- Email: [email protected]
UV assisted stabilization routes for carbon fiber precursors produced from melt-processible polyacrylonitrile terpolymer
Carbon 43 (2005) 1065–1072 www.elsevier.com/locate/carbon
UV assisted stabilization routes for carbon fiber precursors produced from melt-processible polyacrylonitrile terpolymer Amit K. Naskar, Robert A. Walker, Sarah Proulx, Dan D. Edie, Amod A. Ogale
*
Department of Chemical Engineering, Center for Advanced Engineering Fibers and Films, Clemson University, 203 Earle Hall, Clemson, SC 29634-0910, USA Received 11 May 2004; accepted 28 November 2004 Available online 25 January 2005
Abstract A low-cost route for producing PAN-based carbon fibers is being developed. The approach involves forming polyacrylonitrile terpolymers that can be melt-spun into fibers. The fibers are then stabilized and carbonized to yield carbon fibers. Melt-processibility, however, precludes direct thermal stabilization of these polymeric fibers. Therefore, a precursor terpolymer containing acryloyl benzophenone (ABP) is used. The UV sensitivity of ABP moiety enhances the UV crosslinkability of the precursor fibers. After a brief exposure to UV radiation, the melt-spun terpolymer fibers can be oxidatively stabilized at 320 C without melting and subsequently carbonized. UV–visible and ATR-IR spectroscopic analyses suggest that UV radiation induces the formation of free radicals which, in turn, cyclize the PAN. Cyclized PAN was characterized by a strong absorbance in UV–visible region (300–500 nm) due to conjugated >C@C< and >C@N– bonds which were also detected by infrared spectroscopy. 2004 Elsevier Ltd. All rights reserved. Keywords: Carbon fibers; Stabilization; Infrared spectroscopy
1. Introduction Solution spinning is used to produce the precursor fibers for commercial PAN-based carbon fibers [1–6]. Although more costly and more hazardous than melt spinning, solution spinning is necessary because the PAN copolymers employed in these processes degrade before they melt. Once solution-spun, the fibers are heated to 300 C in air to induce cyclization reactions. This converts the linear copolymer to infusible, sixmembered ring structures that do not melt during the final carbonization step. Several research teams have used external plasticization to lower the polymer softening point below its decomposition temperature and, thereby, create a *
Corresponding author. Tel.: +1 864 656 5483; fax: +1 864 656 0784. E-mail address: [email protected] (A.A. Ogale). 0008-6223/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2004.11.047
pseudo melt spinning process [7–10]. However, the plasticizer is difficult to remove at commercial extrusion rates. As a result the pseudo melt-spun precursor fibers often contain voids, reducing the strength of final carbon fibers. A true melt spinning route for PAN precursor fibers could reduce processing costs by eliminating solvent handling and recovery as well as the void formation problem that has plagued the external plasticization approach. Melt spinning also offers the additional advantage of higher process throughput. PAN copolymers of higher comonomer content are, in fact, melt processible. The presence of comonomers in the polymer chain disrupts long-range order in PAN, enabling the polymer to soften before it degrades [11]. Incorporation of comonomers such as methyl acrylate (MA), acrylic acid (AA), and vinyl acetate (VA), at 10–15 mol%, lower the softening point and the viscosity of the PAN copolymer [12] and make the precursor melt spinnable [13–15]. However, melt-spun PAN fibers
1066
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
cannot be subjected to conventional thermal stabilization in air because the fibers melt at lower temperatures than that required to initiate the oxidative reactions. In previous research we have found that photo-induced crosslinking can eliminate premature melting of precursor fibers during stabilization [16]. Photo-induced crosslinking of PAN has been reported in the literature [17–26]. Photo-oxidation of PAN, especially at elevated temperatures, forms a cyclized ladder polymer structure [17,18]. Electron beam irradiation generates free radicals in PAN, which aid cyclization [19]. Radical-induced cyclization and intermolecular crosslinking through the nitrile groups has been reported for PAN via ion beam [20], X-ray [21,22], gamma [23–25], and UV [20,26] irradiation. Although we have demonstrated that UV-assisted stabilization of melt-spun PAN copolymers was feasible [16], the process required a long stabilization time. Conventional melt-spun PAN filaments displayed voids and bubbles throughout the crosssection, including the surface. These defects were generated due to inadequate crosslinking. In the present study, we report on the accelerated UV crosslinking of a terpolymer containing a UV sensitive comonomer, acryloyl benzophenone (ABP), which was synthesized in a companion study by Banu et al. and McGrath et al. [27,28]. The compositions of the polymers used in this study were poly[acrylonitrile (AN)co-methyl acrylate (MA)-co-acryloyl benzophenone (ABP)] of molar ratio 85:14:1 (hereafter denoted as ÔterpolymerÕ) and its control polymer without ABP of 85:15 molar ratio (hereafter denoted as ÔcopolymerÕ). For carbonized fibers obtained from the terpolymer, tensile modulus and strength of 140 GPa and 700 MPa, respectively, were reported in a recent study [29]. The modulus values were close to those found in carbon fibers obtained from conventional, solution-spinning processes. The lateral surfaces of carbonized terpolymer-based fiber were smooth and defect-free (in contrast to copolymer-based fibers), but occlusions were observed in the core where the crosslinking was inadequate, particularly for thicker fibers whose diameter was greater than about 25 lm. Ongoing studies are being conducted to improve carbon fiber properties by reducing defect generation during the heat treatment process. The present study seeks to elucidate the role of ABP in UV-assisted crosslinking and stabilization of the melt-processible precursor through spectroscopic measurements.
2. Experimental 2.1. Materials Two compositions of PAN precursor polymers having AN:MA:ABP mole ratio of 85:14:1 (terpolymer)
X
H3CO
C
N
AN
O
Y
O
O
Z
MA
C
O
ABP Fig. 1. Molecular structure of the terpolymer showing AN, MA and ABP segments.
and AN:MA, 85:15 (copolymer) were used throughout the study (Monomer Polymer, Inc., Feasterville, PA). The synthesis recipes were developed and specified by McGrath and coworkers, and the comonomers were shown to be present in random statistical distribution in the copolymers [15]. The number average molecular weight (Mn) for the terpolymer and copolymers were 22,000 and 26,500, respectively. The molecular structure of the terpolymer is shown in Fig. 1. 2.2. Differential scanning calorimetry (DSC) The thermal stability of the precursor powder samples was evaluated by differential scanning calorimetry (Pyris 1, Perkin Elmer). Isothermal scans were conducted over a temperature range of 220–300 C. Constant rate scans were performed at heating rates of 5, 10, and 20 C/min in a nitrogen atmosphere. The kinetic parameters of reaction were evaluated by analyzing thermograms obtained at various heating rates (5– 20 C/min) according to KissingerÕs method [16,30]. The shape factor of the exotherm peak gives the kinetic order of the reaction [30]. The temperature corresponding to maximum reaction rate (Tmax)for different heating rates (/) were noted. The slopes of ln½/=T 2max vs 1/Tmax plots were used to calculate activation energy for the cyclization reaction. 2.3. Processing of precursor Powder precursors were compacted at 25 C to obtain 1 mm thick disks using a pelletizing die and a Carver model 2926 laboratory press. Pelletized precursors were compressed at 20 MPa and 220 C to obtain 3 cm · 6 cm films with a thickness of 50–60 lm for spectroscopic studies. The precursors were UV irradiated using low power (100 W) and high power (4.5 kW) mercury arc lamps for 3 h and 50 s durations, respectively, at 100 C. After UV stabilization, the terpolymer
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
A Nicolet Avatar 360 FTIR spectrometer, in the transition mode, was used to collect infrared spectra of the terpolymer and copolymer precursors. Attenuated total reflectance (ATR) IR spectra of the precursors were measured using a Nicolet Magna 550 FTIR spectrometer with a high endurance diamond ATR attachment to determine any changes in functionality of the film surfaces caused by UV irradiation. The UV–visible absorption spectra were recorded on a Shimadzu UV3100 spectrophotometer at 25 C. 2.5. Gel fraction Gel fractions of the melt processed and the UV crosslinked polymers were measured after equilibrium swelling (48 h) in dimethyl sulphoxide (DMSO) solvent (Aldrich Chemical Company, Inc. Milwaukee, WI). The swollen gel was dried in vacuum at 50 C for 5 days and weighed to obtain the mass fraction relative to the original mass.
