesis work performed at INNVENTIA AB
5th July 2015
Supervisors
Professor Lennart Salmén, INNVENTIA AB
Daniel Aili, Linköping University
Examiner
Dr. Thomas Ederth, Linköping University
Linköping University Department of Physics, Chemistry and Biology 581 83 Linköping
Abstract
Thermo-stabilized lignin fibers from pH-fractionated softwood kraft lignin were carbonized to various temperatures during thermomechanical analysis (TMA) under static and increasing load and different rates of heating. The aim was to optimize the carbonization process to obtain suitable carbon fiber material with good mechanical strength potential (high tensile strength and high E-modulus). The carbon fibers were therefore mainly evaluated of mechanical strength in Dia-Stron uniaxial tensile testing.
In addition, chemical composition, in terms of functional groups, and elemental (atomic) composition was studied in Fourier transform infrared spectroscopy (FTIR) and in energy-dispersive X-ray spectroscopy (EDS), respectively. The structure of carbon fibers was imaged in scanning electron microscope (SEM) and light microscopy. Thermogravimetrical analysis was performed on thermo-stabilized lignin fibers to evaluate the loss of mass and to calculate the stress-changes and diameter-changes that occur during carbonization.
The TMA-analysis of the deformation showed, for thermo-stabilized lignin fibers, a
characteristic behavior of contraction during carbonization. Carbonization temperatures above 1000°C seemed most efficient in terms of E-modulus and tensile strength whereas rate of heating did not matter considerably. The E-modulus for the fibers was improved significantly by slowly increasing the load during the carbonization. The tensile strength remained however unchanged.
The FTIR-analysis indicated that many functional groups, mainly oxygen containing,
dissociate from the lignin polymers during carbonization. The EDS supported this by showing that the oxygen content decreased. Accordingly, the relative carbon content increased
passively to around 90% at 1000°C. Aromatic structures in the carbon fibers are thought to contribute to the mechanical strength and are likely formed during the carbonization.
However, the FTIR result showed no evident signs that aromatic structures had been formed, possible due to some difficulties with the KBr-method.
In the SEM and light microscopy imaging one could observe that porous formations on the surface of the fibers increased as the temperature increased in the carbonization. These formations may have affected the mechanical strength of the carbon fibers, mainly tensile strength.
The carbonization process was optimized in the sense that any heating rate can be used. No restriction in production speed exists. The carbonization should be run to at least 1000°C to achieve maximum mechanical strength, both in E-modulus and tensile strength. To improve the E-modulus further, a slowly increasing load can be applied to the lignin fibers during carbonization. The earlier the force is applied, to counteract the lignin fiber contraction that occurs (namely around 300°C), the better. However, in terms of mechanical performance, the lignin carbon fibers are still far from practical use in the industry.
Abbreviations
DLaTGS Deuterated L-alanine doped Triglycene Sulphate DSC Differential Scanning Calorimetry
EDS Energy Dispersive X-ray Spectrometer E-modulus Elastic modulus, Young’s modulus FTIR Fourier Transform Infrared Spectroscopy
G Guaiacyl
GC Gas chromatograph
H P-hydroxyphenyl (P=para)
IR Infrared
Mn Number Average Molecular Weight
MPP Mesophase pitch
Mw Average Molecular Weight
PAN Polyacrylonitrile PDI Polydispersity Index
S Syringyl
SEM Scanning Electron Microscope
TE Thermo Electric
Tg Glass Transition Temperature
TGA Thermogravimetric Analysis TMA Thermomechanical Analysis
Ts Softening Point
v Version
i Contents 1 Introduction ... 1 1.1 Aim ... 1 1.2 Background ... 1 1.2.1 Polymer properties ... 3 1.2.2 Carbon fibers ... 4
1.2.2.1 PAN-based carbon fibers ... 4
1.2.2.2 Pitch-based carbon fibers ... 5
1.2.2.3 Structure of carbon fibers ... 6
1.2.3 Lignin polymer ... 7
1.2.3.1 Refining of lignin ... 10
1.2.3.2 Lignin-based carbon fibers... 10
1.2.3.3 Production of lignin carbon fiber ... 11
2 Experimental ... 15
2.1 Stabilization ... 15
2.2 Carbonization and Thermomechanical Analysis (TMA) ... 16
2.2.1 Principles ... 16
2.2.2 Method ... 17
2.3 Uniaxial Tensile Testing (with Dia-Stron) ... 20
2.3.1 Principles ... 20
2.3.2 Method ... 20
2.4 Thermogravimetric Analysis (TGA) ... 22
2.4.1 Principles ... 22
2.4.2 Method ... 22
2.5 Fourier Transform Infrared Spectroscopy (FTIR) ... 23
2.5.1 Principles ... 24
2.5.2 Method ... 25
2.6 Scanning Electron Microscope (SEM) ... 27
2.6.1 Principles ... 27
2.6.2 Method ... 28
2.7 Light Microscopy ... 29
3 Results ... 30
ii
3.1.1 Static measurement ... 30
3.1.2 Measurements with increasing force ... 35
3.2 Dia-stron ... 39
3.2.1 Evaluation of temperature and heating rate ... 39
3.2.2 Evaluation of applied force during carbonization ... 45
3.3 TGA ... 47
3.4 FTIR ... 52
3.5 SEM ... 60
3.5.1 Imaging ... 60
3.5.2 Energy-dispersive X-ray spectroscopy ... 64
3.6 Light Microscopy ... 69
4 Discussion ... 74
4.1 Analysis of the main results ... 74
4.1.1 TMA ... 74 4.1.2 Dia-Stron ... 74 4.1.3 TGA ... 76 4.1.4 FTIR ... 76 4.1.5 SEM ... 78 4.1.6 Microscopy ... 78
4.2 Impact in a broad sense ... 78
4.3 Ethical implications in a broad sense ... 79
4.4 Future perspectives ... 79 5 Conclusion ... 81 6 Acknowledgements ... 82 7 References ... 83 8 Appendix ... 87 8.1 Process ... 87 8.1.1 Timetable ... 88 8.1.2 Planning ... 90 8.1.3 Process results ... 91
8.1.4 Comprehensive analysis of the process ... 95
8.2 Data ... 95
iii
8.2.2 TGA ... 96
8.2.2.1 Calibration ... 96
8.2.2.2 Calculations of diameter and stress change ... 98
8.2.3 Dia-stron ... 98
8.2.3.1 Compliance ... 98
8.2.3.2 Data of carbonized fiber samples under load ... 105
8.2.4 Light microscopy ... 106
8.2.4.1 Carbonization series ... 106
1
1 Introduction
Aim and background of the project are discussed in this section. All information regarding the process is found in appendix 8.1.
1.1 Aim
The aim of this project was to optimize the carbonization process of lignin to obtain suitable carbon fiber material with good mechanical properties (high tensile strength and
high E-modulus).
In order to optimize the process the following parameters were being looked at:
Temperature
Rate of temperature elevation
Applied force
Several tests were conducted to provide information about mechanical properties as well as structural formation (visual morphologies and chemical structures), for each setting of parameters, to determine the optimal carbonization.
The elemental composition of carbon fibers was identified by energy dispersive X-ray spectrometer (EDS). To be considered ‘carbon fibers’, the content of carbon must be at least 90 % of the weight.
The structure and morphology of the carbon fibers was studied in light microscopy and scanning electron microscopy (SEM). Defects and irregularities affect the tensile strength. Structures like skin-core may also affect strength properties.
The change in molecular structure was studied, mainly by looking at disappearing functional groups and molecules in the carbonization, by Fourier transform infrared spectroscopy (FTIR). The loss of functional groups, during carbonization, should result in a graphite skeleton, with little to none non-carbon compounds.
Mechanical properties, e.g. stress, strain, elastic modulus and tensile strength, of the carbon fiber were tested using uniaxial tensile testing.
