Rapid and Continuous Atmospheric Plasma Surface Modification of PAN-Based Carbon Fibers.

In this work, a continuous and rapid atmospheric plasma setup was developed for rapidly modifying the surface of PAN-based carbon fibers (CFs). The interlaminar shear strength (ILSS) of CFs increased from 64.9 to 80.0 MPa with 60 s plasma treatment. Further mechanical and surface structural characterizations revealed that the effect of plasma was different, depending on the treatment time. When the treatment time was lower than 15 s, the effect of plasma was mainly on physically etching the surface of CFs, and the ILSS of CFs increased rapidly. Further extending the plasma treatment time did not increase surface roughness but promoted the addition of oxygen-containing functional groups on the surface of CFs, corresponding to a slower growth rate of ILSS. The atmospheric plasma was generated via a dielectric barrier discharge (DBD) method, and its energy intensity was significantly lower than that of plasma generated under low pressure. Accordingly, a mechanism was proposed for the plasma treatment of CFs: atmospheric plasma was not strong enough to simultaneously etch all the carbon atoms on the surface of CFs; therefore, carbon atoms on the graphitic plane were selectively etched, followed by the attaching of oxygen-containing functional groups on the exposed carbon sites caused by etching.

Introduction

Carbon fiber reinforced plastic (CFRP) is well-known for its high specific strength, high modulus, superior thermal stability, and chemical resistance.[1−4] CFRPs have been widely adopted in aerospace, satellite, wind turbine, and sports equipment.[5,6] However, the surface of PAN-based carbon fibers (CFs) consists of graphitic crystalline structures and shows poor adhesion to resin matrix.[7,8] The outstanding mechanical properties of composites tend to depend on the strong interfacial bonding between fibers and matrix resin.[9,10] Surface modifications are an indispensable process to improve the surface polarity and binding energy of CFs.[11] Traditional treatment methods can be classified into vapor-phase chemical oxidation, liquid-phase chemical oxidation[12−14], and surface coating.[15−17] Ozone oxidation is an effective method to modify the surface of CFs, which takes advantage of the strong oxidizing property of ozone. Li and co-workers demonstrated the ozone treatment increased the amount of carboxyl groups on the carbon fiber surface.[18] Jin et al. discovered the ozone method improved the compressive strength and flexural strength of carbon/carbon composites.[19] In addition, there are surface treatment methods that combine ozone and UV radiation to generate hydroxyl radicals with stronger oxidizing ability. However, the ozone oxidation method is limited by its efficiency, and the ozone needs to be treated before it can be discharged. Among these methods, electrochemical oxidation has been the most universal surface modifications[20,21] because it has the advantages of easy operation,[22] continuous processing, and low-cost. With the assistance of direct current, electrolyte solution can achieve continuous contact treatment on CFs by means of anodic oxidation and cathodic reduction.[23] However, the electrochemical oxidation has the drawback of being time-consuming and requires complicated subsequent washing and drying,[24] and the disposal of electrolyte causes environmental pollution.[25] Plasma treatment has the advantages of being highly efficient, environmentally friendly, and simplistic in operation, which makes it a promising candidate for the surface modification of CFs.[26]

Plasma is a complex mixture of atoms, molecules, electrons, and free radicals and is a state of matter that appears electrically neutral.[27] Several pioneering works have been conducted to modify the surface of CFs using plasma.[28−32] Zhao and co-workers discussed the effect of H2/O2 plasma treatment on CFs/epoxy interfacial adhesion,[33] and the results indicated the roughness and the oxygen content of fiber surfaces increased. Kusano demonstrated the rise of C–O bonding on the surface of CFs with He/O2 plasma.[7] Park and co-workers studied the effect of He/O2 plasma treatment on the surface of CFs.[34] On the basis of these studies, plasma was verified to introduce oxygen-containing functional groups to the surfaces of CFs,[35,36] and the etching tended to concentrate on a few atomic layers of the surfaces of CFs and avoided the reduction of mechanical properties of CFs. Although these results demonstrated the promising feature of plasma modification, most studies adopted batch processing method under high vacuum, which cannot meet the requirements of industrialization. Meanwhile, previous studies mainly relied on inert gas as gas source, which was not practical considering the gas cost. Therefore, developing a continuous and cost-effective plasma modification method was desirable for the practical application of plasma in the carbon fiber industry.