3. Results and discussion
Heat flow/unit mass (w/g) [endoup]
2.4. Spectroscopy
Terpolymer 240°C Copolymer 240°C Terpolymer 220°C Copolymer 220°C
0.00
-0.03
-0.06
-0.09
200
400
600
800
Time (min)
(a)
0.0
Heat flow/unit mass (w/g) [endoup]
precursor film could be oxidatively stabilized by heating at a rate of 2.5 C/min to a temperature of 320 C and then holding at this temperature for 30 min in a forced-circulation oven (ATS 3610). By contrast, the maximum heating rate that could be used to oxidatively stabilize the copolymer precursors was 0.2 C/min. Thus, the photo-crosslinked terpolymer enables a significantly higher stabilization rate than does the copolymer.
1067
Terpolymer Copolymer Terpolymer Copolymer Terpolymer Copolymer
-0.5
260°C 260°C 280°C 280°C 300°C 300°C
-1.0
-1.5 25
(b)
50
75
100
125
Time (min)
Fig. 2. Isothermal DSC thermograms of PAN-based terpolymer and copolymer at different temperatures: (a) 220 and 240 C; (b) 260, 280 and 300 C.
3.1. Thermal stability of precursors Isothermal DSC scans, shown in Fig. 2(a), indicate that at 220 C both the copolymer and the terpolymer are thermally stable for about four hours. However, after this period of stability they show a broad exotherm extending up to 7 h. The order of the overall reaction, as measured by KissingerÕs method, for both the compositions was close to unity (1.16 ± 0.13 for terpolymer and 0.94 ± 0.13 for copolymer). The activation energies (Ea) for the terpolymer and copolymer compositions were also similar (137 and 131 kJ/mol, respectively), and consistent with earlier reports on PAN copolymers used as carbon fiber precursors [31]. These results establish that, in spite of the incorporation of ABP, the terpolymer and the copolymer exhibited similar cyclization kinetics and melt stability. At 220 C, exotherm maxima were observed after about 9 h (540 min). The times corresponding to these exotherm maxima and the heats of reaction (DH) at
Table 1 Isothermal DSC data of precursor powders showing exothermic DH and time corresponding to exotherm maxima Temperature (C)
Terpolymer (DH, J/g; Tmax, min)
Copolymer (DH, J/g; Tmax, min)
220 240 260 280 300
255 ± 10; 265 ± 20; 360 ± 20; 390 ± 10; 405 ± 15;
255 ± 10; 290 ± 20; 400 ± 15; 425 ± 15; 440 ± 10;
545 180 67 30 15
520 170 64 27 14
different temperature scans for both the precursors are summarized in Table 1. At 220 and 240 C, the isothermal DHs for the two compositions were not significantly different (95% confidence level). However, at higher temperatures (260–300 C, Fig. 2(b)) the heat of reaction for the terpolymer was less than that of the copolymer. Although the non-AN comonomer content was 15 mol% in both the terpolymer and the copolymer,
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
the DH/mass of terpolymer was expected to be about 3% less than that of the copolymer because of the higher molar mass of ABP compared to MA. Surprisingly, measurements showed that the DH/mass for the terpolymer was about 10% less than that for the copolymer. Apparently the random distribution of monomer segments disrupts the sequence of PAN more in the case of terpolymer than the copolymer. As a result, fewer adjacent AN repeat units are available to cyclize. For low-temperature scans, the DH values are low due to incomplete cyclization and pyrolytic reaction of PAN that finally leads to ladder structure, consistent with literature studies [32,33].
After ultraviolet (UV) irradiation, the terpolymer precursor turned dark brown, whereas the copolymer developed only a pale brownish tint. Irradiated precursors were immersed in DMSO solvent and the gel contents were measured after equilibrium swelling (48 h). Fig. 3 shows the gel content results for low- and highirradiance (0.03 and 1.0 W/cm2, respectively) exposures. The non-irradiated samples dissolved completely. By comparison, the copolymer sample remained almost completely soluble, even after 180 min of low-power UV irradiation. High-power UV irradiation of copolymer for 1 min yielded only 5% gel content. However, UV irradiated terpolymer produced 30% and 65% gel content after low- and high-power irradiation, respectively. Thus, while low-power radiation required about 60 min for an equilibrium level of crosslinking to occur, high-power radiation achieved a significantly higher level of crosslinking in a considerably shorter duration of 1 min. This establishes the role of ABP moiety in accelerating the crosslinking of the precursor and providing an adequate degree of crosslinking (over
1.0
Terpolymer high power UV Terpolymer low power UV Copolymer high power UV
Gel fraction
Copolymer low power UV 0.6
0.4
0.2
0.0 0.0
0.5
1.0
50
1 min high power (Terpolymer) 180 min low power (Terpolymer) 180 min low power (Copolymer) 180 min 110°C, no UV (Terpolymer) No UV (Terpolymer) No UV (Copolymer)
2.0
1.5
1.0
0.5
0.0 250
300
350
400
450
500
550
600
650
700
Wavelength (nm)
3.2. Influence of UV irradiation
0.8
2.5
Absorbance
1068
100
150
200
Time (min) Fig. 3. Gel fraction of PAN-based terpolymer and copolymer samples after different durations of high- and low-power UV irradiation.
Fig. 4. UV–visible spectroscopy of PAN-based terpolymer and copolymer samples after UV irradiation at different conditions.
60% gel), which prevented the melting of samples during subsequent oxidative stabilization. Next, the UV sensitivity of the precursors was systematically investigated by UV–visible spectroscopic measurements. Fig. 4 displays the UV–visible spectra for terpolymer relative to that of the copolymer at different conditions of UV irradiation. Before irradiation, it was observed that the terpolymer displayed enhanced absorbance at 330 nm relative to that of the copolymer. This difference in the spectra is due to the UV sensitivity of the benzophenone moiety in the terpolymer, and is consistent with literature studies on polymerized benzophenone derivatives [34]. The control terpolymer sample, which was not UV irradiated but underwent only the heating cycle (110 C for 180 min), remained almost colorless and did not show any further change in the UV–visible spectrum. Thus, thermal exposure at low temperature (110 C) does not significantly affect the molecular structure of the terpolymer. After low-power UV irradiation for 180 min, the absorbance of the copolymer appears at wavelengths lower than 300 nm, having a low-level tail extended in the violet region (425 nm) (Fig. 4). The low level of absorbance in the violet region is consistent with the development of only a brownish tint (rather than a significant change of color) for the copolymer. In contrast, the color generation after UV exposure is dramatic for the terpolymer. The terpolymer turned dark brown indicating a strong UV and violet light absorbance. In fact, when the terpolymer was irradiated for only 1 min in the high-power UV source (in contrast to 180 min in low-power), an abrupt violet–blue light absorbance occurred at 450 nm. The intense brown color for the terpolymer likely results from the UV-assisted cyclization reactions that produce chromophores. The UV assisted reactions were accelerated during high power UV irradiation (for short duration 1 min) and the extent of
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
reaction is more than that observed after low-power UV irradiation for 180 min. Takata and Hiroi [35] showed that, for model di-, tri-, and tetrameric acrylonitrile compounds (increasing number of ACN groups), multiple UV absorbance maxima, kmax, appeared after intra-molecular cyclization reactions. They also found that these kmax were higher for the larger ring structures. The kmax at 243, 292 and 336 nm were characteristic of cyclic structures consisting 2, 3, and 4 –CN groups, respectively. Thus, a distribution of kmax would be expected for various ring structures. The spectra displayed in Fig. 4 suggest that the
1.8
Absorbance
1.5
Absorbance
2.0
0h 0.5h 1h 3h 7h
1.5 1.0
0s 2.5 s 12.5 s 50 s
0.5 1700
1650
1600
Wavenumber (cm-1) 1657
1.2
1599
1687 1624
0.9 1579
0.6
1700
1680
1660
1640
1620
1600
1580
Wavenumber (cm-1 ) Fig. 5. Transmittance FTIR spectra of PAN-based terpolymer samples after high- and low-power (inset) UV irradiation for different durations.