1.2 Background
Carbon fibers are defined by the relatively high content of carbon which must be at least 90% wt. (Liu, Kumar 2012)
Carbon fibers have remarkable properties and are often used in composite materials. The mechanical properties are characterized by high tensile strength (2-7 GPa), good compressive strength (up to 3 GPa) and high tensile modulus (200-900 GPa). The density (1.75-2.18 g/cm3) is 4 times lower than steel, whereas the strength of the carbon fiber is much higher. (Liu, Kumar 2012)
2
The high modulus of carbon fibers can be explained by high crystallinity and high alignment of crystals in the fiber direction. The tensile strength of carbon fibers is however affected by defects and crystalline morphologies in the carbon fibers. (Huang 2009)
Additional attributes are good temperature resistance, low thermal expansion, excellent electrical and thermal conductivity and good chemical resistance.(Liu, Kumar 2012) The thermal as well as electric conductivity is due to the high content of delocalized π electrons and the parallel alignment of graphene layers along the fiber axis. The thermal conductivity coefficient is close to that of metals. (Huang 2009)
The market for carbon fibers has been growing steadily over the past 20 years, an average increase of 12% per year. It is estimated to be a 1.6 billion US dollar market by 2017. Some notable manufacturers of carbon fiber are AKSA (AKSACA), Cytec (Thornel®), Formosa Plastics (Tairyfil), Hexcel (HexTowR®,HexForceR®), Mitsubishi Rayon (PYROFIL™), SGL (SIGRAFILR®©), Teijin (TENAX®), Toho Tenax (TenaxR®), Toray
(TORAYCAR®), and Zoltek (PANEXR® ). (Liu, Kumar 2012)
Carbon fibers are today exclusively used in certain types of industry such as aerospace (e.g. aircraft and space systems), military, automobile and sporting goods, to name a few. (Huang 2009) The benefits of using carbon fibers other than using it as a strong composite is to decrease weight in for example vehicles. Lower weight also means lower fuel consumption. Composites reinforced with carbon fibers could offer as much as 60 % part weight reduction. The cost would, however, be ten times higher, due to the expensive raw material of carbon fibers. (Baker, Rials 2013)
In today’s commercial carbon fiber production, polyacrylonitrile (PAN) is the most
domination precursor used. More than 90 % of commercial carbon fibers are manufactured from PAN precursors. The rest are made from pitch, foremost mesophase pitch (MPP). (Liu, Kumar 2012)
Both PAN and pitch precursors are derived from raw fossil materials, such as petroleum or oil. Therefore, the manufacturing cost of carbon fiber is highly dependent on the oil price. In other words, carbon fibers are not cheap. As for the total cost of PAN-based carbon fibers, 51% stands for the cost of the precursor, 18% for utilities, 12% depreciation, 10% labor and the rest is represented by other fixed costs. (Baker, Rials 2013)
Lignin is on the other hand an abundant, cheaper and renewable material found throughout the nature. Around 30 % of the content in wood is lignin and it can be extracted in bio refineries during the pulping process, kraft pulp is often used. (Sjöström 1993) Lignin is a complex polymer that can be used as a cheap precursor alternative for carbon fiber manufacturing. The interest of manufacture carbon fibers out of lignin has been growing over the recent years, as depicted by Figure 1. It illustrates the rising frequency of published articles since the year 1964. (Baker, Rials 2013) But the idea of producing carbon fibers from lignin is nothing new. In the early 1970’s carbon fiber production from lignin had a brief period of small-scale commercialization under the name Kayacarbon by the Nippon Kayaku. Unfortunately, the
3
Kayacarbon could not compete with other precursor materials and had to leave the market due to the high competition. (Baker, Rials 2013)
Figure 1: Frequency of journal articles (dark) and patents (light) about lignin-based carbon fibers by year to 20130131. (Baker, Rials 2013)
Even today, the tensile strength and modulus do not meet the requirement for proper use in the industry. PAN and MPP are still superior regarding strength and mechanical properties. (Baker, Rials 2013)
The standards, set in the automotive industry, require carbon fibers to have tensile strength of 1.72 GPa, and E-modulus of 172 GPa. (Baker, Rials 2013) Current modified lignin carbon fibers, with average strength of 1.07 GPa, E-modulus of 82.7 GPa and extensibilities of 2.03%, do not meet these criteria. (Baker, Rials 2013)
Because of this, much of the lignin carbon fiber research has been focused on improving the mechanical properties. (Baker, Rials 2013)
1.2.1 Polymer properties
Polymers are large macromolecules composed of repeated subunits called monomers. They can be divided into two groups: natural and synthetic polymers. PAN is an example of synthetic polymer whereas lignin is a natural polymer. (McCrum, Bucknall & Buckley 1997) The polymer size and length is often expressed as molecular weight. However,
polymerization yields a product with a wide range of molecular weights. Therefore it is preferred to express the weight statistically in order to describe the distribution of polymer weights in the product. Average molecular weight and number average molecular weight is denoted Mw and Mn, respectively. The ratio, Mw/Mn, between those is termed polydispersity
index (PDI). (McCrum, Bucknall & Buckley 1997)
Often considered as amorphous material, due to the lack of crystallinity order, polymers exhibit a glass transition temperature. The glass transition temperature (Tg) is commonly
defined as the inflection point (onset) of the slope in a differential scanning calorimetry
(DSC) plot or in a dilatometry plot where the polymer material transform from a glassy (solid) to a rubbery state when the temperature is increased. Unlike idealistic solid materials a
distinct melting temperature cannot be determined for polymers. Instead of a direct phase transition from solid to liquid at a critical temperature, the viscosity of amorphous polymers continues to decrease upon heating as the molecules gain movement. This defines the rubbery state. (McCrum, Bucknall & Buckley 1997)
4
The Tg is impacted by the length of the polymer chain. Along with viscosity the Tg tend to
increase with increasing length. (McCrum, Bucknall & Buckley 1997)
Polymers possess viscoelastic properties and exhibit hysteresis during tensile testing.
1.2.2 Carbon fibers
Carbon fibers derived of PAN and MPP precursors dominates the current market. Although they differ in composition the production of carbon fibers is fairly similar.
Extrusion and spinning of polymers into fibers are made after the preparations of the precursors. The fibers are then stabilized by thermo-oxidation, in oxidative environment (oxygen rich gas/atmosphere), transforming it from a thermoplastic into a thermoset material. The final step is carbonization where the stabilized fiber is heated in high temperatures. (Huang 2009)
1.2.2.1 PAN-based carbon fibers
PAN, containing 68% carbon, are made from acrylonitrile (AN) through free radical polymerization (Bajaj, Sreekumar & Sen 2001), with the use of initiators such as peroxides and azo-compounds (Huang 2009).
In the stabilization process, the PAN polymers undergo cyclization, dehydrogenation and finally oxidation. The cyclization is preferred to happen before the oxidative treatment since the oxidation reaction prefers to take place with cyclized PAN chains. (Liu, Kumar 2012) Although the mechanism involved in the cyclization is not fully understood, different models of the final cyclized structure have been proposed (Figure 2). (Huang 2009) Furthermore, the PAN fibers must be stabilized under tension/stretching of the fiber. Otherwise the fiber length will shrink in the process. (Liu, Kumar 2012)
Figure 2: Linear PAN chains are proposed to cyclize into the right hand structure.
The stretching is also often used during the carbonization as well. However, the importance is still under debate. (Huang 2009) Since the stabilized fiber is considered a thermoset, the dimensions should not be affected by temperature changes at this point.
The strength of a carbon fiber is dependent on the carbonization temperature and the maximum strength is observed at around 1,500°C. (Huang 2009)
In the early carbonization, crosslinking reactions take place in the oxidized PAN and the cyclized structure starts to link up sideways during dehydration and denitrogenation. A planar structure is formed along the fiber axis (Figure 3). (Huang 2009)
5
Figure 3: A proposed model of crosslinking and formation of PAN-polymers into graphite structure through dehydrogenation and denitrogenation.
