polyacrylonitrile
fiber through plasma surface modification
Mehmet Gürsoy,
1Fatih Özcan,
2,3Mustafa Karaman
1,21Department of Chemical Engineering, Konya Technical University, Konya 42075, Turkey 2Advanced Technology Research & Application Center, Selcuk University, Konya 42075, Turkey 3Department of Chemistry, Faculty of Science, Selcuk University, 42075, Konya, Turkey Correspondence to: M. Karaman (E-mail: karamanm@selcuk.edu.tr)
ABSTRACT:In this study, surfaces of multiwalled carbon nanotubes (CNTs) were functionalized with poly(hexafluorobutyl acrylate) (PHFBA) thin film using a rotating-bed plasma-enhanced chemical vapor deposition (PECVD) method without imparting any defects on their surfaces. Polyacrylonitrile (PAN) electrospun polymerfiber mats and composite fiber mats with CNTs and functionalized CNTs (f-CNTs) were prepared. The wettability and chemical and morphological properties of the synthesized fiber mats were investigated, and the dispersion of CNTs and f-CNTs in the polymer matrix was compared according to the contact angle results of electrospun poly-mer mats. According to the chemical and morphological characterization results, PHFBA-coated CNTs were dispersed more uniformly in the polymer matrix than the uncoated CNTs. The f-CNTs/PAN compositefiber mat exhibits a lower surface energy than the pristine CNTs/PANfiber mat.© 2019 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2019, 136, 47768.
KEYWORDS:composites; electrospinning; nanotubes, graphene, and fullerenes Received 13 January 2019; accepted 4 March 2019
DOI: 10.1002/app.47768
INTRODUCTION
Electrospinning is considered an efficient technique for fabricating polymeric nanofibers. The resultant mechanical, electrical, and chemical properties of electrospun nanofibers can be improved by incorporation of variousfillers such as metal nanoparticles,1clay,2 graphene oxide,3or carbon nanotubes (CNTs)4into the polymer solution. Among them, CNTs have attracted wide attention for use asfillers in polymer composites because of their unique properties, including their high aspect ratio, high surface area, high elastic modulus, and high tensile strength.5–10In order to take advantage of these unique properties of CNTs for polymer composites, they need to be homogeneously distributed in polymer matrixes with-out clustering. However, CNTs tend to agglomerate in solutions because of their flexible structures and the large surface area resulting from their extremely high aspect ratio.11,12
Moreover, carbon atoms on CNT walls are chemically stable because of the aromatic nature of the bonds. As a result, CNTs can interact with the surrounding polymer matrix mainly via van der Waals interactions, which cannot provide an efficient load transfer across the CNT–polymer interface.13
Thus, various methods have been proposed to improve the disper-sibility of CNTs in a polymer matrix. These methods can be
divided into two main categories of liquid-based and gas-based methods. The main purposes of both methods are to modify or functionalize CNTs to change their surface properties in order to improve their dispersion in a polymer matrix. Surface func-tionalization of CNTs can provide covalent bonds between CNTs and the polymer matrix, instead of much weaker van der Waals physical bonds,14 which can help prevent their agglomeration. However, it is difficult to functionalize small particles (especially below 100 μm) like CNTs without agglomeration while using liquid-based methods.15,16 The use of solvents in liquid-based methods17,18can lead to physical damage to the surfaces of carbon nanotubes. Some corrosive chemicals can change their hybridiza-tion state from sp2to sp3, which would adversely affect the
electri-cal, thermal, and mechanical properties of thefinal products.19 Vapor-based coating methods such as chemical vapor deposition (CVD), on the other hand, have many advantages over wet coat-ing techniques. Due to the solvent-free nature of vapor-based coating techniques, the CVD method prevents problems associ-ated with solvents such as the agglomeration of CNTs. So far, dif-ferent CVD strategies, such as initiated CVD (iCVD)20 and plasma-enhanced CVD (PECVD),21have been employed to mod-ify CNT surfaces. Previously, it was reported that PECVD-modified CNTs show better interface compatibility with an epoxy
Additional Supporting Information may be found in the online version of this article. © 2019 Wiley Periodicals, Inc.
