- Email: [email protected]
Preparation of carbon fibers from a lignin copolymer with polyacrylonitrile
Synthetic Metals 162 (2012) 453–459
Contents lists available at SciVerse ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Preparation of carbon fibers from a lignin copolymer with polyacrylonitrile Sanjeev P. Maradur a , Chang Hyo Kim b , So Yeun Kim b , Bo-Hye Kim a,∗ , Woo Chul Kim c , Kap Seung Yang a,b,d,∗∗ a
Alan G. MacDiarmid Energy Research Institute, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea Department of Advanced Chemicals & Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea Kolon Central Research Park, Kolon Industries, Inc., Republic of Korea d Department of Polymer & Fiber System Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 3 January 2012 Received in revised form 14 January 2012 Accepted 20 January 2012 Available online 16 February 2012 Keywords: Lignin Polyacrylonitrile–lignin copolymer Wet-spinning Carbon fiber
a b s t r a c t In this study, we have developed an economically viable and technologically sound process for the production of low-cost carbon fibers (CFs) made of lignin copolymer with acrylonitrile (AN). Initially, lignin, a by-product of the pulp and paper industry, is copolymerized with AN in dimethysulfoxide (DMSO) by the radical copolymerization. The resulting copolymer was confirmed by a Fourier transform infrared (FT-IR), 13 C, and 1 H nuclear magnetic resonance (NMR) spectroscopy, showing the presence of the C N group of polyacrylonitrile (PAN) co-eluting with ether, hydroxyl, and aromatic groups that are attributed to lignin. This provided evidence that a PAN–lignin copolymer was synthesized. Using a wet-spinning process, the PAN–lignin copolymers are then spun into fibers with an average tensile strength of 2.41 gf/den, a tensile strain of 11.04%, and a modulus of 22.92 gf/den. The CFs are prepared by the subsequent thermal treatment of the spun fibers. Differential scanning calorimeter (DSC) analysis of the PAN–lignin copolymer-based spun fibers displays a downshifted exothermic peak at 285.83 compared with the homopolymer PANbased as-spun fibers, which provides evidence that lignin is cooperated with the oxidative stabilization reactions. The stabilized fibers are carbonized by heating from room temperature to 800 ◦ C in a nitrogen atmosphere. This study shows the potential for a number of recycled and renewable polymers to be incorporated into wet-spun fibers for production of CF feedstocks, thereby reducing the supply cost using the current commercial technology. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Carbon fibers (CFs) are a lightweight, high performance material with high mechanical strength. Due to their superior properties, CFs are widely employed in areas such as applications in aeronautics, aerospace, sports, and leisure industries [1–6]. Light weight composites dramatically reduce the weight, and thus the fuel consumption of an automobile and would also decrease emissions [7–9]. However, the use of CFs in the automotive industry is limited due to their high cost and the limited supply of the precursor materials [10]. Among the precursors for the production of CFs, polyacrylonitrile (PAN) is the predominant material, due to the excellent mechanical properties of PAN-based CFs [7]. However,
∗ Corresponding author at: Alan G. MacDiarmid Energy Research Institute, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea. Tel.: +82 62 530 0774; fax: +82 62 530 1779. ∗∗ Corresponding author at: Department of Polymer & Fiber System Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea. Tel.: +82 62 530 0774; fax: +82 62 530 1779. E-mail addresses: [email protected] (B.-H. Kim), [email protected] (K.S. Yang). 0379-6779/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.synthmet.2012.01.017
the high cost of PAN precursors, which makes up 43% of the CF manufacturing cost, limits its utilization in general performance grade applications in automotive parts. Therefore, it is necessary to develop alternative precursors for general performance CFs that cost $7/kg or less and can be implemented for large scale automotive use [9,11]. With this goal in mind, extensive work has been carried out to find new precursors for replacing expensive petroleum-based feedstock to low cost renewable alternatives [7–18]. There is a strong interest worldwide in developing suitable technologies that can derive chemicals and materials from renewable biomass. Lignin is one of the most abundant substances in nature and is present in all fibrous plants as a byproduct of the pulp and paper industry. Because lignin is a polyaromatic macromolecule that is readily available and comparatively inexpensive, it may fulfill many of the requirements for being a precursor to CFs. Using lignin as a renewable feedstock for fine chemicals and materials has the potential to generate significant environmental and economic benefits necessary for the viability of the main processes. In spite of these large quantities, lignin has not yet been effectively utilized, and this presents environmental problems. Unfortunately, today’s standard for recovering lignin from paper-mill streams is not
454
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
suitable to yield a high volume of lignin with properties for fast and economically viable wet-spinning and eventual conversion to CFs with the quality requirements for industry demands [19]. Moreover, the technical barriers include low-cost purification of lignin to remove hemi-cellulose, salts, particulate contaminants, water, and volatiles. Consequently, a number of systems have been proposed that use lignin as a renewable polymeric material or as a precursor in the synthesis of new materials that overcome biomass environmental problems and produce new products with beneficial properties [18,20]. In particular, a graft copolymerization reaction of lignin by means of specific monomers can be conducted with either addition or condensation copolymerization using either free radicals or the ionic copolymerization process [21–25]. In the present study, we extend our efforts in developing lignin-based CFs. We sought to minimize the precursor cost by synthesizing the hybrid PAN–lignin copolymer using free radical copolymerization of lignin and AN. Furthermore successful incorporation of lignin in PAN precursors would drastically reduce the precursor cost in carbon fiber production. Therefore, the objectives of the present study include synthesizing lignin copolymers that possess vinyl monomer side chains or main chains and developing lignin-based CFs by the subsequent heat treatment of wet-spun fibers.
2. Material and methods 2.1. Materials Hard wood lignin (PC 1369, Induline, MedWestvaco Co, USA), a commercial lignin product, was used after purifying by desalting from the acidic aqueous lignin solution. The ash content of the purified lignin is 1.0 wt% or less. Dimethyl sulfoxide (DMSO, Reagent grade, Yakuri Pure Chemicals Co. Ltd., Japan), anhydrous calcium chloride (CaCl2 , Duksan Pure Chemicals Co. Ltd., Korea), hydrogen peroxide (H2 O2 , 30%, aqueous solution Duksan Pure Chemicals Co. Ltd., Korea), and ␣,␣-azobisisobutyronitrile (AIBN, 98%, Junsei Chemical Co. Ltd.) were used as received without further purification. Acrylonitrile (AN) was purchased from Sigma–Aldrich and distilled using NaCl aqueous solution to remove the inhibitor prior to use.
2.2. Synthesis of PAN–lignin copolymers The PAN–lignin copolymers were synthesized in a two-step process [15,16,26]; AN was oligomerized/polymerized in the presence of the AIBN initiator in the first step. The oligomerized/polymerized AN was copolymerized with lignin that had been previously initiated by a redox reaction involving hydrogen peroxide and chloride ion in the second step. The specific conditions for step 1 included using a total of 19.08 g of AN and 22 mg AIBN that were stirred into 50 g of DMSO solution for 3 h at 70 ◦ C in nitrogen atmosphere to obtain the oligomerized/polymerized AN. The solution was then cooled to 50 ◦ C. The color of the reaction mixture changed from clear to light yellow. In step two, the oligomerized/polymerized AN was added to 40 g of DMSO solution containing 8 g of hard wood lignin and 8 g of calcium chloride. The mixture was stirred for 15 min within a nitrogen-rich environment, and 4.3 mL of 30% aqueous hydrogen peroxide was added to the reaction mixture. The entire reaction mixture was allowed to react for another 6 h at 70 ◦ C. The terminated slurry was added to acidified water at pH = 2, and the product was recovered by filtration. The precipitates were washed with deionized water for a final pH = 7 and dried at 70 ◦ C in a vacuum oven. The reaction yield was 67.20%.