Among limited reports on continuous plasma treatment of CFs, two types of plasma equipment were mainly employed. The first type drives the CFs near the planar DBD (dielectric barrier discharge) electrodes.[26] In this case, the CFs are likely to suffer damage from abrasion, and the conductive CFs would disturb the discharge zone. The other typical method of continuous plasma treatment is directly applying plasma jets on CFs. However, the spot size of the plasma jet is too concentrated and cannot achieve a uniform treatment throughout the CF tow. If an atmospheric-pressure plasma were to be applied in the areas of etching CFs, a uniform plasma over a large area would be required.

In this study, a rapid and continuous atmospheric plasma setup based on DBD method with compressed air as gas source is developed. The plasma was generated in a modified cylindrical plasma reactor where the outer electrode was divided into two half-tubes. Compressed air was purged through the gaps and redirected through a perforated plate to generate an area of uniform plasma. The effect of treatment time on the structure and properties of CFs is examined. ILSS, SEM, AFM, XPS, Raman, and XRD are used to characterize the interfacial energy, surface morphology, microtransformation, and chemical composition of surfaces of CFs, and a mechanism of plasma modification is proposed. Finally, the efficiency and cost of this plasma setup is discussed in comparison with the electrochemical oxidation methods.

Methods

Materials

PAN-based carbon fibers in this study were manufactured by a pilot-scale production line in the lab. The wet-spun precursor fibers (12K per tow) were provided by Jilin Chemical Fiber Company. The average diameter of these CFs was approximately 7.5 μm, and the tensile strength and modulus were about 3.5 and 230 GPa, respectively. The CFs were untreated and unsized.

Atmospheric Plasma Treatment

Carbon fibers were treated in a customized atmospheric plasma setup, as shown in Figure . The plasma was generated in a modified cylindrical plasma reactor, where the outer electrode was divided into two half-tubes. The actual plasma setup and the plasma reactor configuration can be seen in the Supporting Information Figure S1. Compressed air with a pressure of 0.7 MPa was purged through the gaps and redirected through a perforated plate to generate an area of uniform plasma. The flow rate was 2.5 L/min, and the plasma power was 600 W. The plasma jet passed through multiple circular nozzle holes and directly applied onto the CFs. The effective processing area was 280 mm × 20 mm. The distance between fibers and the jet hole was 3 mm. By controlling the wire speed, the processing time of fibers was set to 0 s, 7.5 s, 15 s, 30 s, and 60 s. The samples for ILSS test were sized with 1 wt % epoxy resin.[37]

Figure 1

Schematic diagram of continuous atmospheric plasma modification setup.

Characterizations

X-ray Photoelectron Spectra (XPS)

The fibers were characterized by X-ray photoelectron spectroscopy (Thermo Fisher ESCALAB 250Xi; Waltham, MA) with Al Kα X-ray to determine the surface element content and functional groups. The vacuum degree of the analysis chamber was 2 × 10–7 mbar during work, and the C1S electron binding energy was referenced at 284.8 eV. Origin 2018 was used to analyze data by performing peak fitting. The C1S spectrogram was divided into four peaks, which belonged to C–C bond (284.88 eV); C–OH/C–O–C bond (286.38 eV); C=O bond (287.88 eV); COOR (288.88 eV) respectively. The percentage content ratio (R) was obtained by dividing the intensity of each by the total intensity, for instance, AC–C represents the area of C–C bond peak. Detailed processing of XPS signals for individual samples can be found in Supporting Information Figure S2.