The oxidative stabilization is controlled by oxygen diffusion through the fiber. Due to the insufficient diffusion through the entire fiber a skin-core structure is often formed on the surface after stabilization. The oxygen content in this structure is higher and the form is more compact and ordered than the core. The skin-core in stabilized fibers is also present in the final carbon fiber. (Liu, Kumar 2012)
The excellent mechanical properties of the PAN-based carbon fibers can be explained by intermolecular interactions between the polar nitrile groups (placement shown in Figure 2) in PAN chains. (Chatterjee et al. 2014) The interaction also contributes to a high melting point and the polymer tends to degrade before melting, Therefore melt-spinning is not suitable for PAN and wet-spinning is thus preferred. (Huang 2009)
Today, conventional PAN precursors provide carbon fiber with tensile strengths of 1.03 GPa. (Baker, Rials 2013)
1.2.2.2 Pitch-based carbon fibers
Pitch precursors contains 80% carbon, are isotropic and can be produced in two ways: either by destructive distillation of petroleum and coal or pyrolysis of synthetic polymers. The former is attractive in commercial perspective where petroleum pitch is preferred over coal. Coal derived pitch filaments break more easily during extrusion and thermal treatments due to the higher content of particles. (Huang 2009)
Another common way to produce carbon fibers is by the use of anisotropic pitch, such as mesophase pitch. Mesophase means that the material contains a liquid crystalline phase. In addition, mesophase pitch also contains appreciable amount of anisotropic and isotropic phase. (Liu, Kumar 2012) The mesophase pitch is obtained from heating coal or petroleum at specific temperatures, or by transforming isotropic pitch. (Huang 2009)
The isotropic pitch contains thousands of aromatic hydrocarbons that form spheres and larger aromatic rings during heat treatment. Further heat treatment fuses the spheres together, resulting in a formation with large anisotropic domains combined with isotropic parts, i.e. mesophase. (Matsumoto 1985)
6
The aromatic ring network in the mesophase pitch structure is stacked through π-electron interaction. Due to this structure, a highly oriented basal plane structure can be achieved through carbonization. (Matsumoto 1985)
The graphitic structure in mesophase pitch-based (MPP) carbon fibers gives a higher modulus than PAN-based carbon fibers. Due to the amount of anisotropic phase, and in contrast to isotropic pitch, mesophase pitch exhibit much higher tensile properties. (Liu, Kumar 2012) Although mesophase pitch-based carbon fibers have a higher modulus and elastic modulus, its compressive and tensile strength is lower than PAN. (Liu, Kumar 2012, Huang 2009)
The processing of isotropic and mesophase pitch precursor fibers does not differ much from the conventional processing of PAN precursor. Melt-spinning, thermo-stabilization and carbonization are the same. (Huang 2009) The reactions in pitch during the stabilization process include oxidation, dehydrogenation, cyclization, elimination, condensation and cross-linking. (Matsumoto, Mochida 1992)
The three-dimensional crosslinking in pitch based carbon fibers are accounted by the introduction of oxygen containing groups and the formation of hydrogen bonding between molecules. (Huang 2009)
Stabilization of MPP, as well as PAN, has the challenge of achieving homogenous
stabilization throughout the entire fiber. Skin-core structures have been observed in MPP, just like in PAN. The structure prevents oxygen from penetrating deeper into the structure. This is thought to be caused by the competition of oxidative reactions on the surface of the fiber and the diffusion of oxygen. (Lü et al. 1998, Brodin et al. 2012)
In the pitch based carbon fibers sheet-like morphology can be observed, where the sheets are tens of nanometers thick while the length is considerably longer, 10-100 microns. (Liu, Kumar 2012)
1.2.2.3 Structure of carbon fibers
The layer planes in carbon fibers may be either in turbostratic, graphitic, or a hybrid form depending on precursor and manufacturing process. (Huang 2009)
In graphitic structure, the planes of layers are stacked parallel in a regular and crystalline fashion (Figure 4). The covalent double bonds between carbon atoms are sp2 hybridized while the sheets are held together by relatively weak Van der Waals forces and π-stacking. (Steed, Turner & Wallace 2007) The distance between two graphene layers is about 0.335 nm. (Huang 2009)
In contrast, the graphene sheets are stacked irregular or haphazardly folded, tilted or split in turbostratic structure (Figure 4). The bonds are instead sp3 hybridized which increase the spacing to 0.344 nm. (Huang 2009)
7
Due to the crystallites in MPP carbon fibers tend to be more graphitic and the spacing is in the range of 0.337–0.340 nm. PAN carbon fibers contain mainly turbostratic crystals. (Huang 2009)
Figure 4: Turbostratic and graphitic ordering. 1.2.3 Lignin polymer
Lignin is present in all vascular plants with few exceptions. In fact, lignin is the second, to cellulose, most abundant bio-macromolecule in the world. It is mainly located in the
secondary cell wall of plant cells where it provides stability and support to the plant.(Becker 2008) It also plays a vital role in transportation of water and protects the cell wall from harmful enzymatic degradation. (Gellerstedt, Ek & Henriksson 2009)
The lignin macromolecule is biosynthesized in the plant from three different monolignol monomers: p-coumaryl alcohol, coniferyl alcohol, and sinapyl alcohol. They are incorporated and linked together into the lignin polymer as p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S), respectively (Figure 5). (Freudenberg 1959, Baker, Rials 2013)
Figure 5: The 3 building components of lignin.
R1=OMe, R2=H: Coniferyl alcohol
R1=R2=OMe: Sinapyl alcohol
R1=R2=H: p-Coumaryl alcohol
R1=OMe, R2=H: Guaiacyl unit
R1=R2=OMe: Syringyl unit
8
Lignin polymerization is initiated by oxidation of the phenyl hydroxyl group of the
monolignols by enzymatic dehydrogenation, forming resonance-stabilized free radicals. The radicals then undergo radical coupling reactions, producing dimers called dilignols connected through a variety of linkages (Table 1). The dilignols continue to link to each other in an end-wise polymerization. (Freudenberg 1959)
Table 1: Linkages found in softwood lignin. Linkage
Type
Dimer Structure Percent of Total
Linkages (%)
β-O-4 Phenylpropane β−aryl
ether 45-50 5-5 Biphenyl and Dibenzodioxocin 18-25 β -5 Phenylcoumaran 9-12 β-1 1,2-Diaryl propane 7-10
α-O-4 Phenylpropane α−aryl
ether
6-8
4-O-5 Diaryl ether 4-8
β - β β-β-linked structures 3
Lignin polymers are extracted and separated during the pulping process in biorefineries, mainly from wood. The main component in wood, along with cellulose and hemicellulose, is namely lignin. (Sjöström 1993)
The composition of lignin may vary depending on the plant species. Wood can be divided into two main categories: softwood (gymnosperm, i.e. evergreen trees) and hardwood
(angiosperm, i.e. deciduous trees). Gymnosperms produce their seeds without covering and just release the seeds as they are, whereas the seeds of angiosperms are protected with a covering. Ex. Apple seeds are protected inside the apple, thus apple trees belong to angiosperm. (HowStuffWorks 2001, Mongeau, Brooks 2001)
Softwood lignin is composed almost entirely of G with small amounts of H. An example of chemical structure can be seen in Figure 6. In contrast, hardwood lignin have a blend of G, S and very little of H. (Sjöström 1993)
The syringyl in addition to guaiacyl units contributes to a more linear structure in hardwood. Whereas softwood lignin is more branched and/or cross-linked. (Kubo et al. 1997)
The level of crosslinking has an effect on properties such as glass-transition temperature (Tg).