C
1069
terpolymer readily forms a number of larger ring structures compared to the copolymer. Similar electronic absorption bands were reported for c irradiated PAN [24]. Transmittance infrared spectra of terpolymer precursor samples irradiated in high power UV beam for short periods are shown in Fig. 5. Peaks caused by CH2 and CH stretching (2900–2950 cm1), CN stretching (2242 cm1, due to PAN segments) and C@O stretching (1730 cm1, due to MA segments) were intense and appeared truncated due to saturation of the signal (not shown in Fig. 5). Non-irradiated samples showed a strong peak at 1657 cm1 due to >C@O groups in conjunction with phenyl rings of the ABP termonomer. The intensity of the peak decreased after UV irradiation for 2.5 s due to its conversion to an excited triplet state (through an intermediate singlet state), which creates free radicals. As proposed in the model reaction scheme of Fig. 6, these radicals may convert to secondary alcohols by abstracting hydrogen radicals from the adjacent polymer chain, thereby forming a crosslinked network. Hydrogen abstraction by the triplet state of benzophenone moiety is reported in the literature [28,36,37]. As displayed in Fig. 5 (inset), similar changes in IR spectra of the terpolymer were also observed during prolonged low-power UV irradiation. Gradual decrease in 1657 cm1 peak intensity with increasing UV exposure time (up to 1 h) can be attributed to the slow hydrogen radical abstraction reaction during low-power irradiation. After the disappearance of 1657 cm1 peak from ABP, a broad shoulder appeared at the same position of the spectra (Fig. 5) on further UV irradiation. The
C
O
O
O
O
+ C
H
R
H
R
O *
∗
+
R
CH
OH
R
H
R ϖ
R
H
Excited state ABP moiety after UV irradiation
Polymer chain
Polymer chain with free radicals
Crosslinking Polymer chains with free radicals
Crosslinked polymer network
Fig. 6. Model of free radical induced crosslinking reaction in PAN-based terpolymer.
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
shoulder was due to the formation of conjugated >C@O groups that are generated by the UV induced degradation and recombination in MA segments, consistent with that reported in the literature [38]. Other >C@O functionalities are also generated after UV irradiation as evidenced by formation of 1780 (not shown in Fig. 5) and 1689 cm1 shoulders. A small shoulder at 1624 cm1, as observed in the spectra of non-irradiated sample, may be attributed to the presence of conjugated >C@C< (with acrylate >C@O) in the precursors. These >C@C< groups may be generated during free radical polymerization via disproportionation reaction. After UV exposure, this shoulder intensity decreased initially, due to the reaction of >C@C< with the free radicals. In the case of the terpolymer, the peak at 1599 cm1 is due to a combination of C–C stretching (mCC) and C– C–C in-plane deformation (dCCC) of ABP [39]. With initial UV irradiation, as the planer carbon of benzophenone >C@O group becomes tetrahedral (due to sp3 hybridization) the overall resonance stabilization is lost and dCCC becomes restricted. As a result, the 1599 cm1 peak decreases. However, during the course of UV irradiation, different types of conjugated >C@C< groups are formed. This would account for the broad peak that appears at 1599 cm1 and increases with time. Diamond ATR FTIR spectra of high power UV irradiated and the non-irradiated control samples are shown in Fig. 7(a). In the ATR mode, because of a limited depth of penetration of about 2 lm, changes in the concentration of some abundant functional groups become apparent in this mode. It was observed that the ACN (2242 cm1) stretching absorbance is lowered after high power UV irradiation for 25 s while a broad overlapped peak appears in the frequency range 1560– 1690 cm1. This suggests that the UV assisted reaction causes a decrease in CN groups and development of unsaturations. Fig. 7(b) displays variation of peak intensities caused by CH2 asymmetric stretching (2955 cm1), as well as ACN (2242 cm1), >C@C< (1624 cm1) and >C@N– (1579 cm1) vibrations, respectively, normalized with respect to C–H stretching (2926 cm1). During the course of UV irradiation, the normalized intensity of CH2 asymmetric stretching (2955 cm1) remains almost unaltered. During crosslinking reaction, free radical generation is expected to occur at C–H sites because tertiary radicals are the most stable. Evidently, the change in C–H concentration is small, compared to the total concentration of C–H bonds. Utilization of ACN group with gradual development at 1624 cm1 and 1579 cm1 may be attributed to the formation of AC@C–C@N structure through cyclization of PAN [32]. A schematic of the plausible reactions is shown in Fig. 8. UV radiation creates free radicals in the precursor, which assist in the partial cyclization of PAN.
Absorbance 0.070 terpol 0 pass 0.060
Terpolymer-no UV irradiation
0.050 0.040
2242
0.030
1657 1599
0.020 0.010 0.000 0.070 terpol 20 passes 0.060 0.050
Terpolymer-50 s high power UVirradiation
0.040 0.030 0.020 0.010 0.000
2000
2500
3000
Wavenumbers (cm-1)
(a)
2242 cm-1 2955 cm-1 1624 cm-1 1579 cm-1
1.2
Normalized absorbance.
1070
0.9
0.6
0.3
0 0
10
(b)
20
30
40
50
60
UV exposure time (s)
Fig. 7. (a) ATR-IR spectra of PAN-based terpolymer after highpower UV irradiation. (b) ATR peak intensity (normalized with respect to C–H stretching at 2926 cm1) as a function of high power UV irradiation time for CH2 asymmetric stretching (2955 cm1), C„N stretching (2242 cm1), C@C stretching (1624 cm1), and C@N stretching (1579 cm1) (solid curves are trendlines).
Thermal treatment converts cyclized PAN to a ladder structure. Aggour and Aziz also proposed a similar free radical induced cyclization of PAN [20]. They also proposed that UV and ion beam irradiation of PAN would generate free radicals. In the present study, involving a systematically-designed terpolymer, we find that the ABP moiety accelerates the generation of free radicals, which is consistent with free radical induced crosslinking of precursors discussed earlier (Fig. 3).
4. Conclusions The acrylonitrile (AN)–methyl acrylate (MA)– acryloyl benzophenone (ABP) terpolymer precursor (85:14:1 mole ratio) was found to be thermally stable and melt processible at 220 C. An accelerated UV stabilization route was developed for the precursor, where ABP acts as the UV sensitive component. Free radicals generated during UV irradiation of the terpolymer enhance crosslinking and cyclization of PAN, as evidenced
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072 P∗
P∗
∗
1071
R
H
R + H
P
∗
Polymer chains or their segments containing free radical
P∗
+
CH
CH
CH
CH
C
C
C
C
N
+ N
N
H∗
N
Free radical induced cyclization
CH
CH
CH
C
C
C
N
P
N
CH N
C
NH
C
C
C
C
C
C
C
C
N H
P
N H
N H
Dehydrogenation NH 2
∆
P
C
C
C
C
C
C
N
N
C N
C
NH
Ladder polymer
C CH P
N H
C
C
CH
CH N H
C CH N H
NH 2
Fig. 8. Schematic of model reactions of free radical induced cyclization of PAN.
by UV-visible and IR spectroscopy. In contrast, the copolymer precursor could not be stabilized to any significant extent, because it lacks UV sensitivity. Stabilization of terpolymer was significantly accelerated by exposure to a high energy density UV beam for a very short period (1 min). The UV assisted crosslinking and cyclization inhibits melting of melt-processed terpolymer during subsequent thermo-oxidative stabilization.