X-ray Diffractometer (XRD)

The XRD patterns were recorded by using an Xpert-Pro X-ray diffractometer (PANalytical, Almelo, Netherlands) with Ni-filtered radiation laser that was generated at 40 mA and 50 eV and scanning of 5–60°. The crystallite sizes (Lc) and crystal planar spacings (d002) of CF samples were determined by the following formulas:where k value is a constant of 0.89, λ represents the X-ray wavelength of 0.1542 nm, F is full width at half-maximum (fwhm) of the diffraction peak, and θ is the Bragg angle.

When optical path difference equaled an integer multiple of wavelength, the crystal planar spacings (d002) could be calculated by Bragg’s law, where n is the reflection order of 1, θ refers to the angle between incident X-ray and the corresponding crystal plane, and λ means the X-ray wavelength of 0.1542 nm.

Raman Spectra

Experimental Raman spectra were collected by a Renishaw InVia Raman microscope, which was equipped with a 514 nm Ar-ion laser, and the laser power was set to 10%. The laser used a grating of 2400 lines/mm. The Raman spectrogram was fitted into five peaks,[38,39] which belonged to D1 (amorphous carbon), G (graphite carbon), D2 (defect-related and arises from a complex interaction with the electron band structure near the K point), and D3 and D4 (amorphous and impurity contributions). The D1/G ratio was calculated to characterize the degree of surface nongraphitization (edge carbon atoms) content. The detailed processing of Raman spectra for individual samples can be found in Supporting Information Figure S3.

Interlaminar Shear Strength (ILSS) Test

The ILSS of CFs were tested according to ASTM D2344 method, and the detailed sample preparation procedure is shown in the Supporting Information. In general, the ILSS of composites was measured on a universal testing machine (Instron-5578; Instron, Norwood, MA) using a three-point short-beam bending test method with a span-to-thickness ratio of 5. The loading speed was set to 2 mm/min, and the results were calculated according to the following eq .where P is the maximum compression load at fracture in Newtons, b is the breadth of the specimen in mm, and d is the thickness of the specimen in mm. Each reported ILSS value was the average of more than six successful measurements.

Scanning Electron Microscope (SEM)

SEM images were obtained using Zeiss Merlin-SEM instruments at an accelerating voltage of 5.00 kV with 10.003× magnification.

Atomic Force Microscope (AFM)

A Shimadzu AFM-9700 (Shimadzu Corporation, Kyoto, Japan) atomic force microscope operated in the tapping mode was used to analyze surface roughness and morphology changes of CFs. In the experiment, one single filament of CFs was fixed on a glassy board using double-sided tape and it was detected by a silicon nitride probe. The obtained images were in a 5.0 μm × 5.0 μm area with a height of 190 nm.

Monofilament Tension Test

The monofilament tensile strength and tensile modulus of Toray Corporation T300 reference CFs and plasma-treated CFs were characterized by model XQ-1 tensile tester. The clamping distance and stretching velocity were set to 20 mm and 2 mm/min, respectively. The fiber diameter was taken from the average diameter of 100 fibers.

Results and Discussions

The ILSS test was conducted to investigate the effect of plasma modification on the fiber surfaces. Figure a shows the ILSS gradually increased with treatment time. However, the growth rate of the ILSS was not consistent over time. When the treatment time was less than 15 s, ILSS increased quickly. When the treatment time was higher than 15 s, the growth rate decreased with extended treatment time. It is worth noting that this ILSS behavior was different from previous studies on plasma surface modifications, where an approximately linear increase of ILSS over time was observed. Further analysis of the physical and chemical structures of CFs was carried out to elucidate this phenomenon.

Figure 2

(a) The interlaminar shear strength (ILSS), (b) the tensile strength, and (c) the tensile modulus of CFs with different treatment times; (d) XRD pattern of plasma-treated CFs with different treatment times, (e) crystal planar spacings (d002) of plasma-treated CFs with different treatment times, and (f) crystallite sizes (Lc) of plasma-treated CFs with different treatment times.