Comparing lignin of softwood and hardwood with the same molecular weight and
distribution, the branched and/or cross-linked nature of softwood lignin results in a slightly higher Tg compared to that of hardwood. (Baker, Rials 2013)
9
Figure 6: A possible structure of softwood lignin polymer. Some of the three most occurring linkages and structures are indicated by blue circles. (Gellerstedt, Ek & Henriksson 2009)
β-O-4
β-5
Dibenzodioxocin structure
10 1.2.3.1 Refining of lignin
Lignin is produced and recovered as a byproduct in the bio refinery industry, i.e. in the paper pulping process or the cellulosic ethanol fuel production. As for today, Kraft pulping is the most commonly used process to extract lignin. (Huang 2009) Along with hemicellulose lignin is dissolved and separated from cellulose fibers into a so called black liquor solution using strong alkali reagents and sodium sulfite catalysts. (Sjöström 1993) Therefore lignin molecules often contain traces of sulfur.
Considered waste, lignin is used to supply with low-value energy and is simply burned in the refinery. Burning the black liquor also regenerates inorganic substances. (Baker, Rials 2013) The lignin can be extracted from black liquor through pH fractioning. The lignin molecules precipitate by lowering the pH, in the alkali solution, below 10 pH. (Helander et al. 2013) Organosolv, which is a cleaner alternative to kraft pupling, can also be used to fractionate biomass feedstock into the different lignocellulosic components. (Baker, Rials 2013)
Developed by Innventia AB, Lignoboost is another process of lignin extraction. It is designed to work alongside traditional kraft pulping and allows for high recovery of high purity lignin. The process involves the treatment of industrial black liquor with carbon dioxide which causes precipitation of lignin, which is slurried, washed with dilute acid and recovered. (Baker, Rials 2013)
Through cleaner and purer fractioning the biofuel market has become competitive with oil refining and several products have even been identified as direct replacement for petroleum derived chemicals. That includes: polymer blends, epoxy resins, conduction polymers and antioxidants. The possibility to manufacture carbon fibers from lignin is of particular interest. (Baker, Rials 2013)
1.2.3.2 Lignin-based carbon fibers
Lignin carbon fibers do have advantages over other precursors (MPP and PAN) in the sense of cost and that lignin is a renewable material. (Baker, Rials 2013) Another feature of lignin is that it is already substantially oxidized, which might be beneficial in the carbon fiber
manufacturing. (Baker, Rials 2013)
The best lignin-based carbon fiber samples produced to date have an average strength of 1.07 GPa, modulus of 82.7 GPa, and extensibilities of 2.03%. (Baker, Rials 2013) As mentioned above, these values are, however, inferior to those of PAN and MPP.
The mechanical strength of lignin carbon fibers need to be increased. It is however restricted by the unavailability of suitable and pure lignin. Most industrial lignin contains significant amounts of impurities, which may induce inconsistent rheological and ill-defined properties that hinder processing. (Chatterjee et al. 2014) (Lignoboost is a workaround/solution that produces pure lignin) The lignin needs to meet certain requirement to allow multifilament melt spinning and conversion trials. For instance, the melt-spinning requires a relatively low Tg whereas the stabilization requires high Tg and this causes complications. In the economic
11
perspective of production, the fibers need to be converted into carbon fibers rapidly and at low expenses. (Baker, Rials 2013)
Processing conditions such as stabilization and carbonization have a great influence on the mechanical properties of carbon fibers and should thus be considered to improve the fiber quality. (Baker, Rials 2013)
Furthermore, narrow molecular weight distribution would ensure uniform increases in molecular weight throughout the material during oxidative thermostabilization, and
consequently provide for a more beneficial isotropic/uniform structure during carbonization. (Baker, Rials 2013)
1.2.3.3 Production of lignin carbon fiber
The process of manufacturing carbon fibers from lignin polymers is similar to that of PAN and MPP.
After the extraction and pre-processing, lignin is usually melt spun due to its thermoplastic properties. Pure softwood can be dry or wet spun. (Baker, Rials 2013) Then the produced fiber undergoes thermostabilization to convert it to a thermoset. In the last step non-carbon elements are removed as the fiber is carbonized. (Norberg et al. 2013)
During spinning extrusion, the temperature is often maintained at approximatively 200°C, and in the following stabilization step, the temperature should be between 200-300°C and lastly the carbonization is performed at temperatures higher than 1000°C. (Brodin, Sjöholm & Gellerstedt 2010)
The process flow is illustrated in Figure 7.
Figure 7: Process flow of lignin into carbon fibers. The energy use is different in each of the steps. Carbonization and graphitization tend to take a lot of energy.
It has been stated that increased alkoxy contents and/or carbon contents enhance lignin-based carbon fibers. (Baker, Rials 2013) Accounting for this, softwood lignin has a better potential
12
in carbon fiber manufacturing than hardwood, since its carbon yield is higher. However, the more cross-linked nature of softwood does not allow melt spinning without prior
modification, e.g. blending lignin polymers with plasticizers or softening agents. (Huang 2009)
Thermostabilization
In the thermostabilization of lignin, the sample is often treated in oxidative environment and the temperature is ramped from around 200°C to above 300°C. It is important to maintain the temperature below the Tg so that the integrity of the polymer is kept unchanged and does not
start to degrade or fuse. Oxidative stabilization prior to carbonization causes the lignin to have an increased resistance to further thermal treatment, mainly due to the formation of crosslinks. (Baker, Rials 2013)
In the oxidative thermostabilization, oxygen contents increase as the temperature rise to 250-260°C. This implies that oxidation reactions take place to create oxygenic function groups. After further heating, beyond 260°C, the oxygen content decrease. (Baker, Rials 2013) Meanwhile, the hydrogen content decrease as water, methanol and methane are released upon heating. Passively, due to the decrease of other elements, the carbon content percentage becomes higher and formation of C-C bonds increases. (Li et al. 2013)
In the condensation of carboxylic acid groups, ester and anhydride moieties are formed. (Li et al. 2013) The esters and anhydrides are formed as the thermostabilization temperature rises. This indicates a cross-linking reaction taking place within the lignin macromolecules. (Li et al. 2013)
Also involved in the stabilization is the β-O-4 linkage. It accounts for nearly 48-60% of the total inter-unit linkages of lignin. (Braun, Holtman & Kadla 2005) Homolytic cleavage of the β-O-4 bond, during the oxidative stabilization, is likely the first step in which radicals are formed followed by a series of rearrangement reactions (Figure 8). (Braun, Holtman & Kadla 2005) This is thought to be sufficient to stabilize lignin and transform it into a thermoset. (Norberg 2012) A more aromatic structure is believed to be formed in the lignin molecule. (Li et al. 2013)
13
Studies have shown that Tg is dependent on both cross-linking and incorporation of oxidized
groups. Low temperature and low heating rates favor oxidative weight gain which result in formation of crosslinks. (Braun, Holtman & Kadla 2005)
Softwood lignin is more susceptible to the oxidizing treatment than hardwood. (Norberg et al. 2013), and possesses more cross-linked structure than hardwood lignin. The molecular
structure of softwood lignin is considered to be more branched than hardwood lignin. (Heber S. et al. 2009) The level of crosslinking contributes to a faster stabilization (Lin et al. 2014). On a sidenote, softwood lignin structure contains high amount of oxygen as compared with petroleum pitch and PAN. It might be sufficient for autocatalytic stabilization to take place using only heat. (Norberg et al. 2013)
In order to determine whether the material is thermoset or not, differential scanning calorimetry (DSC) can be used to look for a glass-transition. If no transition or Tg can be
found, it is an indication that the fibers were fully stabilized (i.e., the fusible thermoplastic lignin fiber has become an infusible thermoset structure). (Norberg et al. 2013)
As for lignin, a skin-core structure has also been identified in the fiber. The structure is formed by oxidative reactions near the surface. Formation prevents further diffusion of oxygen into the fiber. As shown for pitch, fibers with thicker diameters have a greater difficulty in achieving full stabilization during the oxidative stabilization. The same may apply to lignin fibers. (Norberg et al. 2013, Matsumoto, Mochida 1992) It is not known how the skin-core affects mechanical properties of the final carbon fiber. It may have negative complications. (Norberg 2012)
Carbonization
The stabilization is followed by the carbonization process. The sample is treated under
pyrolytic conditions in inert atmosphere (N2) at temperatures between 1000-2000°C. The rate
of heating may vary; 5-10°C/min is preferred. The high temperatures cause removal of undesired heteroatoms (non-carbon elements) from the fiber and formation of ordered graphitical structure, and thus improving mechanical properties as well as electrical and thermal properties. Involved in the carbonization is dehydration, decarboxylation, aromatization and condensation reactions (Chatterjee et al. 2014).