Acknowledgments The authors gratefully acknowledge Mr. Dave Warren, DOE program monitor, and the financial support from the Department of Energy through contract no. 4500011036. We also thank Profs. Don Baird and Jim McGrath, Virginia Tech for providing the precursor polymers and Kim Ivey, Clemson University for assistance with diamond ATR spectroscopy.
References [1] Gupta A, Harrison IR. New aspects in the oxidative stabilization of PAN-based carbon fibers. Carbon 1996;34(11):1427–45. [2] Grassie N, Hay JN. Thermal coloration and insolubilization in polyacrylonitrile. J Polym Sci 1962;56:189–202. [3] Bashir Z. A critical review of the stabilization of polyacrylonitrile. Carbon 1991;29(8):1081–90. [4] Edie DD. The effect of processing on the structure and properties of carbon fibers. Carbon 1998;36(4):345–62. [5] Odawara K, Takeuchi S. Process for the melt shaping of acrylonitrile polymers. US Patent 4,108,818, 1978.
[6] Sen K, Bahrami SH, Bajaj P. High-performance acrylic fibers. J Macromol Sci, Rev Macromol Chem Phys 1996;C36(1):1–76. [7] Grove D, Desai P, Abhiraman AS. Exploratory experiments in the conversion of plasticized melt spun PAN-based precursors to carbon fibers. Carbon 1988;26(3):403–11. [8] Daumit GP, Ko YS, Slater CR, Venner JG, Young CC, Zwick MM. Formation of melt-spun acrylic fibers suitable for thermal conversion to high strength carbon fibers. US Patent 4,933,128, 1990. [9] Bashir Z, Atureliya SK, Church SP. Production of oriented polyacrylonitrile films by flow-induced chain extension and crystallization from solution. J Mater Sci 1993;28(10):2721–32. [10] Fishman N. Preparation of acrylonitrile polymer compositions and the molded articles. US Patent 5,434,205, 1995. [11] Frushour BG. Acrylic polymer characterization in the solid state and in solution. In: Masson JC, editor. Acrylic fiber technology and applications. New York: Marcel Dekker, Inc; 1995. p. 218. [12] Gupta AK, Paliwal DK, Bajaj P. Acrylic precursors for carbon fibers. J Macromol Sci, Rev Macromol Chem Phys 1991;C31: 1–89. [13] Smierciak RC, Wardlow Jr E, Ball LE. Process for making a melt processable high nitrile multipolymer of acrylonitrile and olefinically unsaturated monomers. US Patent 5,618,901, 1997. [14] Davidson JA, Jung HT, Hudson SD, Percec S. Investigation of molecular orientation in melt-spun high acrylonitrile fibers. Polymer 2000;41:3357–64. [15] Bhanu VA, Rangarajan P, Wiles K, Bortner M, Sankarpandian M, Godshall D, et al. Synthesis and characterization of acrylonitrile methyl acrylate statistical copolymers as melt processable carbon fiber precursors. Polymer 2002;43:4841–50. [16] Paiva MC, Kotasthane P, Edie DD, Ogale AA. UV stabilization route for melt-processible PAN-based carbon fibers. Carbon 2003;41:1399–409. [17] Ranby B, Rabek JF. Photodegradation photo-oxidation and photostabilization of polymers: principles and applications. London: Wiley; 1975. p. 190–2. [18] Bullock RE. Radiation-enhanced oxidation of PAN-based carbon fibers. Fiber Sci Technol 1974;7(2):157–60.
1072
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
[19] Dietrich J, Hirt P, Herlinger H. Electron-beam-induced cyclisation to obtain C-fiber precursors from polyacrylonitrile homopolymers. Eur Polym J 1996;32(5):617–23. [20] Aggour YA, Aziz MS. Degradation of polyacrylonitrile by low energy ion beam and UV radiation. Polym Test 2000;19:261–7. [21] Murthy MR, Radhakrishna S. Radiation damage in polyacrylonitrile. Pramana 1983;20(1):85–90. [22] Cardoso dos Santos LG, Kawano Y. Degradation of polyacrylonitrile by X-ray radiation. Polym Degrad Stab 1994;44(1):27–32. [23] Badawy SM, Dessouki AM. Cross-linked polyacrylonitrile prepared by radiation-induced polymerization technique. J Phys Chem B 2003;107(41):11273–9. [24] Zhao W, Yamamoto Y, Tagawa S. Regulation of the thermal reactions of polyacrylonitrile by c-irradiation. JAERI Conference 2001 (JCBSRC Õ99, the 8th Japan–China Bilateral symposium on radiation chemistry, 1999) 2000, p. 122–5. [25] Hill DJT, Lang AP, OÕDonnell JH, Pomery PJ. The effects of cradiation on polyacrylonitrile. Polym Degrad Stab 1992;38(3):193–203. [26] Stephenson CV, Lacey Jr JC, Wilcox WS. Ultraviolet irradiation of plastics III. Decomposition products and mechanisms. J Polym Sci 1961;55:477–88. [27] Bhanu VA, Wiles KB, Bortner M, Glass TE, Godshall D, Baird DG, et al. Designing the molecular composition of the melt processable acrylonitrile copolymer carbon fiber precursors. Polym Prep 2002;43(2):1017–8. [28] Mukundan T, Bhanu VA, Wiles KB, Bortner M, Baird DG, McGrath JE. Photocrosslinkable acrylonitrile terpolymers as carbon fiber precursors. Polym Prepr 2003;44(2):651–2. [29] Naskar A, Proulx S, Walker K, Edie D, Ogale A. UV stabilization of melt-processible PAN carbon precursor fibers. Abstracts, 55th
[30] [31]
[32]
[33]
[34]
[35] [36] [37] [38]
[39]
Southeast Regional Meeting of the American Chemical Society, Atlanta, GA, United States, 16–19 November, 2003, p. 869. Kissinger HE. Reaction kinetics in differential thermal analysis. Anal Chem 1957;29(11):1702–6. Fitzer E, Mu¨ller DJ. The influence of oxygen on the chemical reactions during stabilization of PAN as carbon fiber precursor. Carbon 1975;13:63–9. Xue G, Dong J, Zhang J. SERS studies of polymers on metal surfaces. 4. Cyclization and graphitization of polyacrylonitrile on rough silver surfaces. Macromolecules 1992;25:5855–7. Leroy S, Boiziau C, Perreau J, Reynaud C, Zalczer G, Lecayon G, et al. Molecular structure of an electropolymerized polyacrylonitrile film and its pyrolyzed derivatives. J Mol Struct 1985;128(4):269–81. Oh JK, Wu J, Winnik MA, Craun GP, Rademacher J, Farwaha R. Polymerizable benzophenone derivatives for labeling vinyl acetate-butyl acrylate latex particles. J polym Sci, Part A, Polym Chem 2002;40(17):3001–11. Takata T, Hiroi I. Coloration in acrylonitrile polymers. J Polym Sci, Part A 1964;2:1567–85. Walling C, Gibian MJ. Hydrogen abstraction by the triplet state of benzophenone. J Am Chem Soc 1964;86:3902–3. Naguib YMA, Cohen SG, Steel C. Reduction of crystal violet by diphenylketyl radicals. J Am Chem Soc 1986;108:128–33. Cameron GG, Kane DR. The thermal degradation of poly (methyl acrylate) III. Mechanism of volatile formation. Die Makromolekulare Chemie 1968;113:75–84. Kolev TM, Stamboliyska BA. Vibrational spectra and structure of benzophenone and its 18O and d10 labelled derivatives: an ab initio and experimental study. Spectrochim Acta Part A 1999;56:119– 126.