The monofilament tensile strength and modulus of plasma-treated samples are shown in Figure b and 2c. It was demostrated that the tensile strength and the modulus of CFs were maintained after plasma treatment. Thus, the atmospheric plasma, as a mild surface modification method, did not deteriorate the mechanical properties of the fibers. To further confirm the mechanical performance, the tensile properties of the CF-0s and CF-60s, as in the form of resin-impregnated multifilament tows, were tested on the basis of ASTM D4018-17 standard. Meanwhile, a sample of Toray T300-class carbon fiber was also tested as a reference. The results are shown in Table S2. Compared with the untreated CFs, the mechanical properties of CF-60s were not deteriorated.

The crystalline structures of CF samples were first characterized by XRD, and the crystal planar spacings (d002) and the crystallite sizes (Lc) were calculated and plotted in Figure e and 2f, respectively. On the basis of the results of Lc and d002, no obvious changes of the graphitic crystalline structures had been observed, indicating the plasma treatment had little effect on the overall crystalline structure of CFs. It could be speculated that continuous atmospheric plasma treatment on the fiber surfaces acted as low-intensity modifications, and the scope of modification tended to concentrate on several atomic layers of the surfaces.

The surface morphologies of CFs with different treatment time were characterized using SEM and AFM, as shown in Figure . Clear surface grooves, which are typical features for CFs produced by the wet spinning method, can be observed for all samples. Several previous works revealed that the grooves were etched and the surface was smoothed after electrochemical surface modification or low-pressure plasma treatment.[40] In our case, no distinct change of surface morphology was observed for CFs after plasma treatment, indicating the atmospheric plasma treatment was relatively mild and avoided damaging the surface of CFs. The arithmetic mean roughness Ra was calculated on the basis of AFM results and plotted in Figure b. It can be observed that the Ra noticeably increased after a short period of plasma treatment. The increased surface roughness would promote mechanical interlock between resin and fibers and enhance the interface adhesion as well as the ILSS performance. When the treatment time was longer than 15 s, the increase of Ra stopped, yet the ILSS of CFs was still increasing, indicating other factors were functioning to improve the interfacial properties of CFs.

Figure 3

(a–e) SEM images and (f–j) AFM images (with arithmetic mean roughness Ra inserted) of plasma-treated CFs with different treatment time.

Figure 4

(a) Raman spectra and (b) the arithmetic mean roughness Ra and D1/G ratio of plasma-treated CFs with different treatment time.

Raman spectroscopy was adopted to characterize the graphitic structure on the surface of the fiber, as shown in Figure a. A multipeak fit (D1, D2, D3, D4, and G bands) with a linear baseline was performed, and the corresponding fit curves of the multipeak fit can be found in the Supporting Information Figure S3. The D1 peak at 1360 cm–1 and the G peak at 1600 cm–1 correspond to the sp3 and sp2 carbon atoms, respectively. The D1/G ratio calculated using the fwhm of D1 peak and G peak represented the ratio of amorphous carbon to graphitic carbon. As shown in Figure b, the D1/G ratio increased rapidly after a short period of plasma treatment, indicating the portion of amorphous carbon increased after plasma treatment. When the treatment time was longer than 15 s, the D1/G ratio slowly declined. It is worth noting that the D1/G ratio and the Ra demonstrate virtually identical trends over time, as shown in Figure b. Based on the above observation, it was speculated that because of the limited strength of atmospheric plasma, it cannot etch all the surface carbon atoms simultaneously; since the edged carbon atoms on the graphitic plane were more active, plasma preferably etched graphitic carbon faster than amorphous carbon. The graphitic carbon on the surface of CFs was etched within several seconds, which corresponds to the quickly increased D1/G ratio and Ra values. When the conversion of graphitic carbon was completed, the etching of amorphous carbon continued, but proceeded in a much slower manner. Consequently, the D1/G ratio and Ra values decreased slightly.