The outcome of finalized carbon fibers is highly dependent on the thermostabilization. Generally speaking, the yield increases as the stabilization temperature reaches 260°C and it has been shown that the best yield of carbon fiber can be attained from fibers stabilized in between 260-290°C, due to the high amount of aromatic structure formed in the stabilization. Above 290°C the yield tend to decrease as the stabilization temperature goes higher. (Li et al. 2013)
Raman spectroscopy can be used to identify the degree of structure ordering in carbon fiber material. (Li et al. 2013) showed that carbonized fibers, in a temperature of 1400°C, had low ordering when the sample had been thermostabilized below 260°C. Above 290°C the amount of C-C structures were higher but the aromatic ordering was disrupted due to the pyrolysis or
14
thermostabilization. A competition may occur between ordering and disruption in the range 260-290°C. Compared to PAN, lignin does not have the same good graphitic structure when treated under the same conditions. (Li et al. 2013)
Furthermore, in their studies the Raman spectra also indicate that the disordered carbon structure exist mainly as sp3 hybridization rather than sp2 hybridization. (Li et al. 2013) There is a possibility that pores and defects could be formed due to the elimination of heteroatoms in heat treatment. (Huang 2009) Regarding the mechanical strength of carbon fiber, it is a well-known fact that voids and defects weaken the carbon fibers. (Lin et al. 2014) conclude that improved mechanical properties are more likely due to the reduction of voids and defects rather than modification of the microstructure. They also state that anisotropic or oriented form is preferred over isotropic form.
15
2 Experimental
The entire project and its studies were conducted at Innventia AB in Stockholm.
Short overview: Lignin fibers were stabilized prior to carbonization. After measuring the deformation during carbonization in the TMA, the carbonized fibers were evaluated in Dia-Stron, TGA, FTIR, SEM and light microscopy analyses (Figure 9).
Figure 9: Schematic overview of the process. 2.1 Stabilization
Prior the start of the project, pH-fractionated softwood kraft lignin (Mw= 51 kDa, Mn = 11 kDa and PDI= 4.6) from Bäckhammar had been extruded and formed to fibers through
multifilament melt spinning. The lignin was pre-treated in vacuum, 24 hours at 80°C followed by 2 hours at 160°C and lastly in nitrogen gas at 180°C for 1.5 hours prior extrusion. A
multifilament extruder (Alex James & associates Inc., USA) was used.
In order to work with the fibers during the project they were stabilized, increasing the glass transition temperature due to cross-linking and thus retaining a thermoset material.
The stabilization was then carried out in a gas chromatograph (GC) oven (Hewlett Packard 5890 Series II, Agilent Technologies Inc.).
In course of the project three batches of stabilized fibers were made. The first two were stabilized in the same way and the third batch was processed slightly different.
In the preparation of the two first batches, lignin fibers were hinged, with clamps serving as counterweights in each end, over a glass rod put inside a glass bowl. The counterweights purpose was to suppress the dimensional changes of the fibers, i.e. to retain the original length. The entire setup was then covered in aluminum foil to protect it from the turbulent fan inside the GC oven.
Stabilization
TMA Carbonization
Dia-Stron Tensile-testing
TGA Mass loss analysis
FTIR Composition analysis
SEM Imaging and composition
analysis
Light microscopy Imaging
16
In the making of the third batch the glass rod was left out and the fibers was simply placed on the bowls aluminum foil covered bottom (Figure 10). Although with glass rod missing, the clamps were still attached to serve as counterweights. However, in this case and position they were probably less effective.
Figure 10: Not yet stabilized lignin fibers placed on the foil covered bottom.
The entire setup was then inserted into the oven. Starting at 25°C, the temperature was increased to 140°C at 3°C/min, then up to 250°C at 0.2°C/min where it remained for 1 hour. The oven completed the program in 10.8 hours after which the temperature passively
decreased.
2.2 Carbonization and Thermomechanical Analysis (TMA)
The principles of TMA and carbonization methods are described in the following sections.
2.2.1 Principles
In static force TMA, the specimen is strapped into a holder in the instrument, subjected to a small/negligible load. A dilatometer then measure the deformation of the specimen as the temperature continuously increases in a linear fashion (Figure 11). This gives information about the thermal and force dependent expansion or contraction (elongation) of the material. (Brown 2001)
Figure 11: Classic thermo-dilatometry of expansion. Note that the arrow can be in the opposite direction, in which case the material contract. The specimen is fixated in the top as indicated by the dot. A small force is applied via the pushrod of the dilatometer to the specimen.
For length measurements, the linear expansion or dimensional change (∆L/L0, where ∆L is the
length difference L-L0 and L0 is initial length) is related to the temperature changes (∆T) by a
linear expansion coefficient (α) and can be expressed as: L0
∆L Positive
direction
17 𝛥𝐿
𝐿0 = 𝛼𝛥𝑇
The linear expansion coefficient is thermomechanically characteristic for a material and is determined by composition and structure of the solid. (Brown 2001)
The TMA can also be used to determine glass-transition temperatures Tg and softening points
(Ts) for certain materials, polymers in particular. (Brown 2001)
2.2.2 Method
A multiple of fibers were glued together into a bundle by applying carbon paint suspension (SPI Supplies®) in both ends of the fibers (Figure 12 a). After a brief moment of curing, in normal atmosphere (1 at) and room temperature (20°C), clamps of alumina (Al2O3) (Model
TMA40200A06.022-00, NETZSCH GmbH, Germany) were attached to the glued part of the fibers with the help of a mounting station (Model TMA40200A06.023-00, NETZSCH GmbH, Germany) (Figure 12 b).
Figure 12: In a), a bundle of stabilized lignin fibers are placed on two stacked microscope slides. Both ends are glued together. The mounting station in b) was used for alignment fixture of the sample in the clamps (white pieces).
The sample of fibers was then mounted in the TMA 402 F1 Hyperion® (NETZSCH GmbH, Germany) header connected to an alumina pushrod (Model TMA40200A07.023-00,
NETZSCH GmbH, Germany) (Figure 13), to be carbonized at high temperatures. The furnace was lowered upon the header, centering it in the chamber, sealing off the outside environment in total isolation. A thermometer thread was positioned close to the sample, measuring the temperature. The freely moving pushrod was connected to a linear variable displacement transducer (LVDT) in the interior of the instrument. As the sample deformed during the carbonization, the sensor recorded the absolute displacement from the point of origin.
18
Figure 13: The TMA-system. The furnace is in the elevated state above the header. Magnification of the header depicts an inserted sample of fibers. Sprints are holding the setup in place. The lower clamp is connected to the pushrod.
Cooling and regulating temperatures below room temperature was accomplished by a F25-MA Refrigerated/Heating Circulator modulus (Julabo GmbH, Germany) connected to the TMA.
For control and acquisition of the TMA-measurements the NETZSCH Measurement software (v.6.1.0, NETZSCH GmbH, Germany) was employed.
Prior to carbonization the chamber was evacuated of atmosphere, to a pressure of approximate 4 mbar (indicated as 100 % vacuum by the TMA monitor), by a diaphragm pump MPC 105 7 (NETZSCH GmbH, Germany). The chamber was shortly after filled again with pure nitrogen gas from a Nitrogen 5.0 LAB LINE® N2 (N2 ≥ 99.999%) tube (Strandmöllen AB, Sweden) by
opening the TMA by-pass valve. The by-pass valve was left open until the atmosphere gained a slight overpressure and was then closed.