UV assisted stabilization routes for carbon fiber precursors produced from melt-processible polyacrylonitrile terpolymer Amit K. Naskar, Robert A. Walker, Sarah Proulx, Dan D. Edie, Amod A. Ogale
*
Department of Chemical Engineering, Center for Advanced Engineering Fibers and Films, Clemson University, 203 Earle Hall, Clemson, SC 29634-0910, USA Received 11 May 2004; accepted 28 November 2004 Available online 25 January 2005
Abstract A low-cost route for producing PAN-based carbon fibers is being developed. The approach involves forming polyacrylonitrile terpolymers that can be melt-spun into fibers. The fibers are then stabilized and carbonized to yield carbon fibers. Melt-processibility, however, precludes direct thermal stabilization of these polymeric fibers. Therefore, a precursor terpolymer containing acryloyl benzophenone (ABP) is used. The UV sensitivity of ABP moiety enhances the UV crosslinkability of the precursor fibers. After a brief exposure to UV radiation, the melt-spun terpolymer fibers can be oxidatively stabilized at 320 C without melting and subsequently carbonized. UV–visible and ATR-IR spectroscopic analyses suggest that UV radiation induces the formation of free radicals which, in turn, cyclize the PAN. Cyclized PAN was characterized by a strong absorbance in UV–visible region (300–500 nm) due to conjugated >C@C< and >C@N– bonds which were also detected by infrared spectroscopy. 2004 Elsevier Ltd. All rights reserved. Keywords: Carbon fibers; Stabilization; Infrared spectroscopy
1. Introduction Solution spinning is used to produce the precursor fibers for commercial PAN-based carbon fibers [1–6]. Although more costly and more hazardous than melt spinning, solution spinning is necessary because the PAN copolymers employed in these processes degrade before they melt. Once solution-spun, the fibers are heated to 300 C in air to induce cyclization reactions. This converts the linear copolymer to infusible, sixmembered ring structures that do not melt during the final carbonization step. Several research teams have used external plasticization to lower the polymer softening point below its decomposition temperature and, thereby, create a *
Corresponding author. Tel.: +1 864 656 5483; fax: +1 864 656 0784. E-mail address: [email protected] (A.A. Ogale). 0008-6223/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.carbon.2004.11.047
pseudo melt spinning process [7–10]. However, the plasticizer is difficult to remove at commercial extrusion rates. As a result the pseudo melt-spun precursor fibers often contain voids, reducing the strength of final carbon fibers. A true melt spinning route for PAN precursor fibers could reduce processing costs by eliminating solvent handling and recovery as well as the void formation problem that has plagued the external plasticization approach. Melt spinning also offers the additional advantage of higher process throughput. PAN copolymers of higher comonomer content are, in fact, melt processible. The presence of comonomers in the polymer chain disrupts long-range order in PAN, enabling the polymer to soften before it degrades [11]. Incorporation of comonomers such as methyl acrylate (MA), acrylic acid (AA), and vinyl acetate (VA), at 10–15 mol%, lower the softening point and the viscosity of the PAN copolymer [12] and make the precursor melt spinnable [13–15]. However, melt-spun PAN fibers
1066
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
cannot be subjected to conventional thermal stabilization in air because the fibers melt at lower temperatures than that required to initiate the oxidative reactions. In previous research we have found that photo-induced crosslinking can eliminate premature melting of precursor fibers during stabilization [16]. Photo-induced crosslinking of PAN has been reported in the literature [17–26]. Photo-oxidation of PAN, especially at elevated temperatures, forms a cyclized ladder polymer structure [17,18]. Electron beam irradiation generates free radicals in PAN, which aid cyclization [19]. Radical-induced cyclization and intermolecular crosslinking through the nitrile groups has been reported for PAN via ion beam [20], X-ray [21,22], gamma [23–25], and UV [20,26] irradiation. Although we have demonstrated that UV-assisted stabilization of melt-spun PAN copolymers was feasible [16], the process required a long stabilization time. Conventional melt-spun PAN filaments displayed voids and bubbles throughout the crosssection, including the surface. These defects were generated due to inadequate crosslinking. In the present study, we report on the accelerated UV crosslinking of a terpolymer containing a UV sensitive comonomer, acryloyl benzophenone (ABP), which was synthesized in a companion study by Banu et al. and McGrath et al. [27,28]. The compositions of the polymers used in this study were poly[acrylonitrile (AN)co-methyl acrylate (MA)-co-acryloyl benzophenone (ABP)] of molar ratio 85:14:1 (hereafter denoted as ÔterpolymerÕ) and its control polymer without ABP of 85:15 molar ratio (hereafter denoted as ÔcopolymerÕ). For carbonized fibers obtained from the terpolymer, tensile modulus and strength of 140 GPa and 700 MPa, respectively, were reported in a recent study [29]. The modulus values were close to those found in carbon fibers obtained from conventional, solution-spinning processes. The lateral surfaces of carbonized terpolymer-based fiber were smooth and defect-free (in contrast to copolymer-based fibers), but occlusions were observed in the core where the crosslinking was inadequate, particularly for thicker fibers whose diameter was greater than about 25 lm. Ongoing studies are being conducted to improve carbon fiber properties by reducing defect generation during the heat treatment process. The present study seeks to elucidate the role of ABP in UV-assisted crosslinking and stabilization of the melt-processible precursor through spectroscopic measurements.
2. Experimental 2.1. Materials Two compositions of PAN precursor polymers having AN:MA:ABP mole ratio of 85:14:1 (terpolymer)
X
H3CO
C
N
AN
O
Y
O
O
Z
MA
C
O
ABP Fig. 1. Molecular structure of the terpolymer showing AN, MA and ABP segments.
and AN:MA, 85:15 (copolymer) were used throughout the study (Monomer Polymer, Inc., Feasterville, PA). The synthesis recipes were developed and specified by McGrath and coworkers, and the comonomers were shown to be present in random statistical distribution in the copolymers [15]. The number average molecular weight (Mn) for the terpolymer and copolymers were 22,000 and 26,500, respectively. The molecular structure of the terpolymer is shown in Fig. 1. 2.2. Differential scanning calorimetry (DSC) The thermal stability of the precursor powder samples was evaluated by differential scanning calorimetry (Pyris 1, Perkin Elmer). Isothermal scans were conducted over a temperature range of 220–300 C. Constant rate scans were performed at heating rates of 5, 10, and 20 C/min in a nitrogen atmosphere. The kinetic parameters of reaction were evaluated by analyzing thermograms obtained at various heating rates (5– 20 C/min) according to KissingerÕs method [16,30]. The shape factor of the exotherm peak gives the kinetic order of the reaction [30]. The temperature corresponding to maximum reaction rate (Tmax)for different heating rates (/) were noted. The slopes of ln½/=T 2max vs 1/Tmax plots were used to calculate activation energy for the cyclization reaction. 2.3. Processing of precursor Powder precursors were compacted at 25 C to obtain 1 mm thick disks using a pelletizing die and a Carver model 2926 laboratory press. Pelletized precursors were compressed at 20 MPa and 220 C to obtain 3 cm · 6 cm films with a thickness of 50–60 lm for spectroscopic studies. The precursors were UV irradiated using low power (100 W) and high power (4.5 kW) mercury arc lamps for 3 h and 50 s durations, respectively, at 100 C. After UV stabilization, the terpolymer
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
A Nicolet Avatar 360 FTIR spectrometer, in the transition mode, was used to collect infrared spectra of the terpolymer and copolymer precursors. Attenuated total reflectance (ATR) IR spectra of the precursors were measured using a Nicolet Magna 550 FTIR spectrometer with a high endurance diamond ATR attachment to determine any changes in functionality of the film surfaces caused by UV irradiation. The UV–visible absorption spectra were recorded on a Shimadzu UV3100 spectrophotometer at 25 C. 2.5. Gel fraction Gel fractions of the melt processed and the UV crosslinked polymers were measured after equilibrium swelling (48 h) in dimethyl sulphoxide (DMSO) solvent (Aldrich Chemical Company, Inc. Milwaukee, WI). The swollen gel was dried in vacuum at 50 C for 5 days and weighed to obtain the mass fraction relative to the original mass.