matrix, compared to those with chemical functionalization.22 In another study, CNTs were modified by PECVD to improve their dispersion capability in water.23
In this study, CNTs were encapsulated with poly(hexafluorobutyl acrylate) (PHFBA) thin films by PECVD. Thus, the aim is to overcome the various limitations associated with their dispersion in a polymer matrix. Moreover, a rotating-bed type reactor was chosen to increase the uniformity of the thin film coating on CNTs. Polyacrylonitrile (PAN) electrospun polymer fiber mats and their composite fiber mats with CNTs and functionalized CNTs (f-CNTs) were prepared. The wettability and chemical and morphological properties of all fiber mats were investigated, and the dispersion of CNTs and f-CNTs in polymer matrixes were compared according to the contact angle results of electrospun polymer mats.
EXPERIMENTAL
CNT Synthesis
Multiwalled CNTs were synthesized by the CVD method using an Fe/Al catalyst. Iron(III) nitrate nonahydrate [Fe(NO3)39H2O,
≥98%, Sigma-Aldrich, St. Louis, MO] and aluminum nitrate non-ahydrate (≥98%, Sigma-Aldrich) were used as the sources of Fe and Al, respectively. First, 90 g urea and 3 g Al(NO3)39H2O
were dissolved in 450 mL water. Then the solution was stirred at 80C on a magnetic stirrer. Meanwhile, 3 g Fe(NO3)39H2O and
3 g Al(NO3)39H2O were dissolved in 80 mL water and added
into the other solution (the solution of urea, Al(NO3)39H2O,
and water). The obtained solution mixture was allowed to stand at 80C for one day and then filtered using filter paper. Subse-quently, the catalysts were placed in a furnace and allowed to stand at 500C for 10 h. Then, the as-synthesized Fe/Al catalyst was loaded into the CVD reactor, and CNTs were synthesized at 750C for 20 min at atmospheric pressure using an acety-lene/hydrogen (2:1) gas mixture. The obtained CNTs were treated with 5% HF solution for 5 min at room temperature and subse-quently with 4 M HNO3 for 1 h at 50C in order to remove
amorphous carbon and residual metal catalysts. Plasma Polymerization
Polymeric thin films were deposited on a CNT surface using a rotating-bed PECVD system. A more detailed description of the PECVD setup has been presented elsewhere.24,25 A radio fre-quency (RF) plasma discharge was obtained inside a cylindrical Pyrex tube using a 13.56 MHz RF generator at 20 W pulsed plasma power (duty cycle 30%). The monomer 2,2,3,4,4,4-hexafluorobutyl acrylate (HFBA, 95%, Sigma-Aldrich) was used as received without further purification or modification. HFBA was vaporized in a glass jar maintained at room temperature. Plasma polymerization was carried out at a monomerflow rate of 1.2 sccm and a reactor pressure of 0.1 Torr for 30 s. During the deposition, the reactor was rotated at 30 rpm by a stepper motor to improve the uniformity of the PHFBA coating on the CNT surfaces.
Electrospinning
All electrospun polymer fiber mats were prepared on aluminum foil substrates using an electrospinning unit (Basic System Electrospinning Equipment, NANOspinner, Inovenso, Boston,
MA). Polyacrylonitrile (PAN, MW = 150,000, Sigma-Aldrich) at 10 wt % was dissolved in an N,N-dimethylformamide (DMF, 99.8%, Sigma-Aldrich) solution to prepare pure polymer fibers. The mixtures were then stirred at 55C for 3 h using a magnetic stirrer with its speed set to 100 rpm. At this stage, CNT and f-CNTs were added to the mixtures (5 wt % of the polymer weight) to prepare polymer composites with unmodified and modified CNTs. After the addition of CNTs, stirring was continued for another 30 min at the same conditions. During the electrospinning process, the distance between the collector and the tip was held constant at 18 cm. An electric potential difference of 16 kV was applied to the copper tip, and aflow rate of 3 mL/h was used. Characterization
Static, advancing, and receding water contact angle measure-ments were performed atfive points (the center and four edges) on 5× 5 cm2fiber mats using a contact angle goniometer (Model OCA 50, DataPhysics Instruments GmbH, Filderstadt, Germany). Moreover, the contact angles of unmodified and modified CNTs were measured by the Washburn capillary rise method,26and n-heptane was used for the Washburn method as the test liquid. Surface energies offiber mats were determined by the Owens– Wendt–Rabel–Kaelble (OWRK) method.27 For this purpose, the contact angles of three different test liquids [water, diiodomethane (≥99.0%, Aldrich), and ethylene glycol (99.8%, Sigma-Aldrich)] on fiber mats were measured. The surface tension parameters of the test liquids used in the surface energy calculation are given in Table I. The details of the surface energy calculations are given in the Supporting Information.