When the system were set up and the run program had been added via the NETZSCH Measurement software (v.6.1.0, NETZSCH GmbH, Germany), the carbonization could take place. The main inlets were opened; purge 2 (connected to the TMA chamber) and the protective inlet (protecting the LVDT sensor) along with the main outlet. Opening the main outlet normalized the pressure once again. The N2 gas flow was set to 75 ml/min for each inlet
and remained constant throughout the run.
19
In the first phase, the bundle of fibers (samples) underwent carbonization from 20°C to set temperatures between 300-1200°C in steps of 100°C for each bundle, increasing the temperature 1K/min (or 1°C/min) at a load of 5 mN.
In phase two, the increasing of temperature was varied, testing 10, 20 and 40 K/min, to set temperatures (900-1300°C, in steps of 100°C). Same load as before.
During the third and last phase, heating of 10 K/min and a final carbonization temperature of 1000°C were selected as fixed parameters in order to test different forces. The initial load was increased to 10 mN.
However, since the dimension of the fibers changed, both in length and diameter, due to the mass loss and evaporation of substances, as the temperature rose during the carbonization, the above stated preparations of the fibers (i.e. the use of carbon paint) and attachment could not be applied in the force measurements (last phase). As the fibers shrunk they were simply drawn out from the clamps when the force increased.
The fibers had to be fixated in a better way so that they would not move at all. Plates of graphite were used, instead of the clamps, onto which the fibers were placed, one plate for each end, and adhered with graphite adhesive (Figure 14). A thin layer of graphite bonding agent (Graphi-Bond 551-RN, Aremco Products Inc.) was applied over the fibers and plate. The adhesive was then cured in an oven (Gas Chromatograph, Hewlett Packard 5890 Series II, Agilent Technologies Inc, USA) in three steps; first dried at 25°C for 4 hours, secondly the temperature was increased to 130°C at 25°C/min and was held there for 4 hours and thirdly the temperature was increased further up to 260°C at 3°C/min where it remained for 2 hours. The curing was performed in normal atmosphere (1 at). After the curing the plates could be inserted/mounted in the TMA sample holder.
Figure 14: Plates of graphite onto which stabilized lignin fibers were adhered.
The TMA measured the deformation (in this case linear expansion/contraction) through a dilatometer, at an acquisition rate of 25 points per minute, which could only operate in the range of 5000 µm (5 mm). Reaching the upper limit resulted in an underflow error and lower limit in an overflow error. When the fibers sometimes snapped/ruptured the monitor would always show an underflow error. After completing a run the computer software transferred the data directly to another program (NETZSCH Proteus – Thermal Analysis v.6.1.0) for
20
Each sample of carbonized lignin fibers were put in small glass jars sealed with a plastic cap and marked on the side with date, sample number etc. The samples prepared in the TMA were then examined in FTIR-spectroscopy, tensile tested in the Dia-stron and imaged by SEM.
2.3 Uniaxial Tensile Testing (with Dia-Stron)
The principles of uniaxial tensile testing and used methods are described in the following sections.
2.3.1 Principles
A commonly used method to evaluate a material is tensile testing. The principle behind the test is gripping opposite ends of a specimen/sample in a test machine. Force is exerted on the object by the machine, resulting in gradual elongation and ultimately a fracture of the sample. (Davis 2004)
During the process, force-extension data is recorded, which is a measure how the sample deforms under applied force. This data gives information about important mechanical
properties of the sample. Elastic deformation (Young’s modulus or E-modulus), yield strength and ultimate tensile strength are such properties. (Davis 2004)
Stress (s) is defined as the force that acts over a given cross-sectional area and the unit is N/m2 or Pa. Stress allows for strength comparison between materials. (Davis 2004)
𝑠 =𝐹 𝐴
Strain (e) expressed as the ratio of total dimensional change to the initial dimension which describes the deformation. (Davis 2004)
𝑒 =𝛥𝐿 𝐿0 =
𝐿 − 𝐿0 𝐿0
The Young’s modulus or E-modulus (E) is the slope of the linear part of the stress-strain curve. It is a measure of rigidity and is derived from the stress to strain ratio, which is also expressed in Pa units. (Davis 2004)
𝐸 =𝑠 𝑒= 𝐹 ∕ 𝐴0 𝛥𝐿 ∕ 𝐿0 = 𝐹𝐿0 𝐴0𝛥𝐿
In this project the stress and strain are referred to as engineering stress and strain. (The true stress and strain is defined respectively by the instantaneous change of area and length.) (Davis 2004)
2.3.2 Method
The tensile testing was carried out on an automated system, called Dia-stron (Dia-stron Limted, Andover, UK). The system is composed of several modules; an automated sample loader (ALS1500), a linear extensometer (LEX820) and a laser diffraction system (LDS0200) (Figure 15). Pneumatically driven, the moving arm of the automated sample loader was provided with air from a PG28L-Dual purifier (PEAK Sientific Instruments Ltd., Renfrew,
21
UK) connected to the system, controlled and regulated by a PU1100 Pneumatic unit stron Limted, Andover, UK). The system was controlled through a UV1000 control unit (Dia-stron Limted, Andover, UK).
Figure 15: Modules constituting the Dia-Stron instrument.
The tests were performed on stabilized and carbonized lignin fibers made in the TMA. Prior the strength measurement, plastic plates were placed in designated slots on a cassette, 20 slots in total. Each slot consisted of two plastic plates, in between the distance or gauge length was 12, 20 or 30 mm depending on the cassette. For each slot single fibers were placed, one end in the well of one plate (Figure 16). The wells were then filled with UV-adhesive
(DyMAX 3193, Dymax Corporation, USA) with a Performus™ III dispenser (Nordson EFD, Westlake, USA) set to 3.36 Bar and 0.335 seconds flow. The adhesive was then curated for roughly 4 seconds with a pE-100 ultraviolet pen (CoolLED Limited, Andover, UK) set to an intensity of 50% and a wavelength of 400 nm.
Figure 16: Cassette of gauge length 12 mm with plastic plates. The lines in the magnification indicate the alignment and placement of the fibers.
Automated sample loader
Cassette platform Linear extensometer
22
The whole cassette was then placed on the cassette holder of the automated tensile testing system Dia-stron (Dia-stron Limted, Andover, UK). The automated sample loader picked up the couple of plates and moved the sample to the linear extensometer modulus coupled with the laser diffraction system. Through light diffraction the diameter was measured, in which the modulus “rocked” back and forth to find the position where the signal/intensity of the diffraction was strongest. The laser diffraction system was very sensitive and the software could not detect/calculate any diameter from the diffraction pattern if the samples were irregular in form, insufficiently round or transparent. Regardless of whether the diameter measurements were successful or not the LEX modulus performed, afterwards, an extension of the fiber till it ultimately broke. From the test the UvWin software (v. 3.35.0000 build 2, Dia-stron Limted) plotted the stress versus strain in a graph. By using the analytical tool provided by the same software, the corrected E-modulus could be determined. (The mathematical models are given above, in section 2.3.1).
The compliance for the system was calculated according the appendix (section 8.2.3.1). However, the compliance was neglected in the calculations of E-modulus.
The LDS was set to a tension force of 3 gram force (gmf) prior to measurement and to an extension rate of 0.1 mm/sec. The linear extensometer had a gauge force of 2 gmf prior to measurement and extended the sample to 7% of the gauge length at 0.1 mm/sec. The threshold of breakage was set to 5 gmf and the max force allowed was 500 gmf. When the sample snapped/broke the instrument informed the user by sending out a loud beep noise where after the automated sample loader arm moved and discarded the sample in the originating slot and continued to the next slot.
2.4 Thermogravimetric Analysis (TGA)
The principles of TGA and used methods are described in the following sections.