3. Results and discussion
Heat flow/unit mass (w/g) [endoup]
2.4. Spectroscopy
Terpolymer 240°C Copolymer 240°C Terpolymer 220°C Copolymer 220°C
0.00
-0.03
-0.06
-0.09
200
400
600
800
Time (min)
(a)
0.0
Heat flow/unit mass (w/g) [endoup]
precursor film could be oxidatively stabilized by heating at a rate of 2.5 C/min to a temperature of 320 C and then holding at this temperature for 30 min in a forced-circulation oven (ATS 3610). By contrast, the maximum heating rate that could be used to oxidatively stabilize the copolymer precursors was 0.2 C/min. Thus, the photo-crosslinked terpolymer enables a significantly higher stabilization rate than does the copolymer.
1067
Terpolymer Copolymer Terpolymer Copolymer Terpolymer Copolymer
-0.5
260°C 260°C 280°C 280°C 300°C 300°C
-1.0
-1.5 25
(b)
50
75
100
125
Time (min)
Fig. 2. Isothermal DSC thermograms of PAN-based terpolymer and copolymer at different temperatures: (a) 220 and 240 C; (b) 260, 280 and 300 C.
3.1. Thermal stability of precursors Isothermal DSC scans, shown in Fig. 2(a), indicate that at 220 C both the copolymer and the terpolymer are thermally stable for about four hours. However, after this period of stability they show a broad exotherm extending up to 7 h. The order of the overall reaction, as measured by KissingerÕs method, for both the compositions was close to unity (1.16 ± 0.13 for terpolymer and 0.94 ± 0.13 for copolymer). The activation energies (Ea) for the terpolymer and copolymer compositions were also similar (137 and 131 kJ/mol, respectively), and consistent with earlier reports on PAN copolymers used as carbon fiber precursors [31]. These results establish that, in spite of the incorporation of ABP, the terpolymer and the copolymer exhibited similar cyclization kinetics and melt stability. At 220 C, exotherm maxima were observed after about 9 h (540 min). The times corresponding to these exotherm maxima and the heats of reaction (DH) at
Table 1 Isothermal DSC data of precursor powders showing exothermic DH and time corresponding to exotherm maxima Temperature (C)
Terpolymer (DH, J/g; Tmax, min)
Copolymer (DH, J/g; Tmax, min)
220 240 260 280 300
255 ± 10; 265 ± 20; 360 ± 20; 390 ± 10; 405 ± 15;
255 ± 10; 290 ± 20; 400 ± 15; 425 ± 15; 440 ± 10;
545 180 67 30 15
520 170 64 27 14
different temperature scans for both the precursors are summarized in Table 1. At 220 and 240 C, the isothermal DHs for the two compositions were not significantly different (95% confidence level). However, at higher temperatures (260–300 C, Fig. 2(b)) the heat of reaction for the terpolymer was less than that of the copolymer. Although the non-AN comonomer content was 15 mol% in both the terpolymer and the copolymer,
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
the DH/mass of terpolymer was expected to be about 3% less than that of the copolymer because of the higher molar mass of ABP compared to MA. Surprisingly, measurements showed that the DH/mass for the terpolymer was about 10% less than that for the copolymer. Apparently the random distribution of monomer segments disrupts the sequence of PAN more in the case of terpolymer than the copolymer. As a result, fewer adjacent AN repeat units are available to cyclize. For low-temperature scans, the DH values are low due to incomplete cyclization and pyrolytic reaction of PAN that finally leads to ladder structure, consistent with literature studies [32,33].
After ultraviolet (UV) irradiation, the terpolymer precursor turned dark brown, whereas the copolymer developed only a pale brownish tint. Irradiated precursors were immersed in DMSO solvent and the gel contents were measured after equilibrium swelling (48 h). Fig. 3 shows the gel content results for low- and highirradiance (0.03 and 1.0 W/cm2, respectively) exposures. The non-irradiated samples dissolved completely. By comparison, the copolymer sample remained almost completely soluble, even after 180 min of low-power UV irradiation. High-power UV irradiation of copolymer for 1 min yielded only 5% gel content. However, UV irradiated terpolymer produced 30% and 65% gel content after low- and high-power irradiation, respectively. Thus, while low-power radiation required about 60 min for an equilibrium level of crosslinking to occur, high-power radiation achieved a significantly higher level of crosslinking in a considerably shorter duration of 1 min. This establishes the role of ABP moiety in accelerating the crosslinking of the precursor and providing an adequate degree of crosslinking (over
1.0
Terpolymer high power UV Terpolymer low power UV Copolymer high power UV
Gel fraction
Copolymer low power UV 0.6
0.4
0.2
0.0 0.0
0.5
1.0
50
1 min high power (Terpolymer) 180 min low power (Terpolymer) 180 min low power (Copolymer) 180 min 110°C, no UV (Terpolymer) No UV (Terpolymer) No UV (Copolymer)
2.0
1.5
1.0
0.5
0.0 250
300
350
400
450
500
550
600
650
700
Wavelength (nm)
3.2. Influence of UV irradiation
0.8
2.5
Absorbance
1068
100
150
200
Time (min) Fig. 3. Gel fraction of PAN-based terpolymer and copolymer samples after different durations of high- and low-power UV irradiation.
Fig. 4. UV–visible spectroscopy of PAN-based terpolymer and copolymer samples after UV irradiation at different conditions.
60% gel), which prevented the melting of samples during subsequent oxidative stabilization. Next, the UV sensitivity of the precursors was systematically investigated by UV–visible spectroscopic measurements. Fig. 4 displays the UV–visible spectra for terpolymer relative to that of the copolymer at different conditions of UV irradiation. Before irradiation, it was observed that the terpolymer displayed enhanced absorbance at 330 nm relative to that of the copolymer. This difference in the spectra is due to the UV sensitivity of the benzophenone moiety in the terpolymer, and is consistent with literature studies on polymerized benzophenone derivatives [34]. The control terpolymer sample, which was not UV irradiated but underwent only the heating cycle (110 C for 180 min), remained almost colorless and did not show any further change in the UV–visible spectrum. Thus, thermal exposure at low temperature (110 C) does not significantly affect the molecular structure of the terpolymer. After low-power UV irradiation for 180 min, the absorbance of the copolymer appears at wavelengths lower than 300 nm, having a low-level tail extended in the violet region (425 nm) (Fig. 4). The low level of absorbance in the violet region is consistent with the development of only a brownish tint (rather than a significant change of color) for the copolymer. In contrast, the color generation after UV exposure is dramatic for the terpolymer. The terpolymer turned dark brown indicating a strong UV and violet light absorbance. In fact, when the terpolymer was irradiated for only 1 min in the high-power UV source (in contrast to 180 min in low-power), an abrupt violet–blue light absorbance occurred at 450 nm. The intense brown color for the terpolymer likely results from the UV-assisted cyclization reactions that produce chromophores. The UV assisted reactions were accelerated during high power UV irradiation (for short duration 1 min) and the extent of
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
reaction is more than that observed after low-power UV irradiation for 180 min. Takata and Hiroi [35] showed that, for model di-, tri-, and tetrameric acrylonitrile compounds (increasing number of ACN groups), multiple UV absorbance maxima, kmax, appeared after intra-molecular cyclization reactions. They also found that these kmax were higher for the larger ring structures. The kmax at 243, 292 and 336 nm were characteristic of cyclic structures consisting 2, 3, and 4 –CN groups, respectively. Thus, a distribution of kmax would be expected for various ring structures. The spectra displayed in Fig. 4 suggest that the
1.8
Absorbance
1.5
Absorbance
2.0
0h 0.5h 1h 3h 7h
1.5 1.0
0s 2.5 s 12.5 s 50 s
0.5 1700
1650
1600
Wavenumber (cm-1) 1657
1.2
1599
1687 1624
0.9 1579
0.6
1700
1680
1660
1640
1620
1600
1580
Wavenumber (cm-1 ) Fig. 5. Transmittance FTIR spectra of PAN-based terpolymer samples after high- and low-power (inset) UV irradiation for different durations.