Table 1 The wet spinning process specifications. Spec
Spinning conditions
Polymer concentration Dope degassing time Nozzle Gear pump Spinning speed Stretch ratio Coagulation bath composition Stepwise stretching temperatures Thermal behavior (temp/shrinkage) Spinning time
16 wt% 12 h 150 holes/0.05˚ 20 RPM 6.0–8.0 m/min 1.76–2.40 100% H2 O 60 ◦ C, 70 ◦ C, 80 ◦ C, 90 ◦ C 150 ◦ C/6% 4h
2.3. Preparation of lignin based CFs Using a wet-spinning process, the lignin-based CFs were spun from the spinning dope that consisted of 16 wt% PAN–lignin copolymers in DMSO. This spinning dope was deaerated and pumped through a spinneret (150 holes, 0.05 mm/hole) to a coagulation bath. In wet-spinning (Fig. S1), the spinning solution was extruded through the spinneret into a coagulation bath that contained a coagulant such as water. The coagulation process was performed gradually by four coagulation baths in which the coagulation temperature ascended gradually from 60 to 90 ◦ C. Coagulation occurred by diffusion of the solvent out of the fiber and diffusion of the coagulant into the fiber. The fibers were then wound onto a traversing take-up roller, which had variable speed control for stretching to the fibers as they came out of the coagulation bath. The spinning parameters are summarized in Table 1. The wet-spun fibers were first thermo-stabilized prior to carbonization. The heat treatment experiments were performed in a tube furnace furnished with gas regulators with a 10 cm heating zone. Tension was applied during the heat treatment process to prevent the fibers from becoming brittle. A 10 cm sample fiber tied with commercial CFs was placed in the center of the heating zone. A tension of 21.2 Pa was applied at one end of the CF bundle and extended to the tube furnace by hanging corresponding weights. The other end of the CF bundle was fixed as shown in Fig. S2. The 10 cm fibers were placed in the center of the heating zone and heated in an air atmosphere. The thermostabilization process was initially carried out at 105 ◦ C for 1 h in air with a heating rate of 1 ◦ C/min to remove moisture in the sample and then heated to 280 ◦ C for 1 h in nitrogen. The thermo-stabilized fibers pulled to a tension of 7.86 Pa were then carbonized at 800 ◦ C with a heating rate of 5 ◦ C/min under a nitrogen atmosphere. 2.4. Characterization Viscosity measurement were performed using a Brookfield viscometer (DV-II+) at a temperature of 25 ◦ C. The surface functionalities of the copolymer were examined by Fourier transform infrared spectroscopy (FT-IR, Nicolet 200 instrument). All the samples were analyzed using the KBr pellet technique and scanned in
H3C
O
HO
-H
O H H3C O
H3C
O
HO
O
.
OH Phenolic hydroxy
H3C O
Benzyl alcohol
OH Methoxy
Fig. 1. Potential sites for hydrogen abstraction for free-radical grafting from lignin [29].
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
455
Fig. 2. FT-IR spectrum of the PAN–lignin copolymer.
the range from 4000 to 400 cm−1 . 13 C and 1 H nuclear magnetic resonance (NMR) spectra were obtained in DMSO-d6 on a Varian Unity Plus 300 spectrometer operating at 300 MHz. Tetramethylsilane was used as internal standard. Samples for differential scanning calorimeter (DSC) experiments were heated to 500 ◦ C under nitrogen atmosphere at a heating rate of 10 ◦ C/min. Thermogravimetric analysis (TGA) was performed using a Shimadzu TGA 50 (Shimadzu, Inc.). The samples were heated to 800 ◦ C in N2 in an electric furnace at a rate of 10 ◦ C/min. The mechanical analysis of the fibers was performed with a universal tensile tester (INSTRON 3365 with tension rate of 60 mm/min). Surface layer morphologies were analyzed using a scanning electron microscope (SEM) model Hitachi, S-4700. As-spun and stabilized fibers were sputter-coated with a platinum layer before being subjected to a SEM analysis. 3. Results and discussion There is the potential for lignin to be used in free-radical copolymerization with unsaturated vinyl monomers. The major functional groups identified are methoxy, phenolic hydroxy, and benzyl alcohol (hydroxyl group on the alpha carbon) groups, as shown in Fig. 1. Among these functional groups, in particular, the benzylic sites of lignin may undergo hydrogen abstraction to make room for grafting sites, while the phenolic hydroxyl groups act as radical scavengers and can limit the free radical copolymerization, resulting in the formation of quinonic structures [27,28]. The acrylonitrile molecule has a highly polar nitrile group that may cause a polymerization via free radical or anionic polymerization. Therefore, the PAN–lignin copolymers can be synthesized by following the procedure in sequential order to obtain sufficient viscosity of the spinning dope for the wet-spinning process. We applied a two-step process wherein in step 1, an oligomerized/polymerized AN was formed using AIBN as an initiator at 70 ◦ C. In step 2, this oligomerized/polymerized AN with the active radical chain was allowed to
react with the lignin radical generated by CaCl2 –H2 O2 redox initiator, resulting in the PAN–lignin copolymers. The FT-IR spectrum of the PAN–lignin copolymers is reported in Fig. 2. We can observe absorption at 3439.08 cm−1 (OH broadened band of either alcoholic or hydroxyl groups of the lignin molecule), 2939.52 cm−1 (CH stretching band of the methyl groups in lignin and AN), 2243.21 cm−1 (C N stretching band in oligomerized/polymerized AN), 1612.49/1454.33 cm−1 (C C and C C stretching bands in the aromatic range, carbonyl), and 1112.93 cm−1 (C O stretching band of ether) [18]. The 13 C NMR spectrum (Fig. 3a) of typical PAN shows three groups of peaks centered around ␦ 27.0, 32.0, and 120.0 ppm, which correspond to the ␣, , and nitrile carbon atom, respectively. Additionally, a peak assigned to C O of lignin is shown at ␦ 58.0 ppm. Furthermore, the 1 H NMR spectrum (Fig. 3b) on the PAN–lignin copolymer shows signals at ␦ 1.9–2.4 (aliphatic proton), ␦ 3.0–4.0 (alcohol and ether protons), and ␦ 6.0–7.5 ppm (aromatic proton) [30]. Hence, the FTIR, 13 C, and 1 H NMR spectrum shows the presence of the C N group of PAN co-eluting with peaks associated with lignin. These spectra, taken together, provide evidence that PAN–lignin copolymers were synthesized. There is a prerequisite condition that a precursor for wetspinning must have enough viscosity so that continuous fibers can be drawn and undergo stretching in subsequent processing. Additional experiments were conducted to identify optimum reaction conditions that would maximize yield and viscosity from the reaction. As shown in Table 2, the synthesis of the PAN–lignin copolymer with the highest viscosity (16,191 cP at 25 ◦ C) can be achieved by free radical copolymerization of AN and lignin at a weight ratio of 8:2. Performing this reaction in the presence of CaCl2 and H2 O2 will induce the configuration of vinyl homopolymers on the lignin backbone. The reaction temperature in both procedures is kept at 70 ◦ C to increase the yield and reduce the reaction time. The spinning dope with the hybrid PAN–lignin copolymer in a DMSO
456
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Table 2 Synthesis data of the PAN–lignin copolymer. Experiment No.