XPS analysis was used to characterize the content of elements and functional groups on the surface of CFs, as shown in Figure . The ratios of O/C and N/C represent the content of active functional groups on the surface of fibers. Within 30 s of processing time, only small changes of O/C and N/C were observed. When the treatment time was more than 30 s, the ratios of O/C and N/C showed a sharp rise. It was evident that the grafting of active functional groups was difficult on graphitic planes. Thus, it can be inferred that because of the high degree of graphitization on the surface of CFs, the attaching of oxygen-containing functional groups was restricted at the initial stage of plasma treatment. When the graphitic carbon atoms were etched, the exposed carbon provided active sites for the grafting of reactive oxygen-containing functional groups. These groups could form stable chemical bonds with the resin, which led to the enhanced ILSS performance.[8] A qualitative wettability test of CF samples was performed to examine the above speculation, as shown in Figure S5. Initially, the spreading of water droplets were poor for CFs treated less than 30 s, and the wettability was greatly improved with prolonged treatment time.

Figure 5

(a–e) XPS peak fitting images and (f) the ratio of O/C and N/C ratios of plasma-treated CFs with different treatment time.

To sum up, the effect of atmospheric plasma treatment on CFs could be divided into two stages. In the early stage (treatment time less than 15 s), the effect of plasma was mainly on rapidly etching the graphitic carbon atoms on the surface of CFs, and the ILSS quickly improved as a result of increased surface roughness. In the later stage (treatment time longer than 30 s), the etching of amorphous carbon was much slower, and the effect of plasma was mainly on chemically modifying the surface of CFs by introducing oxygen-containing functional groups, which contributed to the continuously increased ILSS.

Plasma and electrochemical oxidation surface treatments in the same laboratory were compared to estimate the time and cost consumption. Photos of both equipment setups are shown in the Supporting Information Figure S6. It took 7.9 min for electrochemical oxidation (including washing and drying) to process the same length of CFs, which was 7 times more than plasma treatment. The detailed cost estimation can be found in the Supporting Information Figure S7 and Table S3.[40] The cost of plasma surface treatment was $1.52–1.67 per kilogram in the lab, which can save more than two-thirds of the cost compared with traditional electrochemical oxidation treatment. For the industrial production of carbon fibers, the carbon fibers industry benefits from economy of scale,[41,42] which mean that the production costs can be greatly reduced by increasing the throughput of the production line.[43] On the basis of not significantly changing the production line, the way to increase the output is to raise the speed of the line. If the wire speed is doubled, the existing output can be increased to about 1.7 times the original, thus reducing the production cost of carbon fibers. However, it is difficult for traditional electrochemical treatment to match the increased processing speed due to the large size and complexity of the equipment.[44,45] Compared with traditional electrochemical treatment, plasma treatment is compact and can be easily adjusted to match the line speed without significantly modifying the equipment and waste treatments, thus increasing the efficiency and lowering the cost of a CF production line.[46] Plasma treatment has the advantages of being highly efficient, environmentally friendly, and simplistic in operation, which makes it a promising candidate for the surface modification of CFs.

Conclusion

In this work, an atmospheric plasma setup based on DBD method was developed to perform continuous surface treatment on PAN-based carbon fibers. The influence of the physical and chemical structures of the surface on the ILSS of carbon fibers was systematically studied under different treatment time conditions. The ILSS of carbon fibers increased from 64.9 to 80.0 MPa with 60 s plasma treatment. The energy of atmospheric plasma was relatively weak compared with low-pressure plasma and electrochemical oxidation, thus demonstrating a different mechanism of modification. At the starting period of plasma treatment (less than 15 s), the plasma preferably etched the graphitic carbon on the surface of CFs, accompanied by the quickly increased surface roughness and ILSS values. When the etching of graphitic carbon was completed, the etching of amorphous carbon was much slower, and the increase of interfacial adhesion between the fibers and the resin was mainly attributed to the attaching of oxygen-containing functional groups on the exposed carbon sites. This continuous plasma treatment was a facile, effective, and environmentally friendly surface modification method, which is promising for the carbon fibers industry.