C
1069
terpolymer readily forms a number of larger ring structures compared to the copolymer. Similar electronic absorption bands were reported for c irradiated PAN [24]. Transmittance infrared spectra of terpolymer precursor samples irradiated in high power UV beam for short periods are shown in Fig. 5. Peaks caused by CH2 and CH stretching (2900–2950 cm1), CN stretching (2242 cm1, due to PAN segments) and C@O stretching (1730 cm1, due to MA segments) were intense and appeared truncated due to saturation of the signal (not shown in Fig. 5). Non-irradiated samples showed a strong peak at 1657 cm1 due to >C@O groups in conjunction with phenyl rings of the ABP termonomer. The intensity of the peak decreased after UV irradiation for 2.5 s due to its conversion to an excited triplet state (through an intermediate singlet state), which creates free radicals. As proposed in the model reaction scheme of Fig. 6, these radicals may convert to secondary alcohols by abstracting hydrogen radicals from the adjacent polymer chain, thereby forming a crosslinked network. Hydrogen abstraction by the triplet state of benzophenone moiety is reported in the literature [28,36,37]. As displayed in Fig. 5 (inset), similar changes in IR spectra of the terpolymer were also observed during prolonged low-power UV irradiation. Gradual decrease in 1657 cm1 peak intensity with increasing UV exposure time (up to 1 h) can be attributed to the slow hydrogen radical abstraction reaction during low-power irradiation. After the disappearance of 1657 cm1 peak from ABP, a broad shoulder appeared at the same position of the spectra (Fig. 5) on further UV irradiation. The
C
O
O
O
O
+ C
H
R
H
R
O *
∗
+
R
CH
OH
R
H
R ϖ
R
H
Excited state ABP moiety after UV irradiation
Polymer chain
Polymer chain with free radicals
Crosslinking Polymer chains with free radicals
Crosslinked polymer network
Fig. 6. Model of free radical induced crosslinking reaction in PAN-based terpolymer.
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
shoulder was due to the formation of conjugated >C@O groups that are generated by the UV induced degradation and recombination in MA segments, consistent with that reported in the literature [38]. Other >C@O functionalities are also generated after UV irradiation as evidenced by formation of 1780 (not shown in Fig. 5) and 1689 cm1 shoulders. A small shoulder at 1624 cm1, as observed in the spectra of non-irradiated sample, may be attributed to the presence of conjugated >C@C< (with acrylate >C@O) in the precursors. These >C@C< groups may be generated during free radical polymerization via disproportionation reaction. After UV exposure, this shoulder intensity decreased initially, due to the reaction of >C@C< with the free radicals. In the case of the terpolymer, the peak at 1599 cm1 is due to a combination of C–C stretching (mCC) and C– C–C in-plane deformation (dCCC) of ABP [39]. With initial UV irradiation, as the planer carbon of benzophenone >C@O group becomes tetrahedral (due to sp3 hybridization) the overall resonance stabilization is lost and dCCC becomes restricted. As a result, the 1599 cm1 peak decreases. However, during the course of UV irradiation, different types of conjugated >C@C< groups are formed. This would account for the broad peak that appears at 1599 cm1 and increases with time. Diamond ATR FTIR spectra of high power UV irradiated and the non-irradiated control samples are shown in Fig. 7(a). In the ATR mode, because of a limited depth of penetration of about 2 lm, changes in the concentration of some abundant functional groups become apparent in this mode. It was observed that the ACN (2242 cm1) stretching absorbance is lowered after high power UV irradiation for 25 s while a broad overlapped peak appears in the frequency range 1560– 1690 cm1. This suggests that the UV assisted reaction causes a decrease in CN groups and development of unsaturations. Fig. 7(b) displays variation of peak intensities caused by CH2 asymmetric stretching (2955 cm1), as well as ACN (2242 cm1), >C@C< (1624 cm1) and >C@N– (1579 cm1) vibrations, respectively, normalized with respect to C–H stretching (2926 cm1). During the course of UV irradiation, the normalized intensity of CH2 asymmetric stretching (2955 cm1) remains almost unaltered. During crosslinking reaction, free radical generation is expected to occur at C–H sites because tertiary radicals are the most stable. Evidently, the change in C–H concentration is small, compared to the total concentration of C–H bonds. Utilization of ACN group with gradual development at 1624 cm1 and 1579 cm1 may be attributed to the formation of AC@C–C@N structure through cyclization of PAN [32]. A schematic of the plausible reactions is shown in Fig. 8. UV radiation creates free radicals in the precursor, which assist in the partial cyclization of PAN.
Absorbance 0.070 terpol 0 pass 0.060
Terpolymer-no UV irradiation
0.050 0.040
2242
0.030
1657 1599
0.020 0.010 0.000 0.070 terpol 20 passes 0.060 0.050
Terpolymer-50 s high power UVirradiation
0.040 0.030 0.020 0.010 0.000
2000
2500
3000
Wavenumbers (cm-1)
(a)
2242 cm-1 2955 cm-1 1624 cm-1 1579 cm-1
1.2
Normalized absorbance.
1070
0.9
0.6
0.3
0 0
10
(b)
20
30
40
50
60
UV exposure time (s)
Fig. 7. (a) ATR-IR spectra of PAN-based terpolymer after highpower UV irradiation. (b) ATR peak intensity (normalized with respect to C–H stretching at 2926 cm1) as a function of high power UV irradiation time for CH2 asymmetric stretching (2955 cm1), C„N stretching (2242 cm1), C@C stretching (1624 cm1), and C@N stretching (1579 cm1) (solid curves are trendlines).
Thermal treatment converts cyclized PAN to a ladder structure. Aggour and Aziz also proposed a similar free radical induced cyclization of PAN [20]. They also proposed that UV and ion beam irradiation of PAN would generate free radicals. In the present study, involving a systematically-designed terpolymer, we find that the ABP moiety accelerates the generation of free radicals, which is consistent with free radical induced crosslinking of precursors discussed earlier (Fig. 3).
4. Conclusions The acrylonitrile (AN)–methyl acrylate (MA)– acryloyl benzophenone (ABP) terpolymer precursor (85:14:1 mole ratio) was found to be thermally stable and melt processible at 220 C. An accelerated UV stabilization route was developed for the precursor, where ABP acts as the UV sensitive component. Free radicals generated during UV irradiation of the terpolymer enhance crosslinking and cyclization of PAN, as evidenced
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072 P∗
P∗
∗
1071
R
H
R + H
P
∗
Polymer chains or their segments containing free radical
P∗
+
CH
CH
CH
CH
C
C
C
C
N
+ N
N
H∗
N
Free radical induced cyclization
CH
CH
CH
C
C
C
N
P
N
CH N
C
NH
C
C
C
C
C
C
C
C
N H
P
N H
N H
Dehydrogenation NH 2
∆
P
C
C
C
C
C
C
N
N
C N
C
NH
Ladder polymer
C CH P
N H
C
C
CH
CH N H
C CH N H
NH 2
Fig. 8. Schematic of model reactions of free radical induced cyclization of PAN.
by UV-visible and IR spectroscopy. In contrast, the copolymer precursor could not be stabilized to any significant extent, because it lacks UV sensitivity. Stabilization of terpolymer was significantly accelerated by exposure to a high energy density UV beam for a very short period (1 min). The UV assisted crosslinking and cyclization inhibits melting of melt-processed terpolymer during subsequent thermo-oxidative stabilization.