ANa (g)
Lignin (g)
Ratio of AN:ligin
CaCl2 (g)
H2 O2 (mL)
Yield (%)
Viscosityb (cP)
1 2 3 4 5
9.35 14.31 19.08 19.08 108.32
9.44 6.00 8.00 8.00 27.08
5:5 7:3 7:3 7:3 8:2
9.40 6.00 8.00 8.00 27.08
8.50 3.00 4.30 4.30 28.80
89.06 68.93 62.77 67.20 64.62
– 0.85 0.27 11,115 16,191
a b
AN, acrylonitrile. 16 wt% of copolymer in DMF solution.
Table 3 The mechanical properties of the PAN–lignin copolymer-based as-spun fibers.
1 2 3 4 5 6 7 Mean
Denier (den)
Max. load (gf)
Tensile strain (%)
Tenacity (gf/den)
Mod (auto Young’s) (gf/den)
1.79 2.01 2.62 2.44 2.26 2.64 2.73 2.36
4.15 4.55 6.38 5.57 6.17 6.43 6.57 5.69
12.40 8.10 11.00 10.50 13.10 10.10 12.10 11.04
2.32 2.26 2.43 2.28 2.73 2.43 2.41 2.41
23.55 31.21 22.72 18.64 19.53 24.18 20.62 22.92
solution is generally used for the production of lignin-based fibers by the wet-spinning process. The schematic diagram of the wetspinning shown in Fig. S1 consists of a coagulation bath, spinneret, gear pump, roller, and wind-up drum. The unique feature of this wet-spinning process is that it uses 100% water in the coagulation bath, which has been guaranteed to be environmentally friendly
Fig. 3. (a) 13 C NMR spectrum, (b) 1 H NMR of the PAN–lignin copolymer-based asspun fiber.
and economical [31]. The 16 wt% of the PAN–lignin copolymer in a DMSO solution used as the spinning dope is wet-spun into fibers, as shown in Fig. 4. In this work, the polymer concentration chosen as the spinning dope is 16 wt%, due to the good spinnability of spun fibers under these conditions. Throughout the experiment, the spun fibers are drawn with a low stretch ratio of 1.76–2.40 in a warm water bath to reduce fiber diameter and improve the mechanical properties of fibers. The coagulation process is performed gradually by four coagulation baths in which the coagulation temperature ascends gradually from 60 to 90 ◦ C. The coagulation bath temperature has a substantial effect on the coagulation process, which is responsible for controlling the mass transfer and the counter diffusion of the solvent and the non-solvent [32]. When the coagulation process is performed gradually by four coagulation baths, the large molecule chain segments under high drawing rates can move easily and arrange along the draw direction. Thus, the structure becomes more symmetrical and compact, compared to the structure when coagulation occurs using one high temperature [33–35]. For the mechanical properties of the as-spun fibers, the mean values from 7 tests are shown in Table 3. The finesse of the as-spun fibers is 2.36 den, with an average tensile strength of 2.4 gf/den, a modulus of 22.9 gf/den and a tensile strain of 11.04%.
Fig. 4. Photograph of the as-spun fiber.
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Fig. 5. (a) Thermal properties of the PAN–lignin copolymer based as-spun fiber. (b) DSC curves of the PAN–lignin copolymer-based as-spun fiber and homopolymer PAN-based as-spun fiber.
It is necessary to acquire information about the thermal properties of the PAN–lignin copolymer-based as-spun fibers before using heat treatment that will transform the fibers to CFs. The thermal properties of the as-spun fibers are investigated by DSC and TGA experiments in a nitrogen atmosphere. The as-spun fibers exhibit a strong exothermic peak, as observed in the DSC plots in Fig. 5a, which is regarded as the result of the cyclization of a nitrile group, thus causing a large amount of heat release from the PAN [36]. Fig. 5a shows TGA curves of the PAN–lignin copolymer based asspun fibers at the heating rate of 10 ◦ C/min to 800 ◦ C. The main mass
457
loss from 280 to 450 ◦ C, attributed to the pyrolysis of the fibers, is in agreement with the peak on the DSC curve of as-spun fibers. The loss slows down above 450 ◦ C to give a yield of 48% at 800 ◦ C. The DSC experiments investigated the effect of incorporation of lignin in the copolymer on the copolymer’s thermal properties. The PAN–lignin copolymer-based and homopolymer PAN-based as-spun fibers display very sharp exothermic peaks at 285.83 and 308.26 ◦ C, respectively, in Fig. 5b. A linearly downshifted exothermic peak for the PAN–lignin copolymer-based as-spun fibers provides strong experimental evidence that having lignin incorporated into the as-spun fibers accelerates the oxidative stabilization reactions. It proves that the oxidative stabilization processes of the PAN–lignin copolymer-based as-spun fibers are kinetically faster than for homopolymer PAN-based as-spun fibers. Pyrolysis is used during the conversion process from spun fibers into CFs. In general, when lignin, cellulose, and phenol fibers are used as precursors, the stabilization process under oxidative conditions may be omitted from the manufacturing process due to the oxygen in the material itself, and the fibers are subjected directly to carbonization [14]. The chemistry of the stabilization process of PAN consists of cyclization, dehydrogenation, and crosslinking. As a result of the conversion of C N bonds to C N bonds, a fully aromatic cyclized ladder type structure forms [37]. In addition, auto-oxidation can take place to introduce the carbonyl and carboxyl groups into the lignin structure in the presence of air [38]. The carbonyl and carboxyl functionalities of lignin lead to keto, ester, and anhydride linkages through dehydration reactions, condensation, crosslinking and elimination reactions during the stabilization process. Therefore, thermo-stabilization involves cyclization of the PAN component and oxidation of the lignin component in the hybrid precursor fiber, enabling the spun fibers to be thermally stable (infusible) for the subsequent carbonization. From the above results, we can infer that lignin with carbonyl and carboxyl functionalities acts as an initiator for the stabilization reaction and facilitates the process. Thus, cyclization occurs at lower temperatures, as shown in Fig. 5b. A variety of physical changes take place during the thermo-stabilization process, which cause the initial changes in color of spun fibers from pale yellow to dark brown, as shown in Fig. 6. The spun fibers shrink considerably during the thermo-stabilization process. Stabilization shrinks the fiber by approximately 13–19%, and the fiber diameter is also reduced by 18 m when as-spun fibers are stabilized at 280 ◦ C. The results are attributed to a realignment of the composite molecules that were previously aligned and stretched during the fiber spinning in the thermo-stabilization process. The stabilized fibers are then carbonized at a higher temperature in an inert atmosphere, which
Fig. 6. Photographs of (a) as-spun fibers and (b) stabilized fibers.
