Enhancing the Interfacial Strength of Carbon Fiber/Poly(ether ether ketone) Hybrid Composites by Plasma Treatments

As a promising alternative to traditional prepreg, carbon fiber/poly(ether ether ketone) (CF/PEEK) hybrid composites have attracted wide public interest for their flexibility and conformability. However, modification methods focused on the hybrid premix have not been previously studied. In the present work, the interfacial strength of the hybrid composite was improved by treating the carbon and PEEK fibers together in a radiofrequency (RF) plasma containing one of the following gases to achieve surface activation: air, Ar, or Ar–air. After plasma treatment, the increased roughness of CF and the grafted chemical groups of CFs and PEEK fibers were propitious to the mechanical interlocking and interfacial strength. Significant interfacial shear strength (IFSS) enhancement was achieved after Ar 1 min, air 1 min plasma treatment. This study offers an alternative method for improving the interfacial properties of CF/PEEK composites by focusing on the boundary layer and modifying and controlling the fiber–matrix interface.


Introduction
With the increasing demand for high-performance composites for aerospace, automobile, and other structural applications in engineering, great efforts have been devoted to the study and development of fiber-reinforced thermoplastic composites [1][2][3]. Among these, carbon fiber/poly(ether ether ketone) (CF/PEEK) composites have attracted wide attention for the superior properties of their semicrystalline PEEK matrix, including their high service temperature, outstanding mechanical properties, and dimensional stability [4][5][6]. However, the high melting viscosity and the difficulty of infiltrating carbon fibers during CF/PEEK composite fabrication must be addressed. Many studies have investigated the 2-D and 3-D hybrid CF/PEEK premixes, a subject that has attracted wide attention [4,7]. The resulting mechanical properties of CF/PEEK depend mainly on the interfacial property, which plays an important role in stress transfer between the reinforcement and the matrix [8][9][10]. The physical state and chemical structure of CFs and PEEK fiber determine the interfacial strength [11]. Hence, the low interfacial bonding strength between CF and PEEK fiber matrices due to the intrinsically smooth, hydrophobic, and chemically inert CF surface is a problem that must be addressed. Recently, significant scientific efforts have focused on the surface modification of CF to improve the interfacial Hefei, China), which is usually used to clean nanoscale organic pollutants on substrates or crystals. This equipment is composed of a vacuum chamber, a needle valve, a resistance vacuum gauge, a vacuum pump, and a radio source (Figure 1a). The plasma treatment was conducted in an inductively coupled radio frequency (13.56 MHz) plasma reactor with a high-power mode (18 W). Gas was fed into the vacuum chamber at a flow rate of about 6-8 standard cubic centimeters per minute (SCCM). The operational pressure was set at 1 Torr. Dozens of carbon and PEEK fiber tows were carefully fixed across a glass frame and treated using air plasma (2 min), Ar plasma (2 min), and Ar-air plasma (Ar 15 s, air 15 s; Ar 1 min, air 1 min; Ar 5 min, air 5 min) for different times. Samples treated by Arair were labeled as AA15, AA1, and AA5. After the samples were subjected to different treatments, they were stored in a desiccator at room temperature until they were analyzed by SEM, Raman spectroscopy, DSC/TGA, and XPS.
By definition, plasma is a neutrally ionized gas on a spatial scale, containing an equal number of negative and positive ions, a small percentage of free radicals, a few parts of atoms, excited molecules and free electrons, and a large amount of extremely energetic vacuum ultraviolet light ( Figure 1b). The basic theoretical foundation in plasma treatment of materials is free-radical chemistry. Gases are activated and dissociated using a radio source, which acts via electron bombardment and photochemical processes to create a high density of gas-phase free radicals. Among them, free radicals in electronically excited states carry a great deal of energy that is sufficient to break any organic bond. This results in the abstraction of atoms or molecular fragments that can react further in the plasma to form volatile species. In addition, abstraction causes a progressive ablation of organic surfaces and the formation of residual free radicals on that surface. These can either react with themselves, resulting in surface cross-linking, or react with the plasma gas, even ground state molecules, to form new chemical species on the surface. If the gas used to produce plasma is inert (like argon), after treatment, the surface would contain a large number of stable radicals that can further react until exposed to reactive gases. Thus, plasma treatments are effective for breaking and creating bonds on the surface of polymers or organically contaminated substrates [23].

Surface Morphology
Scanning electron microscope (Merlin Compact, ZEISS, Dresden, Germany) measurements were used to observe surface topography and provide a general analysis for the material before and after modification.