2.4.1 Principles
TGA is a thermal analysis in which the specimen is subjected to elevated temperatures. In this fashion the chemical and physical properties can be determined by measuring the change of weight as a function of increasing temperature. This provides information about thermal stability, degradation, decomposition and vaporization of substances. (Coats, Redfern 1963) In particular, for this project, decomposition in the form of pyrolysis (in inert atmosphere) is the most interesting aspect.
2.4.2 Method
During carbonization of the stabilized lignin fibers mainly the oxygen and hydrogen content of the fibers will be reduced, a pyrolysis, meaning that different oxygen containing substances will be released from the fibers. In order to monitor the change in mass a thermogravimetric analysis was carried out in a TGA 7 thermogravimetric analyser (PerkinElmer Inc, Waltham, USA).
A few stabilized fibers were hacked into small pieces and put into a little platinum dish. In turn the dish was placed on the internal balance of the TGA. The glass cylinder tube was then elevated so it would contain the sample. In the elevated state gas of pure nitrogen from a
23
Nitrogen 5.0 LAB LINE® N2 (N2 ≥ 99.999%) tube (Strandmöllen AB, Sweden) flowed
through the system. Afterwards the balance was tared and the instrument was then left
untouched alone in the room for a brief moment so that the balance would adjust and stabilize itself. (The sensitive balance fluctuated and would react to even the slightest movement in the room.)
The instrument was connected to a TAC 7/DX thermal analysis control unit (PerkinElmer Inc, Waltham, USA). Via a computer, measurement data of the weight loss was collected by the Pyris software (v. 8.0.0.0172, PerkinElmer Inc., Waltham, USA) over a set temperature range. The program was set to run the TGA from 20°C to 1000°C at 10°C/min as the weight was measured at each point. However, the TGA managed only to reach 885°C before cooling. The initial mass was 0.6290 mg.
However, TGA-instrument was very unreliable since it provided illogical readings (e.g. some measurements showed negative mass). This was probably caused by air leaking into the system, resulting in combustion of the stabilized lignin fibers. In other word the instrument was not sealed completely and the environment inside the system was not kept inert or oxygen free.
As a measure, a reference carbonization was done in the TMA-instrument for comparison. 30 fibers, glued together with carbon paint, were weighted on an AE200 balance (Mettler
Toledo, USA) before and after carbonization up to temperatures of 300, 600 and 900°C (10K/min). Initial masses were 2.7, 2.2 and 2.7 mg respectively for each carbonization temperature. The results were thought to serve as calibration point for the mass-loss curve in the TGA. The weight loss graph was corrected accordingly.
As a second measure of obtaining a decent measurement of weight loss, stabilized lignin fibers were sent to the NETZSCH Applications Laboratory, section of thermal analysis, in Selb, Germany for investigation, in mission of Innventia AB (Stockholm, Sweden). Two measurements were performed with the thermogravimetric analyzer,
thermo-microbalance TG 209 F1 Libra® (NETZSCH GmbH, Selb, Germany). Precise internal mass flow controllers of the vacuum tight instrument and a resolution of 0.1 µg enabled high precision measurements. For control and data acquisition the PROTEUS® software (NETZSCH GmbH, Selb, Germany) was employed.
The samples were placed into an Al2O3 crucible, and the system was evacuated and refilled
with nitrogen (20 ml/min) before measurement. The program was set to run from 20°C to 1100°C at 10°C/min as the weight was measured at each point. Initial masses of the both samples were 1.972 and 2.038 mg.
2.5 Fourier Transform Infrared Spectroscopy (FTIR)
24
2.5.1 Principles
Infrared spectroscopy works on the principle that molecules and atoms interact with electromagnetic radiation. Infrared light causes atoms and groups of atoms to vibrate with increased amplitude about the connecting covalent bond. (Solomons, Fryhle 2011a)
The vibration or oscillation occurs only at certain frequencies, meaning that the vibrational states are quantized energy levels. A transition between states can only occur when the
frequency of the radiation match the transition energy (i.e. the frequencies resonate), in which case the radiation is absorbed by the compound. (Solomons, Fryhle 2011a)
Since different functional groups and molecules are arranged in specific ways, light will be absorbed at frequencies that correspond to the molecules and functional groups. The frequency is related to the masses of the bonded atoms and the stiffness of the bond. (Solomons, Fryhle 2011a)
In order for a molecule to be ‘IR-active’, the dipole moment must change as the vibration occurs. Thus symmetric molecules, such as homodiatomic molecules (e.g. N2 and O2), does
not absorb IR-light. (Solomons, Fryhle 2011a)
The vibration can occur in a variety of ways: symmetric stretching, asymmetric stretching, scissoring and twisting etc. (Solomons, Fryhle 2011a)
An IR-spectrometer operates by irradiating a sample with IR-light. The transmittance of the radiation is compared to the transmittance in absence of sample, a so called reference. The comparison indicates the frequencies where the sample absorbed light. (Solomons, Fryhle 2011a)
FTIR spectrometers employ a Michelson interferometer. The radiation is split into two beams, by a beam splitter. The beams are reflected back on two mirrors, of which one is mobile, and recombine which causes interference. The radiation passes through the sample to the detector which records the signal as an ‘interferogram’. Through a Fourier transform the interferogram is then converted into a spectrum. (Solomons, Fryhle 2011a)
The absorption can be observed as a peak in the spectrum where the position is specified in units of wavenumbers, which is the reciprocal of the wavelength, often expressed as cm-1. (Solomons, Fryhle 2011a)
𝑣̅ =1 𝜆
This is due to the energy (∆E) of absorption being directly proportional to the frequency (f) according the equation:
𝛥𝐸 = ℎ𝑓 =∕ 𝑓 = 𝑐 𝜆∕=
ℎ𝑐
25
2.5.2 Method
Fiber samples were carbonized in the TMA prior analysis in FTIR-spectroscopy. The carbonization temperature and rate was 1000°C and 10 K/min, respectively.
By using an agat mortar the carbonized fibers were grinded thoroughly. Small fragments of the fibers spurt away when grinding, thus the mortar had to be contained within a hand folded dome of aluminum foil. This prevented fiber fragments from escaping.
Depending on the degree of carbonization different amount of sample was used when grinding. The more carbonized the fibers were the less amount had to be used in order to get distinctive spectrum. The weight of the fiber sample was determined by using a highly sensitive balance (Ultramicro Type 4504 MP8-1, Sartorius AG, Goettingen, Germany), measuring weights less than 1 mg, 4 digits precision.
Previously dried for 30 min in a Memmert Type U30 oven (Gemini BV, Apeldoorn, Netherlands), set to 105°C, potassium bromide (KBr) powder (Uvasol®, Merck Millipore, Billerica, USA) was added and mixed/grinded together with the sample. The amount was roughly the same for each sample, i.e. approximately 300 mg.
The mixture was then transferred to a hydraulic press to form a slightly transparent thin black disc-like pellet, 13 mm in diameter and approximatively 2 mm thick, under pressure of 250 kg/cm2 (recalculated to nearly 250 bar) and slight vacuum (Figure 17). Pellets of lignin powder, stabilized or carbonized lignin fibers up to 300°C were, instead, somewhat brown in color. The reference pellet was made in a similar manner with solely KBr powder, 300 mg, resulting in a clear transparent pellet.
26
Due to the hygroscopic nature of KBr (van der Maas, J. H., Tolk 1962, Abo 2010) the full procedure had to be done quickly. The pellet was then rapidly inserted with the holder into the FTIR-spectrometer, since the pellet had the tendency to become “cloudy” rather fast due to it soaking up moisture directly from the atmosphere.
A Varian 680-IR FTIR Spectrometer (Varian Incorporate, Palo Alto, USA) was used for this study (Figure 18).
Figure 18: The Varian spectroscopy instrument.
During the FTIR-spectrometry the atmosphere in the sample compartment was continuously purged with dry nitrogen gas, eliminating substances like carbon dioxide that may interfere with the spectra. The level of moisture in the gas was regulated by an external filter connected to the device keeping the atmosphere as dry as possible.