458
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Fig. 7. FE-SEM images of (a and d) as-spun, (b and e) stabilized, and (c and f) carbonized fibers, the inset shows a cross-sectional field-emission SEM image.
proceeds in the solid state. The thermoset, stabilized fibers are successfully carbonized by heating from room temperature up to 800 ◦ C at a heating rate of 5 ◦ C/min in a stream of nitrogen. High purity nitrogen gas is used to dilute the toxic waste gas as well as to prevent ingress of atmospheric air and combustion of the ligninbased fiber during the carbonization stage. During carbonization, the intermolecular crosslinking occurs through oxygen-containing groups, or the cyclized sections coalesce by crosslinking. The overall yield after carbonization is 56% by numerous evolution gasses during the pyrolysis, such as H2 O, NH3 , HCN, CO, CO2 , N2 , H2 , and CH4 . The SEM images obtained at low and high magnification of the asspun, stabilized, and carbonized fibers are presented in Fig. 7. The average diameter of the fibers decreases from 20 to 15 to 11 m with increasing heat treatment. This is likely the result of drastic volumetric changes involving complex physical and chemical reactions, such as carbon densification and gas evolution. At high magnification, the morphology of the carbonized fibers becomes rougher with stripe patterns, and the carbonized fibers appear more porous than the stabilized fibers through pyrolysis process at high temperature. In reference to the SEM image (inset of Fig. 7c) of the cross-section of the lignin-based CFs that was thermally treated at 800 ◦ C, the carbonized fibers have a flower-like-shaped crosssection with relatively few visible defects.
4. Conclusion Here, we have described our efforts to develop lignin-based CFs by synthesizing a PAN–lignin copolymer using free radical copolymerization of lignin and AN. A spinning dope consisting of 16 wt% PAN–lignin copolymer in a DMSO solution was wet-spun into fibers and drawn at a low stretch ratio of 1.76–2.40 in a warm water bath, which is performed gradually by four coagulation bath in a rage of 60–90 ◦ C, to reduce fiber diameter and improve the mechanical properties of fibers. After the as-spun fibers were heated to 280 ◦ C for 1 h in nitrogen atmosphere right, by which time the fibers were transformed such that they had an infusible property, the stabilized fibers were carbonized to 800 ◦ C in a stream of nitrogen with
the overall yield of 56%. Findings from this study suggest that the introduction of lignin in PAN will be an efficient method for cost reduction in CFs production. Additional work on quality improvement and process optimization of lignin-based CFs is currently underway. Acknowledgements This work was supported by the New Process Development for Cheap Carbon Fibers via Precursor Fibers from Natural/Waste Resources (No. 10033168) and Chonnam National University, 2011. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.synthmet.2012.01.017. References [1] A.K. Gupta, D.K. Paliwal, P. Bajaj, J. Macromol. Sci., Rev. Macromol. Chem. Phys. C31 (1991) 1–89. [2] A. Serkov, G. Budnitskii, M. Radishevskii, V. Medvedev, L. Zlatoustova, Fibre Chem. 35 (2003) 117–121. [3] K.E. Perepelkin, Fibre Chem. 35 (2003) 409–416. [4] L. Peebles, Carbon Fibres, CRC Press, Boca Raton, 1995, pp. 7–26. [5] O. Bahl, Z. Shen, J. Lavin, R. Ross, in: J.B. Donnet, T. Wang, J. Peng, S. Reboyillat (Eds.), Carbon Fibres, Marcel Dekker, New York, 1998, pp. 1–19. [6] J.-H. Yun, B.-H. Kim, K.S. Yang, Y.H. Bang, S.R. Kim, H.-G. Woo, Bull. Korean Chem. Soc. 30 (2009) 2253–2258. [7] K. Sudo, K. Shimizu, J. Appl. Polym. Sci. 44 (1992) 127–134. [8] W. Qin, J.F. Kadla, Ind. Eng. Chem. Res. 50 (2011) 12548–12555. [9] J.F. Kadla, S. Kubo, R.A. Venditti, R.D. Gilbert, A.L. Compere, W. Griffith, Carbon 40 (2002) 2913–2920. [10] A.L. Compere, W.L. Griffith, C.F. Leitten, J.T. Shaffer, Adv. Affordable Mater. Technol. 33 (2001) 1306–1314. [11] D.A. Baker, N.C. Gallego, F.S. Baker, J. Appl. Polym. Sci. 124 (2012) 227–234. [12] Q. Shen, T. Zhang, W.-X. Zhang, S. Chen, M. Mezgebe, J. Appl. Polym. Sci. 121 (2011) 989–994. [13] S. Kubo1, J.F. Kadla, J. Polym. Environ. 13 (2005) 97–105. [14] S. Kubo, Y. Uraki, Y. Sano, Carbon 36 (1998) 1119–1124. [15] M.-J. Chen, D.W. Gunnells, D.J. Gardner, O. Milstein, R. Gersonde, H.J. Feine, A. Hulttermann, R. Frund, H.D. Luldemann, J.J. Meister, Macromolecules 29 (1996) 1389–1398.
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459 [16] C. Bonini, M. D’auria, R. Ferri, R. Pucciariello, A.R. Sabia, J. Appl. Polym. Sci. 90 (2003) 1163–1171. [17] J.J. Meister, C.T. Li, Macromolecules 25 (1992) 611–616. [18] M.N.M. Ibrahim, M.R. Ahmed-Haras, C.S. Sipaut, H.Y. Aboul-Enein, A.A. Mohamed, Carbohydr. Polym. 80 (2010) 1102–1110. [19] F.G. Calvo-Flores, J. Dobado, ChemSusChem 3 (2010) 1227–1235. [20] W.G. Glasser, R.A. Northey, T.P. Schultz, Lignin: Historical, Biological and Materials Perspectives, American Chemical Society, Washington, DC, 2000. [21] S.S. Kelley, T.C. Ward, W.G. Glasser, J. Appl. Polym. Sci. 41 (1990) 2813–2828. [22] H. Hatakeyama, S. Hirose, T. Hatakeyama, K. Nakamura, K. Kobashigawa, N. Morohoshi, J. Macromol. Sci., Pure Appl. Chem. 32 (1995) 743–750. [23] R.W. Thring, M.N. Vanderlaan, S.L. Griffin, Biomass Bioenergy 13 (1997) 125–132. [24] V. Marchetti, P. Gerardin, P. Tekely, B. Loubinoux, Horzfors-chung 52 (1998) 654–660. [25] J.H. Lora, W.G. Glasser, J. Polym. Environ. 10 (2002) 39–48.
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]
459
J.J. Meister, M.-J. Chen, Macromolecules 24 (1991) 6843–6848. L.R.C. Barclay, F. Xi, J.Q. Norris, J. Wood Chem. Technol. 17 (1997) 73–90. F.J. Lu, L.-H. Chu, R.-J. Gau, Nutr. Cancer 30 (1998) 31–38. W.O.S. Doherty, P. Mousavioun, C.M. Fellows, Ind. Crops Prod. 33 (2011) 259–276. V.R. Pai Verneker, B. Shaha, Macromolecules 19 (1986) 1851–1856. N. Yusof, A.F. Ismail, Int. J. Chem. Environ. Eng. 1 (2010) 79–84. A.F. Ismail, M.A. Rahman, A. Mustafa, T. Matsuura, Mater. Sci. Eng. A 485 (2008) 251–257. C.-g. Wang, X.-g. Dong, Q.-f. Wang, J. Polym. Res. 16 (2009) 719–724. W. Liu, G. Gao, J. Appl. Polym. Sci. 93 (2004) 956–960. J. Chen, C.W.H. Ge, Y. Bai, Y. Wang, J. Polym. Res. 14 (2007) 223–228. W.-x. Zhang, Y.-z. Wang, C.-f. Sun, J. Polym. Res. 14 (2007) 467–474. D. Esrafilzadeh, M. Morshed, H. Tavanai, D. Esrafilzadeh, M. Morshed, H. Tavanai, Synth. Met. 159 (2009) 267–272. J.L. Braun, K.M. Holtman, J.F. Kadla, Carbon 43 (2005) 385–394.