X-ray photoelectron spectroscopy (XPS) analysis and the Brunauer– Emmett–Teller (BET) method were employed to analyze the chemi-cal and morphologichemi-cal structures of the modified and unmodified CNT surfaces. Transmission electron microscopy (TEM) was used to analyze the morphological properties of both CNTs and electrospun fibers.
XPS was performed using a SPECS spectrometer (Berlin, Germany) equipped with a monochromatic Al source. The TEM images were acquired using a JEOL JEM-2100 (Tokyo, Japan) operating at an acceleration voltage of 200 kV. The BET data were measured on an Autosorb-IQ2 instrument (Quantachrome, Boynton Beach, FL). Scanning electron microscopy (SEM, Zeiss LS-10, Oberkochen, Ger-many) coupled with energy dispersive X-ray spectroscopy (Bruker 123 eV EDX sensor, Karlsruhe, Germany) was employed to analyze the chemical properties of allfiber mats.
Table I. Surface Tension Parameters of Test Liquids Used in Surface Energy Calculations Test liquid Surface tension (total) Surface tension (dispersive) Surface tension (polar) Water 72.30 18.70 53.60 Diiodomethane 50.80 49.50 1.30 Ethylene glycol 47.70 26.40 21.30
RESULTS AND DISCUSSION
CNT Results
TEM images of pristine CNTs and PHFBA-coated CNTs are shown in Figure 1(a,b). According to Figure 1(a), CNTs are multiwalled with an average external diameter of about 10 nm. In Figure 1(b), a PHFBA polymer thinfilm of approximately 10 nm thickness was clearly observed to be deposited on the outer walls of the CNTs. The surface areas of the pristine CNTs and polymer-coated CNTs were calculated by the BET method based on nitrogen adsorption–
desorption measurements. Although their surface areas are found to be quite similar, the surface area of PHFBA-coated CNTs (202.2 m2/g) is a little less than that of pristine ones (212.9 m2/g). The small
decrease in the surface area can be caused by the thickness of the CNTs being increased by the polymer coating.
XPS analysis was performed to reveal the surface chemical structure of CNTs before and after plasma modification. XPS survey scans of pristine CNTs and PHFBA-coated CNTs are given in Figure S1 (in the Supporting Information) and Figure 2(a), respectively. The
Figure 1.TEM images of (a) pristine CNT and (b) PHFBA-coated CNT.
Figure 2.(a) XPS wide-scan survey spectra of PHFBA-coated CNT; (b) high-resolution scan of PHFBA-coated CNT; (c) high-resolution scan of O1s peak with two resolved peaks; (d) high-resolution scan of C1s peak with seven resolved peaks.
calculation of atomic percentages from survey scans shows that pris-tine CNTs contain 99.2 at. % C and 0.8 at. % O. The atomic percent-ages of C, O, and F were found to be 48.1, 12.5, and 39.4 at. % on the PHFBA-coated CNT surface, respectively. These values are very similar to the theoretical values calculated from the chemical for-mula of the monomer (46.7 at. % C, 13.3 at. % O, 40.0 at. % F). Moreover, a more detailed chemical investigation of PHFBA-coated CNTs [Figure 2(b)] was performed using the high-resolution mode of XPS. The high-resolution C1s spectrum [Figure 2(c)] can be curve-fitted with seven peak components at binding energies of 293.3, 291.1, 289.2, 286.8, 286.4, 285.7, and 285.0, which can be attributed to ─C*F3, ─C*F2─, ─C*─O, ─CH2─CF2─C*HF,
─O─C*H2─, ─C*H─CO─, and ─C─C*H2─C─, respectively.28,29
The O1s spectrum [Figure 2(d)] can be curve-fitted with two major peak components at binding energies of 532.8 and 534.1, attributed to the groups ─C─O* and ─O*─CH2, respectively.30 Observed
binding energies in the C1s and O1s spectra are given with attrib-uted groups and their theoretical values in Table II. The XPS results can be considered as evidence for the presence of PHFBA coating on CNT surfaces.