Since the compartment was opened when the sample was inserted into the pellet holder, air leaked in. By looking at the live spectra, the scan was manually started when the carbon dioxide peak had disappeared. The time between insertion and spectral scan was estimated to 60 sec.
The spectral resolution was set to 2 cm-1 and the sample was scanned in range 4000-600 cm-1. A thermo electric (TE) cooled DLaTGS (deuterated L-alanine doped triglycene sulphate) detector was used to collect the absorbance data at 5 kHz frequency. The samples were illuminated by a mid-infrared source (MIR) and the beam splitter was of KBr.
Both the background and samples were scanned 32 times each. One background spectra was made (150115) and used for all the samples throughout the study. However, it should be noted that the correct procedure is to collect a background spectrum for each measurement,
preferable as close to the measurement as possible.
The collected spectrums were visualized in the Varian Resolutions Pro software (v.5.1.0.829, Varian Incorporate, Palo Alto, USA).
27
2.6 Scanning Electron Microscope (SEM)
The principles of SEM and used methods are described in the following sections.
2.6.1 Principles
The principle behind electron microscope is bombarding a surface of a specimen with a stream of electrons. The scattering of the electrons can be used to generate an image of the scanned area.
The electrons are produced by an electron gun which consists of an electron source (usually tungsten), called cathode, and an electron accelerator. There are different methods to generate the electrons from the cathode. In a field emission gun, the principle of field emission is utilized. I.e. the electrostatic field at the tip of the cathode is increased to the point that the potential barrier becomes small enough to allow electrons to escape through this barrier by quantum-mechanical tunneling. After emission the electrons are accelerated in an electric field, parallel to the optic axis, to a final kinetic energy. (Egerton 2005)
An objective in the SEM condenses the incident beam into a so called electron probe, typically 10 nm in diameter. The beam is scanned over a rectangular area of the specimen, known as raster scanning. (Egerton 2005)
Figure 19: Elastic and inelastic scattering. The electrons are incident from above, indicated by arrows.
Primary electrons from the incident beam penetrate the specimen and scatter both elastically and inelastically (Figure 19). The respective scattering is caused by interaction with the nucleus or with electrons surrounding the nucleus. Primary electrons that backscatter (i.e. scattering through an angle greater than 90° to the incident beam) can be collected by a backscattering detector placed above the specimen. The signal increases with atomic number and thus contrast between chemical compositions can be observed in the imaging of materials. (Egerton 2005)
28
Through inelastic interaction and transfer of kinetic energy, primary electrons cause ejection of electrons orbiting around the nucleus. As moving particles, the secondary electrons
themselves will interact with other atomic electrons and scatter inelastically, gradually losing kinetic energy until most of them are brought to rest within the material. Secondary electrons that instead escape close to the surface are collected for imaging by an Everhart-Thornley detector. Hence the image is a property of the surface structure (topography) and displays topographical contrast. (Egerton 2005)
The same inelastic effect that ‘knock out’ electrons (in other words causes the electrons to undergo transition to a higher energy state) in the inner shell (e.g. K-shell), results in a
vacancy in the shell. An electron in a higher shell de-excite, in a downward energy transition, to fill the hole. The transition causes release/emission of energy in form of photons with wavelengths in the X-ray spectrum. Since the X-rays are characteristic to atomic species, it allows for measurement of the elemental composition of the specimen. A silicon drift detector is often used to detect the X-rays. (Egerton 2005)
Transition to the innermost shell is called K-emission and is denoted α if the electron de-excite from the L shell, β from the M shell etc. Similarly, transition between the M and L shell would result in Lα-emission. (Egerton 2005)
2.6.2 Method
A few stabilized and carbonized lignin fibers from the TMA carbonization process were selected for imaging and X-ray analyses. The analysis was carried out using a scanning electron microscope (Quanta FEG 650, FEI™, USA) combined with an energy dispersive X-ray spectrometer (EDS) (Figure 20). The instrument was located in the Department of Geological Sciences at the Stockholm University.
Figure 20: The scanning electron microscope, Quanta FEG 650.
Main components of the instrument include a field emission gun (FEG), a concentric
29
silicon X-ray, Oxford Instruments). The instrument was operated through the AZtec software (Oxford Instruments).
Prior the study, the fiber samples were attached to the sample holders with double coated PELCO carbon tape (Ted Pella Inc., Redding, USA). Parts of the axial fiber surface and fiber cross-section was imaged and analyzed of content. In the former, the fibers were positioned horizontally on the holder and in the latter the fibers were positioned vertically in a special holder.
The following day the fiber samples were placed inside the SEM chamber which was
evacuated of atmosphere to a pressure of approximatively 1 mbar. The system operated at an accelerating voltage of 15 kV and a working distance (WD) of 10 mm and the fibers could be observed in 1000-5000 times magnification.
2.7 Light Microscopy
Prior and after carbonization treatment in the TMA a light microscope (Axioplan 2 Imaging, Carl Zeiss) was used to look at the structure, integrity, to measure diameter changes and to spot defects. Placed on top of the microscope was a CT5 camera (ProgRes®, Jenoptik AG, Jena, Germany) which was used to capture images of the fibers in the computer software Capture Pro 2.8.8 (ProgRes®, Jenoptik AG, Jena, Germany).
The fibers were simply put on a microscope slide which was placed under the microscope and illuminated from above. The magnification was 10, 20 or 50x and the camera captured the images in resolution of 2592x1944 and 1296x972 2x Bin. In the software the diameters was be measured by drawing a straight perpendicular line between two opposite points of the fiber contour.
30
3 Results
In the following section the outcome of each experiment is described.
3.1 TMA
The TMA provided information about the behavior of stabilized lignin fibers in temperature changes and under static or varied load during carbonization. Each sample is represented by one carbonization in the TMA instrument.
The data from the carbonization is presented as linear dimensional change over temperature or time.
3.1.1 Static measurement
By studying the expansion/contraction (red curve) in Figure 21, where the slope/derivative (blue curve) is represented as the linear expansion coefficient α, we can observe a slight expansion up to 250°C, followed by a fast shrinkage. Thereafter the rate of contraction diminishes slightly around 420°C and continues in a linear fashion until it begins to disappear around 700°C. After 1100°C the slope cease completely, as illustrated in Figure 22 and Figure 23, where some of the samples have been carbonized further.
Figure 21: Deformation of approximatively 50 fibers, 20 mm each, during carbonization to 1000°C at 1 K/min and 5 mN load. The red curve represents the length changes, while the blue curve represents the slope/alpha value of the blue curve.
Hereafter, the carbonization in Figure 21 will be referred to as the standard carbonization, since it represents the general deformation pattern of all fibers. It is color marked as red in the following charts, allowing for easy comparisons.
-0.0010 -0.0005 0.0000 0.0005 -25% -20% -15% -10% -5% 0% 5% 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 α (1 /K ) ∆L/L 0 Temperature (°C)
31
Carbonizations of stabilized fibers with different lengths are compared in Figure 22. The curves follow approximatively the same path. There is no clear difference between 15 and 20 mm fibers. However, the curve of 10 mm deviates from the path and possess a nick around 600°C (possibly due to a fiber or multiple fiber breaking). The final difference, after completion, is around 2 percentage units compared to the 15 and 20 mm fiber samples.
Figure 22: Comparison of samples with varying lengths. The number of fiber in the samples, corresponding to the blue, green and red curve, was respectively 10, 20 and 50. The blue curve possesses a nick around 600°C. The carbonization was performed at 1 K/min, under load of 5 mN.
The following Figure 23 is compiled of sample data with varied amount of fibers. We can see that the majority of the curves fall into the same path.
-25% -20% -15% -10% -5% 0% 5% 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 ∆L/L 0 Temperature (°C) 10 mm 15 mm 20 mm