Contents lists available at SciVerse ScienceDirect
Synthetic Metals journal homepage: www.elsevier.com/locate/synmet
Preparation of carbon fibers from a lignin copolymer with polyacrylonitrile Sanjeev P. Maradur a , Chang Hyo Kim b , So Yeun Kim b , Bo-Hye Kim a,∗ , Woo Chul Kim c , Kap Seung Yang a,b,d,∗∗ a
Alan G. MacDiarmid Energy Research Institute, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea Department of Advanced Chemicals & Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea Kolon Central Research Park, Kolon Industries, Inc., Republic of Korea d Department of Polymer & Fiber System Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 3 January 2012 Received in revised form 14 January 2012 Accepted 20 January 2012 Available online 16 February 2012 Keywords: Lignin Polyacrylonitrile–lignin copolymer Wet-spinning Carbon fiber
a b s t r a c t In this study, we have developed an economically viable and technologically sound process for the production of low-cost carbon fibers (CFs) made of lignin copolymer with acrylonitrile (AN). Initially, lignin, a by-product of the pulp and paper industry, is copolymerized with AN in dimethysulfoxide (DMSO) by the radical copolymerization. The resulting copolymer was confirmed by a Fourier transform infrared (FT-IR), 13 C, and 1 H nuclear magnetic resonance (NMR) spectroscopy, showing the presence of the C N group of polyacrylonitrile (PAN) co-eluting with ether, hydroxyl, and aromatic groups that are attributed to lignin. This provided evidence that a PAN–lignin copolymer was synthesized. Using a wet-spinning process, the PAN–lignin copolymers are then spun into fibers with an average tensile strength of 2.41 gf/den, a tensile strain of 11.04%, and a modulus of 22.92 gf/den. The CFs are prepared by the subsequent thermal treatment of the spun fibers. Differential scanning calorimeter (DSC) analysis of the PAN–lignin copolymer-based spun fibers displays a downshifted exothermic peak at 285.83 compared with the homopolymer PANbased as-spun fibers, which provides evidence that lignin is cooperated with the oxidative stabilization reactions. The stabilized fibers are carbonized by heating from room temperature to 800 ◦ C in a nitrogen atmosphere. This study shows the potential for a number of recycled and renewable polymers to be incorporated into wet-spun fibers for production of CF feedstocks, thereby reducing the supply cost using the current commercial technology. © 2012 Elsevier B.V. All rights reserved.
1. Introduction Carbon fibers (CFs) are a lightweight, high performance material with high mechanical strength. Due to their superior properties, CFs are widely employed in areas such as applications in aeronautics, aerospace, sports, and leisure industries [1–6]. Light weight composites dramatically reduce the weight, and thus the fuel consumption of an automobile and would also decrease emissions [7–9]. However, the use of CFs in the automotive industry is limited due to their high cost and the limited supply of the precursor materials [10]. Among the precursors for the production of CFs, polyacrylonitrile (PAN) is the predominant material, due to the excellent mechanical properties of PAN-based CFs [7]. However,
∗ Corresponding author at: Alan G. MacDiarmid Energy Research Institute, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea. Tel.: +82 62 530 0774; fax: +82 62 530 1779. ∗∗ Corresponding author at: Department of Polymer & Fiber System Engineering, Chonnam National University, Yong-Bong dong 300, 500-757 Gwangju, Republic of Korea. Tel.: +82 62 530 0774; fax: +82 62 530 1779. E-mail addresses: [email protected] (B.-H. Kim), [email protected] (K.S. Yang). 0379-6779/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.synthmet.2012.01.017
the high cost of PAN precursors, which makes up 43% of the CF manufacturing cost, limits its utilization in general performance grade applications in automotive parts. Therefore, it is necessary to develop alternative precursors for general performance CFs that cost $7/kg or less and can be implemented for large scale automotive use [9,11]. With this goal in mind, extensive work has been carried out to find new precursors for replacing expensive petroleum-based feedstock to low cost renewable alternatives [7–18]. There is a strong interest worldwide in developing suitable technologies that can derive chemicals and materials from renewable biomass. Lignin is one of the most abundant substances in nature and is present in all fibrous plants as a byproduct of the pulp and paper industry. Because lignin is a polyaromatic macromolecule that is readily available and comparatively inexpensive, it may fulfill many of the requirements for being a precursor to CFs. Using lignin as a renewable feedstock for fine chemicals and materials has the potential to generate significant environmental and economic benefits necessary for the viability of the main processes. In spite of these large quantities, lignin has not yet been effectively utilized, and this presents environmental problems. Unfortunately, today’s standard for recovering lignin from paper-mill streams is not
454
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
suitable to yield a high volume of lignin with properties for fast and economically viable wet-spinning and eventual conversion to CFs with the quality requirements for industry demands [19]. Moreover, the technical barriers include low-cost purification of lignin to remove hemi-cellulose, salts, particulate contaminants, water, and volatiles. Consequently, a number of systems have been proposed that use lignin as a renewable polymeric material or as a precursor in the synthesis of new materials that overcome biomass environmental problems and produce new products with beneficial properties [18,20]. In particular, a graft copolymerization reaction of lignin by means of specific monomers can be conducted with either addition or condensation copolymerization using either free radicals or the ionic copolymerization process [21–25]. In the present study, we extend our efforts in developing lignin-based CFs. We sought to minimize the precursor cost by synthesizing the hybrid PAN–lignin copolymer using free radical copolymerization of lignin and AN. Furthermore successful incorporation of lignin in PAN precursors would drastically reduce the precursor cost in carbon fiber production. Therefore, the objectives of the present study include synthesizing lignin copolymers that possess vinyl monomer side chains or main chains and developing lignin-based CFs by the subsequent heat treatment of wet-spun fibers.
2. Material and methods 2.1. Materials Hard wood lignin (PC 1369, Induline, MedWestvaco Co, USA), a commercial lignin product, was used after purifying by desalting from the acidic aqueous lignin solution. The ash content of the purified lignin is 1.0 wt% or less. Dimethyl sulfoxide (DMSO, Reagent grade, Yakuri Pure Chemicals Co. Ltd., Japan), anhydrous calcium chloride (CaCl2 , Duksan Pure Chemicals Co. Ltd., Korea), hydrogen peroxide (H2 O2 , 30%, aqueous solution Duksan Pure Chemicals Co. Ltd., Korea), and ␣,␣-azobisisobutyronitrile (AIBN, 98%, Junsei Chemical Co. Ltd.) were used as received without further purification. Acrylonitrile (AN) was purchased from Sigma–Aldrich and distilled using NaCl aqueous solution to remove the inhibitor prior to use.
2.2. Synthesis of PAN–lignin copolymers The PAN–lignin copolymers were synthesized in a two-step process [15,16,26]; AN was oligomerized/polymerized in the presence of the AIBN initiator in the first step. The oligomerized/polymerized AN was copolymerized with lignin that had been previously initiated by a redox reaction involving hydrogen peroxide and chloride ion in the second step. The specific conditions for step 1 included using a total of 19.08 g of AN and 22 mg AIBN that were stirred into 50 g of DMSO solution for 3 h at 70 ◦ C in nitrogen atmosphere to obtain the oligomerized/polymerized AN. The solution was then cooled to 50 ◦ C. The color of the reaction mixture changed from clear to light yellow. In step two, the oligomerized/polymerized AN was added to 40 g of DMSO solution containing 8 g of hard wood lignin and 8 g of calcium chloride. The mixture was stirred for 15 min within a nitrogen-rich environment, and 4.3 mL of 30% aqueous hydrogen peroxide was added to the reaction mixture. The entire reaction mixture was allowed to react for another 6 h at 70 ◦ C. The terminated slurry was added to acidified water at pH = 2, and the product was recovered by filtration. The precipitates were washed with deionized water for a final pH = 7 and dried at 70 ◦ C in a vacuum oven. The reaction yield was 67.20%.