The effect of the PHBA coating on the wettability of CNTs was determined using the Washburn capillary method. For this pur-pose, the following equation is used25:
h2=ReffγLcosθ
2η t ð1Þ
in which h is the height of the liquid; t is the time to reach that height; Reffis the effective pore radius;θ is the contact angle of the
liquid;η and γLare the dynamic viscosity and surface tension of the
liquid, respectively; h2/t is calculated from the slope of the linear plot of distance squared versus time. Note thatη and γLare known liquid
properties. In the equation, there are still two unknowns: Reffandθ.
Therefore, the Reffvalue must be determined in order to calculateθ
for water. For this purpose,first n-heptane was used as the test liq-uid because of its low surface tension, and the contact angle of n-heptane was assumed to be zero. Then the experiment was carried out using water (Figure 3). Using eq. (1), the contact angle of pris-tine CNTs was calculated as 74.8from Figure 3. When the experi-ment was carried out for PHFBA-coated CNTs, water did not penetrate through the capillary column. According to this observa-tion, modified CNTs can be considered highly hydrophobic. Fiber Results
Photographs of fiber mats with their contact angles, total surface energies, and EDX results of the pure PANfiber mat, CNT/PAN compositefiber mat, and f-CNT/PAN composite mat are given in Figure 4. The contact angle values were measured fromfive differ-ent points (the cdiffer-enter and four edges) on 5× 5 cm2fiber mats.
Based on the water contact angle results, both CNT/PAN composite and f-CNT/PAN compositefiber mats exhibited a higher contact angle and lower contact angle hysteresis than pure PANfiber mat. The incorporation of CNTs into polymerfibers usually decreases the size of thefiber diameter.31,32In this study, the same trend was also observed (see Figure 5). It can be expected that a thinnerfiber can minimize the contact area between its surface and a water drop, resulting in an increase in contact angle. Moreover, the increase in contact angle with the presence of CNTs could also be attributed to a heterogeneous surface effect.33,34Increasing contact angles with the incorporation of CNTs into polymeric composites have also been reported in previous studies.35–37
Among all thefiber mats, the f-CNT/PAN composite fibers showed the highest contact angle values. As expected, in addition to the carbon and oxygen peaks observed in both pure PANfiber and CNT/PAN compositefiber mats, fluorine was also observed in the f-CNTs/PAN compositefibers based on the results of SEM/EDX analysis (SEM/EDX spectra of allfiber mats are given in the Supporting Information). The presence offluorine elements could be evidence of CF3groups with
low surface energy in PHFBA. As compared to the contact angle results, not surprisingly the total surface energy calculations showed the exact opposite trend. The values for polymer fiber mats were ranked from higher to lower as follows: pure polymer fibers, CNT/polymer compositefibers, and f-CNTs/polymer composite fibers. As can be seen from the photographs of allfiber mats, differences in their overall physical appearances can be detected with even the naked eye. While big black spots resulting from the aggrega-tion of carbon nanotubes can be seen on the polymer composite
Table II.XPS Elemental Analyses of Pristine CNT and PHFBA-Coated CNT
Origin Binding energy (eV) Theoretical Binding energy (eV) Experimental C1s 1 ─C*F3 293.3 293.3 2 ─C*F2─ 291.2 291.1 3 ─C*─O 289.2 289.2 4 ─CH2─CF2─C*HF 286.8 286.8 5 ─O─C*H2─ 286.7 286.4 6 ─C*H─CO─ 285.7 285.7 7 ─C─C*H2─C─ 285.0 285.0 O1s 1 ─C O* 532.2 532.8 2 ─O*─CH2 533.7 534.1
Figure 3. Distance squared versus time for pristine and fluoropolymer-coated CNTs.
fiber mat with CNTs, these black spots are not observed on the polymer compositefiber mat with f-CNTs. This observation indi-cates that f-CNTs were more homogeneously distributed in the polymer matrix compared to pristine CNTs. The reason for this observation could be attributed to the dispersion difference between pristine CNTs and f-CNTs in the polymer matrix. Pristine CNTs tend to aggregate and cluster. This aggregation can originate from van der Waals interactions between neighbor-ing CNTs. Accordneighbor-ing to the TEM images (Figure 5), aggregated CNTs are observed inside the CNT/PAN compositefiber. On the other hand, f-CNTs are dispersed in the f-CNT/PAN composite fiber without agglomeration. A fluoropolymer coating could help reduce van der Waals interactions among CNTs.38Moreover, the steric repulsion of fluorine can also play a role in reducing the tendency of CNTs to agglomerate.39On the other hand, encapsu-lation of CNTs withfluoropolymer could increase the interaction between CNTs and polymer matrixes,40 so the tendency for CNTs to reagglomerate might be decreased.