Acknowledgments The authors gratefully acknowledge Mr. Dave Warren, DOE program monitor, and the financial support from the Department of Energy through contract no. 4500011036. We also thank Profs. Don Baird and Jim McGrath, Virginia Tech for providing the precursor polymers and Kim Ivey, Clemson University for assistance with diamond ATR spectroscopy.
References [1] Gupta A, Harrison IR. New aspects in the oxidative stabilization of PAN-based carbon fibers. Carbon 1996;34(11):1427–45. [2] Grassie N, Hay JN. Thermal coloration and insolubilization in polyacrylonitrile. J Polym Sci 1962;56:189–202. [3] Bashir Z. A critical review of the stabilization of polyacrylonitrile. Carbon 1991;29(8):1081–90. [4] Edie DD. The effect of processing on the structure and properties of carbon fibers. Carbon 1998;36(4):345–62. [5] Odawara K, Takeuchi S. Process for the melt shaping of acrylonitrile polymers. US Patent 4,108,818, 1978.
[6] Sen K, Bahrami SH, Bajaj P. High-performance acrylic fibers. J Macromol Sci, Rev Macromol Chem Phys 1996;C36(1):1–76. [7] Grove D, Desai P, Abhiraman AS. Exploratory experiments in the conversion of plasticized melt spun PAN-based precursors to carbon fibers. Carbon 1988;26(3):403–11. [8] Daumit GP, Ko YS, Slater CR, Venner JG, Young CC, Zwick MM. Formation of melt-spun acrylic fibers suitable for thermal conversion to high strength carbon fibers. US Patent 4,933,128, 1990. [9] Bashir Z, Atureliya SK, Church SP. Production of oriented polyacrylonitrile films by flow-induced chain extension and crystallization from solution. J Mater Sci 1993;28(10):2721–32. [10] Fishman N. Preparation of acrylonitrile polymer compositions and the molded articles. US Patent 5,434,205, 1995. [11] Frushour BG. Acrylic polymer characterization in the solid state and in solution. In: Masson JC, editor. Acrylic fiber technology and applications. New York: Marcel Dekker, Inc; 1995. p. 218. [12] Gupta AK, Paliwal DK, Bajaj P. Acrylic precursors for carbon fibers. J Macromol Sci, Rev Macromol Chem Phys 1991;C31: 1–89. [13] Smierciak RC, Wardlow Jr E, Ball LE. Process for making a melt processable high nitrile multipolymer of acrylonitrile and olefinically unsaturated monomers. US Patent 5,618,901, 1997. [14] Davidson JA, Jung HT, Hudson SD, Percec S. Investigation of molecular orientation in melt-spun high acrylonitrile fibers. Polymer 2000;41:3357–64. [15] Bhanu VA, Rangarajan P, Wiles K, Bortner M, Sankarpandian M, Godshall D, et al. Synthesis and characterization of acrylonitrile methyl acrylate statistical copolymers as melt processable carbon fiber precursors. Polymer 2002;43:4841–50. [16] Paiva MC, Kotasthane P, Edie DD, Ogale AA. UV stabilization route for melt-processible PAN-based carbon fibers. Carbon 2003;41:1399–409. [17] Ranby B, Rabek JF. Photodegradation photo-oxidation and photostabilization of polymers: principles and applications. London: Wiley; 1975. p. 190–2. [18] Bullock RE. Radiation-enhanced oxidation of PAN-based carbon fibers. Fiber Sci Technol 1974;7(2):157–60.
1072
A.K. Naskar et al. / Carbon 43 (2005) 1065–1072
[19] Dietrich J, Hirt P, Herlinger H. Electron-beam-induced cyclisation to obtain C-fiber precursors from polyacrylonitrile homopolymers. Eur Polym J 1996;32(5):617–23. [20] Aggour YA, Aziz MS. Degradation of polyacrylonitrile by low energy ion beam and UV radiation. Polym Test 2000;19:261–7. [21] Murthy MR, Radhakrishna S. Radiation damage in polyacrylonitrile. Pramana 1983;20(1):85–90. [22] Cardoso dos Santos LG, Kawano Y. Degradation of polyacrylonitrile by X-ray radiation. Polym Degrad Stab 1994;44(1):27–32. [23] Badawy SM, Dessouki AM. Cross-linked polyacrylonitrile prepared by radiation-induced polymerization technique. J Phys Chem B 2003;107(41):11273–9. [24] Zhao W, Yamamoto Y, Tagawa S. Regulation of the thermal reactions of polyacrylonitrile by c-irradiation. JAERI Conference 2001 (JCBSRC Õ99, the 8th Japan–China Bilateral symposium on radiation chemistry, 1999) 2000, p. 122–5. [25] Hill DJT, Lang AP, OÕDonnell JH, Pomery PJ. The effects of cradiation on polyacrylonitrile. Polym Degrad Stab 1992;38(3):193–203. [26] Stephenson CV, Lacey Jr JC, Wilcox WS. Ultraviolet irradiation of plastics III. Decomposition products and mechanisms. J Polym Sci 1961;55:477–88. [27] Bhanu VA, Wiles KB, Bortner M, Glass TE, Godshall D, Baird DG, et al. Designing the molecular composition of the melt processable acrylonitrile copolymer carbon fiber precursors. Polym Prep 2002;43(2):1017–8. [28] Mukundan T, Bhanu VA, Wiles KB, Bortner M, Baird DG, McGrath JE. Photocrosslinkable acrylonitrile terpolymers as carbon fiber precursors. Polym Prepr 2003;44(2):651–2. [29] Naskar A, Proulx S, Walker K, Edie D, Ogale A. UV stabilization of melt-processible PAN carbon precursor fibers. Abstracts, 55th
[30] [31]
[32]
[33]
[34]
[35] [36] [37] [38]
[39]
Southeast Regional Meeting of the American Chemical Society, Atlanta, GA, United States, 16–19 November, 2003, p. 869. Kissinger HE. Reaction kinetics in differential thermal analysis. Anal Chem 1957;29(11):1702–6. Fitzer E, Mu¨ller DJ. The influence of oxygen on the chemical reactions during stabilization of PAN as carbon fiber precursor. Carbon 1975;13:63–9. Xue G, Dong J, Zhang J. SERS studies of polymers on metal surfaces. 4. Cyclization and graphitization of polyacrylonitrile on rough silver surfaces. Macromolecules 1992;25:5855–7. Leroy S, Boiziau C, Perreau J, Reynaud C, Zalczer G, Lecayon G, et al. Molecular structure of an electropolymerized polyacrylonitrile film and its pyrolyzed derivatives. J Mol Struct 1985;128(4):269–81. Oh JK, Wu J, Winnik MA, Craun GP, Rademacher J, Farwaha R. Polymerizable benzophenone derivatives for labeling vinyl acetate-butyl acrylate latex particles. J polym Sci, Part A, Polym Chem 2002;40(17):3001–11. Takata T, Hiroi I. Coloration in acrylonitrile polymers. J Polym Sci, Part A 1964;2:1567–85. Walling C, Gibian MJ. Hydrogen abstraction by the triplet state of benzophenone. J Am Chem Soc 1964;86:3902–3. Naguib YMA, Cohen SG, Steel C. Reduction of crystal violet by diphenylketyl radicals. J Am Chem Soc 1986;108:128–33. Cameron GG, Kane DR. The thermal degradation of poly (methyl acrylate) III. Mechanism of volatile formation. Die Makromolekulare Chemie 1968;113:75–84. Kolev TM, Stamboliyska BA. Vibrational spectra and structure of benzophenone and its 18O and d10 labelled derivatives: an ab initio and experimental study. Spectrochim Acta Part A 1999;56:119– 126.