Table 1 The wet spinning process specifications. Spec
Spinning conditions
Polymer concentration Dope degassing time Nozzle Gear pump Spinning speed Stretch ratio Coagulation bath composition Stepwise stretching temperatures Thermal behavior (temp/shrinkage) Spinning time
16 wt% 12 h 150 holes/0.05˚ 20 RPM 6.0–8.0 m/min 1.76–2.40 100% H2 O 60 ◦ C, 70 ◦ C, 80 ◦ C, 90 ◦ C 150 ◦ C/6% 4h
2.3. Preparation of lignin based CFs Using a wet-spinning process, the lignin-based CFs were spun from the spinning dope that consisted of 16 wt% PAN–lignin copolymers in DMSO. This spinning dope was deaerated and pumped through a spinneret (150 holes, 0.05 mm/hole) to a coagulation bath. In wet-spinning (Fig. S1), the spinning solution was extruded through the spinneret into a coagulation bath that contained a coagulant such as water. The coagulation process was performed gradually by four coagulation baths in which the coagulation temperature ascended gradually from 60 to 90 ◦ C. Coagulation occurred by diffusion of the solvent out of the fiber and diffusion of the coagulant into the fiber. The fibers were then wound onto a traversing take-up roller, which had variable speed control for stretching to the fibers as they came out of the coagulation bath. The spinning parameters are summarized in Table 1. The wet-spun fibers were first thermo-stabilized prior to carbonization. The heat treatment experiments were performed in a tube furnace furnished with gas regulators with a 10 cm heating zone. Tension was applied during the heat treatment process to prevent the fibers from becoming brittle. A 10 cm sample fiber tied with commercial CFs was placed in the center of the heating zone. A tension of 21.2 Pa was applied at one end of the CF bundle and extended to the tube furnace by hanging corresponding weights. The other end of the CF bundle was fixed as shown in Fig. S2. The 10 cm fibers were placed in the center of the heating zone and heated in an air atmosphere. The thermostabilization process was initially carried out at 105 ◦ C for 1 h in air with a heating rate of 1 ◦ C/min to remove moisture in the sample and then heated to 280 ◦ C for 1 h in nitrogen. The thermo-stabilized fibers pulled to a tension of 7.86 Pa were then carbonized at 800 ◦ C with a heating rate of 5 ◦ C/min under a nitrogen atmosphere. 2.4. Characterization Viscosity measurement were performed using a Brookfield viscometer (DV-II+) at a temperature of 25 ◦ C. The surface functionalities of the copolymer were examined by Fourier transform infrared spectroscopy (FT-IR, Nicolet 200 instrument). All the samples were analyzed using the KBr pellet technique and scanned in
H3C
O
HO
-H
O H H3C O
H3C
O
HO
O
.
OH Phenolic hydroxy
H3C O
Benzyl alcohol
OH Methoxy
Fig. 1. Potential sites for hydrogen abstraction for free-radical grafting from lignin [29].
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
455
Fig. 2. FT-IR spectrum of the PAN–lignin copolymer.
the range from 4000 to 400 cm−1 . 13 C and 1 H nuclear magnetic resonance (NMR) spectra were obtained in DMSO-d6 on a Varian Unity Plus 300 spectrometer operating at 300 MHz. Tetramethylsilane was used as internal standard. Samples for differential scanning calorimeter (DSC) experiments were heated to 500 ◦ C under nitrogen atmosphere at a heating rate of 10 ◦ C/min. Thermogravimetric analysis (TGA) was performed using a Shimadzu TGA 50 (Shimadzu, Inc.). The samples were heated to 800 ◦ C in N2 in an electric furnace at a rate of 10 ◦ C/min. The mechanical analysis of the fibers was performed with a universal tensile tester (INSTRON 3365 with tension rate of 60 mm/min). Surface layer morphologies were analyzed using a scanning electron microscope (SEM) model Hitachi, S-4700. As-spun and stabilized fibers were sputter-coated with a platinum layer before being subjected to a SEM analysis. 3. Results and discussion There is the potential for lignin to be used in free-radical copolymerization with unsaturated vinyl monomers. The major functional groups identified are methoxy, phenolic hydroxy, and benzyl alcohol (hydroxyl group on the alpha carbon) groups, as shown in Fig. 1. Among these functional groups, in particular, the benzylic sites of lignin may undergo hydrogen abstraction to make room for grafting sites, while the phenolic hydroxyl groups act as radical scavengers and can limit the free radical copolymerization, resulting in the formation of quinonic structures [27,28]. The acrylonitrile molecule has a highly polar nitrile group that may cause a polymerization via free radical or anionic polymerization. Therefore, the PAN–lignin copolymers can be synthesized by following the procedure in sequential order to obtain sufficient viscosity of the spinning dope for the wet-spinning process. We applied a two-step process wherein in step 1, an oligomerized/polymerized AN was formed using AIBN as an initiator at 70 ◦ C. In step 2, this oligomerized/polymerized AN with the active radical chain was allowed to
react with the lignin radical generated by CaCl2 –H2 O2 redox initiator, resulting in the PAN–lignin copolymers. The FT-IR spectrum of the PAN–lignin copolymers is reported in Fig. 2. We can observe absorption at 3439.08 cm−1 (OH broadened band of either alcoholic or hydroxyl groups of the lignin molecule), 2939.52 cm−1 (CH stretching band of the methyl groups in lignin and AN), 2243.21 cm−1 (C N stretching band in oligomerized/polymerized AN), 1612.49/1454.33 cm−1 (C C and C C stretching bands in the aromatic range, carbonyl), and 1112.93 cm−1 (C O stretching band of ether) [18]. The 13 C NMR spectrum (Fig. 3a) of typical PAN shows three groups of peaks centered around ␦ 27.0, 32.0, and 120.0 ppm, which correspond to the ␣, , and nitrile carbon atom, respectively. Additionally, a peak assigned to C O of lignin is shown at ␦ 58.0 ppm. Furthermore, the 1 H NMR spectrum (Fig. 3b) on the PAN–lignin copolymer shows signals at ␦ 1.9–2.4 (aliphatic proton), ␦ 3.0–4.0 (alcohol and ether protons), and ␦ 6.0–7.5 ppm (aromatic proton) [30]. Hence, the FTIR, 13 C, and 1 H NMR spectrum shows the presence of the C N group of PAN co-eluting with peaks associated with lignin. These spectra, taken together, provide evidence that PAN–lignin copolymers were synthesized. There is a prerequisite condition that a precursor for wetspinning must have enough viscosity so that continuous fibers can be drawn and undergo stretching in subsequent processing. Additional experiments were conducted to identify optimum reaction conditions that would maximize yield and viscosity from the reaction. As shown in Table 2, the synthesis of the PAN–lignin copolymer with the highest viscosity (16,191 cP at 25 ◦ C) can be achieved by free radical copolymerization of AN and lignin at a weight ratio of 8:2. Performing this reaction in the presence of CaCl2 and H2 O2 will induce the configuration of vinyl homopolymers on the lignin backbone. The reaction temperature in both procedures is kept at 70 ◦ C to increase the yield and reduce the reaction time. The spinning dope with the hybrid PAN–lignin copolymer in a DMSO
456
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Table 2 Synthesis data of the PAN–lignin copolymer. Experiment No.