Furthermore, the contact angle results from different points on both compositefibers with CNTs and with f-CNTs were compared to each other. The dispersion differences between pristine CNTs and f-CNTs were also reflected in the contact angle values of as-obtained fibers. The measured contact angle values from different points on
f-CNT/PAN compositefibers are much closer to each other than to those of CNT/PAN compositefibers. This comparison indicates that f-CNTs were dispersed more uniformly in the polymer matrix.
CONCLUSIONS
A rotating-bed PECVD method provided conformal PHFBA coatings around CNTs having asymmetric nanostructures with
(a)
(b)
(c)
Figure 4.Photographs offiber mats with contact angles, total surface energies, and SEM/EDX results: (a) pure PAN fiber mat, (b) CNT/PAN composite fiber mat, (c) f-CNT/PAN composite mat (W: water, E: ethylene glycol, D: diiodomethane).
Figure 5.TEM images of (a) PAN electrospunfiber, (b) CNT/PAN com-positefiber, and (c) f-CNT/PAN composite fiber.
high aspect ratios without any defects on their surfaces. PHFBA-coated CNTs were dispersed more uniformly in the polymer matrix compared to uncoated CNTs. The f-CNTs/PAN composite fiber mat exhibited a lower surface energy than CNTs/PAN, due to the presence of fluorine. Moreover, the rotating-bed PECVD approach developed in this study can be applied to modify other nanoparticles or functionalize CNTs with other functional groups.
ACKNOWLEDGMENTS
This project was supported by the Scientific and Technological Research Council of Turkey (TÜB_lTAK) with grant 213M399 and by the Selcuk University Scientific Research Foundation. M. Gürsoy was supported by the TÜBITAK BIDEB National Doctoral Fellow-ship Program. The authors would also like to thank MCM ARGE Ltd. for their help in reactor design and optimization.
REFERENCES
1. Moradi, G.; Rajabi, L.; Dabirian, F.; Zinadini, S. J. Appl. Polym. Sci. 2018, 135, 45738.
2. Xue, J.; Niu, Y.; Gong, M.; Shi, R.; Chen, D.; Zhang, L.; Lvov, Y. ACS Nano. 2015, 9, 1600.
3. Zhang, S.; Li, D.; Kang, J.; Ma, G.; Liu, Y. J. Appl. Polym. Sci. 2018, 135, 46443.
4. Zhao, Y.; Wang, S.; Guo, Q.; Shen, M.; Shi, X. J. Appl. Polym. Sci. 2013, 127, 4825.
5. Prasanna, D.; Selvaraj, V. J. Colloid Interface Sci. 2016, 472, 116.
6. Staudinger, U.; Thoma, P.; Lüttich, F.; Janke, A.; Kobsch, O.; Gordan, O. D.; Pötschke, P.; Voit, B.; Zahn, D. R. T. Compos. Sci. Technol. 2017, 138, 134. 7. Liu, X.; Yang, Q. S.; Liew, K. M.; He, X. Q. Carbon. 2017,
115, 220.
8. Bhattacharya, M. Materials (Basel). 2016, 9, E262.
9. Priyanka, R.; Chandra Mohan Singh, N.; Ajay Singh, V.; Amarjeet, S.; Gayatri, C.; Anto Regis, I.; Saral, K. G. Mater. Res. Express. 2017, 4, 085905.
10. Cen-Puc, M.; Oliva-Avilés, A. I.; Avilés, F. Phys. E (Amsterdam, Netherlands). 2018, 95, 41.
11. Chen, C.; Chen, X.; Xu, L.; Yang, Z.; Li, W. Carbon. 2005, 43, 1660.
12. Hilding, J.; Grulke, E. A.; George Zhang, Z.; Lockwood, F. J. Dispersion Sci. Technol. 2003, 24, 1.
13. Ma, P.-C.; Siddiqui, N. A.; Marom, G.; Kim, J.-K. Compos-ites, Part A. 2010, 41, 1345.
14. Khosla, A.; Gray, B. L. Mater. Lett. 2009, 63, 1203.
15. Kage, H.; Abe, R.; Hattanda, R.; Zhou, T.; Ogura, H.; Matsuno, Y. Powder Technol. 2003, 130, 203.
16. Karaman, M.; Gürsoy, M.; Kus, M.; Özel, F.; Yenel, E.; Sahin, Ö. G.; Kivrak, H. D. In Surface Treatments for Biological, Chemical and Physical Applications; Gürsoy, M., Karaman, M., Eds.; John Wiley & Sons: Weinheim, 2017; pp 23.
17. Dujardin, E.; Ebbesen, T. W.; Krishnan, A.; Treacy, M. M. Adv. Mater. 1998, 10, 611.
18. Hiura, H.; Ebbesen, T. W.; Tanigaki, K. Adv. Mater. 1995, 7, 275.
19. Alam, A.; Wan, C.; McNally, T. Eur. Polym. J. 2017, 87, 422. 20. Lau, K.; Gleason, K. Adv. Mater. 2006, 18, 1972.
21. Shi, D.; Lian, J.; He, P.; Wang, L.; van Ooij, W. J.; Schulz, M.; Liu, Y.; Mast, D. B. Appl. Phys. Lett. 2002, 81, 5216.
22. Krushnamurty, K.; Srikanth, I.; Rao, G. H.; Prasad, P. S. R.; Ghosal, P.; Subrahmanyam, C. RSC Adv. 2015, 5, 62941. 23. Chen, C.; Ogino, A.; Wang, X.; Nagatsu, M. Appl. Phys.
Lett. 2010, 96, 131504.
24. Gürsoy, M.; Karaman, M. Chem. Eng. J. 2016, 284, 343. 25. Gürsoy, M.; Karaman, M. Prog. Org. Coat. 2018, 120, 190. 26. Washburn, E. W. Phys. Rev. 1921, 17, 273.
27. Owens, D. K.; Wendt, R. C. J. Appl. Polym. Sci. 1969, 13, 1741.
28. Beamson, G.; Briggs, D. J. Chem. Educ. 1993, 70, A25. 29. Gupta, M.; Gleason, K. K. Langmuir. 2006, 22, 10047. 30. Karaman, M.; Yenice, E. Chem. Vap. Deposition. 2015,
21, 188.
31. Ra, E. J.; An, K. H.; Kim, K. K.; Jeong, S. Y.; Lee, Y. H. Chem. Phys. Lett. 2005, 413, 188.
32. Kaur, N.; Kumar, V.; Dhakate, S. R. SpringerPlus. 2016, 5, 483.
33. Kim, S. H. J. Adhes. Sci. Technol. 2008, 22, 235.
34. Park, J.-M.; Wang, Z.-J.; Jang, J.-H.; Gnidakoung, J. R. N.; Lee, W.-I.; Park, J.-K.; Lawrence DeVries, K. Composites, Part A. 2009, 40, 1722.
35. Rungraeng, N.; Cho, Y.-C.; Yoon, S. H.; Jun, S. J. Food Eng. 2012, 111, 218.
36. Kaur, T.; Thirugnanam, A. RSC Adv. 2016, 6, 39982. 37. Tijing, L. D.; Park, C.-H.; Choi, W. L.; Ruelo, M. T. G.;
Amarjargal, A.; Pant, H. R.; Im, I.-T.; Kim, C. S. Compos-ites, Part B. 2013, 44, 613.
38. Yang, Y.; Dai, J.; Li, Y.; Jia, M.; Huang, X. Carbon. 2017, 115, 285.
39. Lagow, R. J.; Wei, H.-C. In Fluoropolymers 1: Synthesis; Hougham, G.; Cassidy, P. E.; Johns, K.; Davidson, T., Eds., Springer US: Boston, MA, 2002; p 209.
40. Choi, I. H.; Park, M.; Lee, S.-S.; Hong, S. C. Eur. Polym. J. 2008, 44, 3087.