ANa (g)
Lignin (g)
Ratio of AN:ligin
CaCl2 (g)
H2 O2 (mL)
Yield (%)
Viscosityb (cP)
1 2 3 4 5
9.35 14.31 19.08 19.08 108.32
9.44 6.00 8.00 8.00 27.08
5:5 7:3 7:3 7:3 8:2
9.40 6.00 8.00 8.00 27.08
8.50 3.00 4.30 4.30 28.80
89.06 68.93 62.77 67.20 64.62
– 0.85 0.27 11,115 16,191
a b
AN, acrylonitrile. 16 wt% of copolymer in DMF solution.
Table 3 The mechanical properties of the PAN–lignin copolymer-based as-spun fibers.
1 2 3 4 5 6 7 Mean
Denier (den)
Max. load (gf)
Tensile strain (%)
Tenacity (gf/den)
Mod (auto Young’s) (gf/den)
1.79 2.01 2.62 2.44 2.26 2.64 2.73 2.36
4.15 4.55 6.38 5.57 6.17 6.43 6.57 5.69
12.40 8.10 11.00 10.50 13.10 10.10 12.10 11.04
2.32 2.26 2.43 2.28 2.73 2.43 2.41 2.41
23.55 31.21 22.72 18.64 19.53 24.18 20.62 22.92
solution is generally used for the production of lignin-based fibers by the wet-spinning process. The schematic diagram of the wetspinning shown in Fig. S1 consists of a coagulation bath, spinneret, gear pump, roller, and wind-up drum. The unique feature of this wet-spinning process is that it uses 100% water in the coagulation bath, which has been guaranteed to be environmentally friendly
Fig. 3. (a) 13 C NMR spectrum, (b) 1 H NMR of the PAN–lignin copolymer-based asspun fiber.
and economical [31]. The 16 wt% of the PAN–lignin copolymer in a DMSO solution used as the spinning dope is wet-spun into fibers, as shown in Fig. 4. In this work, the polymer concentration chosen as the spinning dope is 16 wt%, due to the good spinnability of spun fibers under these conditions. Throughout the experiment, the spun fibers are drawn with a low stretch ratio of 1.76–2.40 in a warm water bath to reduce fiber diameter and improve the mechanical properties of fibers. The coagulation process is performed gradually by four coagulation baths in which the coagulation temperature ascends gradually from 60 to 90 ◦ C. The coagulation bath temperature has a substantial effect on the coagulation process, which is responsible for controlling the mass transfer and the counter diffusion of the solvent and the non-solvent [32]. When the coagulation process is performed gradually by four coagulation baths, the large molecule chain segments under high drawing rates can move easily and arrange along the draw direction. Thus, the structure becomes more symmetrical and compact, compared to the structure when coagulation occurs using one high temperature [33–35]. For the mechanical properties of the as-spun fibers, the mean values from 7 tests are shown in Table 3. The finesse of the as-spun fibers is 2.36 den, with an average tensile strength of 2.4 gf/den, a modulus of 22.9 gf/den and a tensile strain of 11.04%.
Fig. 4. Photograph of the as-spun fiber.
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Fig. 5. (a) Thermal properties of the PAN–lignin copolymer based as-spun fiber. (b) DSC curves of the PAN–lignin copolymer-based as-spun fiber and homopolymer PAN-based as-spun fiber.
It is necessary to acquire information about the thermal properties of the PAN–lignin copolymer-based as-spun fibers before using heat treatment that will transform the fibers to CFs. The thermal properties of the as-spun fibers are investigated by DSC and TGA experiments in a nitrogen atmosphere. The as-spun fibers exhibit a strong exothermic peak, as observed in the DSC plots in Fig. 5a, which is regarded as the result of the cyclization of a nitrile group, thus causing a large amount of heat release from the PAN [36]. Fig. 5a shows TGA curves of the PAN–lignin copolymer based asspun fibers at the heating rate of 10 ◦ C/min to 800 ◦ C. The main mass
457
loss from 280 to 450 ◦ C, attributed to the pyrolysis of the fibers, is in agreement with the peak on the DSC curve of as-spun fibers. The loss slows down above 450 ◦ C to give a yield of 48% at 800 ◦ C. The DSC experiments investigated the effect of incorporation of lignin in the copolymer on the copolymer’s thermal properties. The PAN–lignin copolymer-based and homopolymer PAN-based as-spun fibers display very sharp exothermic peaks at 285.83 and 308.26 ◦ C, respectively, in Fig. 5b. A linearly downshifted exothermic peak for the PAN–lignin copolymer-based as-spun fibers provides strong experimental evidence that having lignin incorporated into the as-spun fibers accelerates the oxidative stabilization reactions. It proves that the oxidative stabilization processes of the PAN–lignin copolymer-based as-spun fibers are kinetically faster than for homopolymer PAN-based as-spun fibers. Pyrolysis is used during the conversion process from spun fibers into CFs. In general, when lignin, cellulose, and phenol fibers are used as precursors, the stabilization process under oxidative conditions may be omitted from the manufacturing process due to the oxygen in the material itself, and the fibers are subjected directly to carbonization [14]. The chemistry of the stabilization process of PAN consists of cyclization, dehydrogenation, and crosslinking. As a result of the conversion of C N bonds to C N bonds, a fully aromatic cyclized ladder type structure forms [37]. In addition, auto-oxidation can take place to introduce the carbonyl and carboxyl groups into the lignin structure in the presence of air [38]. The carbonyl and carboxyl functionalities of lignin lead to keto, ester, and anhydride linkages through dehydration reactions, condensation, crosslinking and elimination reactions during the stabilization process. Therefore, thermo-stabilization involves cyclization of the PAN component and oxidation of the lignin component in the hybrid precursor fiber, enabling the spun fibers to be thermally stable (infusible) for the subsequent carbonization. From the above results, we can infer that lignin with carbonyl and carboxyl functionalities acts as an initiator for the stabilization reaction and facilitates the process. Thus, cyclization occurs at lower temperatures, as shown in Fig. 5b. A variety of physical changes take place during the thermo-stabilization process, which cause the initial changes in color of spun fibers from pale yellow to dark brown, as shown in Fig. 6. The spun fibers shrink considerably during the thermo-stabilization process. Stabilization shrinks the fiber by approximately 13–19%, and the fiber diameter is also reduced by 18 m when as-spun fibers are stabilized at 280 ◦ C. The results are attributed to a realignment of the composite molecules that were previously aligned and stretched during the fiber spinning in the thermo-stabilization process. The stabilized fibers are then carbonized at a higher temperature in an inert atmosphere, which
Fig. 6. Photographs of (a) as-spun fibers and (b) stabilized fibers.
458
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459
Fig. 7. FE-SEM images of (a and d) as-spun, (b and e) stabilized, and (c and f) carbonized fibers, the inset shows a cross-sectional field-emission SEM image.
proceeds in the solid state. The thermoset, stabilized fibers are successfully carbonized by heating from room temperature up to 800 ◦ C at a heating rate of 5 ◦ C/min in a stream of nitrogen. High purity nitrogen gas is used to dilute the toxic waste gas as well as to prevent ingress of atmospheric air and combustion of the ligninbased fiber during the carbonization stage. During carbonization, the intermolecular crosslinking occurs through oxygen-containing groups, or the cyclized sections coalesce by crosslinking. The overall yield after carbonization is 56% by numerous evolution gasses during the pyrolysis, such as H2 O, NH3 , HCN, CO, CO2 , N2 , H2 , and CH4 . The SEM images obtained at low and high magnification of the asspun, stabilized, and carbonized fibers are presented in Fig. 7. The average diameter of the fibers decreases from 20 to 15 to 11 m with increasing heat treatment. This is likely the result of drastic volumetric changes involving complex physical and chemical reactions, such as carbon densification and gas evolution. At high magnification, the morphology of the carbonized fibers becomes rougher with stripe patterns, and the carbonized fibers appear more porous than the stabilized fibers through pyrolysis process at high temperature. In reference to the SEM image (inset of Fig. 7c) of the cross-section of the lignin-based CFs that was thermally treated at 800 ◦ C, the carbonized fibers have a flower-like-shaped crosssection with relatively few visible defects.
4. Conclusion Here, we have described our efforts to develop lignin-based CFs by synthesizing a PAN–lignin copolymer using free radical copolymerization of lignin and AN. A spinning dope consisting of 16 wt% PAN–lignin copolymer in a DMSO solution was wet-spun into fibers and drawn at a low stretch ratio of 1.76–2.40 in a warm water bath, which is performed gradually by four coagulation bath in a rage of 60–90 ◦ C, to reduce fiber diameter and improve the mechanical properties of fibers. After the as-spun fibers were heated to 280 ◦ C for 1 h in nitrogen atmosphere right, by which time the fibers were transformed such that they had an infusible property, the stabilized fibers were carbonized to 800 ◦ C in a stream of nitrogen with
the overall yield of 56%. Findings from this study suggest that the introduction of lignin in PAN will be an efficient method for cost reduction in CFs production. Additional work on quality improvement and process optimization of lignin-based CFs is currently underway. Acknowledgements This work was supported by the New Process Development for Cheap Carbon Fibers via Precursor Fibers from Natural/Waste Resources (No. 10033168) and Chonnam National University, 2011. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.synthmet.2012.01.017. References [1] A.K. Gupta, D.K. Paliwal, P. Bajaj, J. Macromol. Sci., Rev. Macromol. Chem. Phys. C31 (1991) 1–89. [2] A. Serkov, G. Budnitskii, M. Radishevskii, V. Medvedev, L. Zlatoustova, Fibre Chem. 35 (2003) 117–121. [3] K.E. Perepelkin, Fibre Chem. 35 (2003) 409–416. [4] L. Peebles, Carbon Fibres, CRC Press, Boca Raton, 1995, pp. 7–26. [5] O. Bahl, Z. Shen, J. Lavin, R. Ross, in: J.B. Donnet, T. Wang, J. Peng, S. Reboyillat (Eds.), Carbon Fibres, Marcel Dekker, New York, 1998, pp. 1–19. [6] J.-H. Yun, B.-H. Kim, K.S. Yang, Y.H. Bang, S.R. Kim, H.-G. Woo, Bull. Korean Chem. Soc. 30 (2009) 2253–2258. [7] K. Sudo, K. Shimizu, J. Appl. Polym. Sci. 44 (1992) 127–134. [8] W. Qin, J.F. Kadla, Ind. Eng. Chem. Res. 50 (2011) 12548–12555. [9] J.F. Kadla, S. Kubo, R.A. Venditti, R.D. Gilbert, A.L. Compere, W. Griffith, Carbon 40 (2002) 2913–2920. [10] A.L. Compere, W.L. Griffith, C.F. Leitten, J.T. Shaffer, Adv. Affordable Mater. Technol. 33 (2001) 1306–1314. [11] D.A. Baker, N.C. Gallego, F.S. Baker, J. Appl. Polym. Sci. 124 (2012) 227–234. [12] Q. Shen, T. Zhang, W.-X. Zhang, S. Chen, M. Mezgebe, J. Appl. Polym. Sci. 121 (2011) 989–994. [13] S. Kubo1, J.F. Kadla, J. Polym. Environ. 13 (2005) 97–105. [14] S. Kubo, Y. Uraki, Y. Sano, Carbon 36 (1998) 1119–1124. [15] M.-J. Chen, D.W. Gunnells, D.J. Gardner, O. Milstein, R. Gersonde, H.J. Feine, A. Hulttermann, R. Frund, H.D. Luldemann, J.J. Meister, Macromolecules 29 (1996) 1389–1398.
S.P. Maradur et al. / Synthetic Metals 162 (2012) 453–459 [16] C. Bonini, M. D’auria, R. Ferri, R. Pucciariello, A.R. Sabia, J. Appl. Polym. Sci. 90 (2003) 1163–1171. [17] J.J. Meister, C.T. Li, Macromolecules 25 (1992) 611–616. [18] M.N.M. Ibrahim, M.R. Ahmed-Haras, C.S. Sipaut, H.Y. Aboul-Enein, A.A. Mohamed, Carbohydr. Polym. 80 (2010) 1102–1110. [19] F.G. Calvo-Flores, J. Dobado, ChemSusChem 3 (2010) 1227–1235. [20] W.G. Glasser, R.A. Northey, T.P. Schultz, Lignin: Historical, Biological and Materials Perspectives, American Chemical Society, Washington, DC, 2000. [21] S.S. Kelley, T.C. Ward, W.G. Glasser, J. Appl. Polym. Sci. 41 (1990) 2813–2828. [22] H. Hatakeyama, S. Hirose, T. Hatakeyama, K. Nakamura, K. Kobashigawa, N. Morohoshi, J. Macromol. Sci., Pure Appl. Chem. 32 (1995) 743–750. [23] R.W. Thring, M.N. Vanderlaan, S.L. Griffin, Biomass Bioenergy 13 (1997) 125–132. [24] V. Marchetti, P. Gerardin, P. Tekely, B. Loubinoux, Horzfors-chung 52 (1998) 654–660. [25] J.H. Lora, W.G. Glasser, J. Polym. Environ. 10 (2002) 39–48.
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]
459
J.J. Meister, M.-J. Chen, Macromolecules 24 (1991) 6843–6848. L.R.C. Barclay, F. Xi, J.Q. Norris, J. Wood Chem. Technol. 17 (1997) 73–90. F.J. Lu, L.-H. Chu, R.-J. Gau, Nutr. Cancer 30 (1998) 31–38. W.O.S. Doherty, P. Mousavioun, C.M. Fellows, Ind. Crops Prod. 33 (2011) 259–276. V.R. Pai Verneker, B. Shaha, Macromolecules 19 (1986) 1851–1856. N. Yusof, A.F. Ismail, Int. J. Chem. Environ. Eng. 1 (2010) 79–84. A.F. Ismail, M.A. Rahman, A. Mustafa, T. Matsuura, Mater. Sci. Eng. A 485 (2008) 251–257. C.-g. Wang, X.-g. Dong, Q.-f. Wang, J. Polym. Res. 16 (2009) 719–724. W. Liu, G. Gao, J. Appl. Polym. Sci. 93 (2004) 956–960. J. Chen, C.W.H. Ge, Y. Bai, Y. Wang, J. Polym. Res. 14 (2007) 223–228. W.-x. Zhang, Y.-z. Wang, C.-f. Sun, J. Polym. Res. 14 (2007) 467–474. D. Esrafilzadeh, M. Morshed, H. Tavanai, D. Esrafilzadeh, M. Morshed, H. Tavanai, Synth. Met. 159 (2009) 267–272. J.L. Braun, K.M. Holtman, J.F. Kadla, Carbon 43 (2005) 385–394.