Raman Spectroscopy
Raman spectroscopy is an effective experimental technique used to determine information on carbon fiber microstructure. Raman spectroscopy data were collected using a INVIA010410 Laser micro confocal Raman spectrometer (Renishaw, London, UK) with a wavelength and laser beam size of 514.5 nm of argon laser (green) and ~2 μm diameter. A 500× magnification microscope was used

Surface Morphology
Scanning electron microscope (Merlin Compact, ZEISS, Dresden, Germany) measurements were used to observe surface topography and provide a general analysis for the material before and after modification.

Raman Spectroscopy
Raman spectroscopy is an effective experimental technique used to determine information on carbon fiber microstructure. Raman spectroscopy data were collected using a INVIA010410 Laser micro confocal Raman spectrometer (Renishaw, London, UK) with a wavelength and laser beam size of 514.5 nm of argon laser (green) and~2 µm diameter. A 500× magnification microscope was used to select a point of focus for the laser beam spot on the fiber's surface. The scanning range was set at 500-4000 cm −1 to obtain the first-and second-order Raman bands for CF. The band intensity, band width, and band position were obtained from the first-order Raman spectra using a Gauss curve-fitting procedure.

Single-Fiber Tensile Test
Single-fiber tensile tests were performed on CFs treated with different plasma conditions using a Sintech universal testing machine following the ASTM D3379 standard test method. The test sample was obtained by cutting a section of 12 cm fiber, pulling out to get the monofilament, and then fixing it on the paper frame using glue and tensile force to keep it straight, as shown in Figure 2. During the test, the paper frame was fixed vertically in the upper and lower clamps of the electronic universal testing machine. The paper frame was cut off before loading. After debugging the instrument, the load was applied, and the peak load was recorded automatically by the testing machine. In this experiment, paper frame gauge was selected as 2 cm.
Polymers 2019, 11, 753 4 to select a point of focus for the laser beam spot on the fiber's surface. The scanning range was set at 500-4000 cm −1 to obtain the first-and second-order Raman bands for CF. The band intensity, band width, and band position were obtained from the first-order Raman spectra using a Gauss curvefitting procedure.

Single-Fiber Tensile Test
Single-fiber tensile tests were performed on CFs treated with different plasma conditions using a Sintech universal testing machine following the ASTM D3379 standard test method. The test sample was obtained by cutting a section of 12 cm fiber, pulling out to get the monofilament, and then fixing it on the paper frame using glue and tensile force to keep it straight, as shown in Figure 2. During the test, the paper frame was fixed vertically in the upper and lower clamps of the electronic universal testing machine. The paper frame was cut off before loading. After debugging the instrument, the load was applied, and the peak load was recorded automatically by the testing machine. In this experiment, paper frame gauge was selected as 2 cm. The chemical structure of PEEK before and after plasma treatment was examined based on the Nicolet iS50 Fourier transform infrared spectra (Thermo Fisher, Waltham, MA, USA), using the KBr pellet method at 500-4000 cm −1 wavelength.

Thermal Analyses
Differential scanning calorimetry measurements were performed in order to study the effect of plasma treatment on the melting point of PEEK fibers using STA449F3 synchronous thermal analyzer (NETZSCH, Selb, Germany). The fibrous PEEK, cut into small length sizes with ~10 mg weight, were thermally cycled between 25 °C and 400 °C twice at a constant heating/cooling rate of 10 °C min −1 in a nitrogen atmosphere. The thermogravimetric analysis tests for PEEK fibers (~10-15 mg) were conducted in a nitrogen atmosphere from 25 °C to 800 °C at a heating rate of 10 °C min −1 to study changes in the PEEK fiber's thermal stability induced by plasma treatment.
2.3.6. X-ray Photoelectron Spectroscopy (XPS) XPS analyses were conducted using the ESCALAB 250Xi (Thermo Fisher, Waltham, MA, USA) to study the surface elements and groups of CFs and PEEK fibers after plasma treatment. XPS spectra were obtained using an Al Ka (hv = 1486.6 eV) monochromated X-ray source at 15 kV and 150 W. The pass energy and energy step were 20.0 eV and 0.05 eV, respectively. A nonlinear least-square curvefitting program (XPSPEAK software 4.1, Raymund W.M.Kwork) was used to deconvolve the XPS data. To compensate for the charging effects, all spectra were calibrated with graphitic carbon as the reference at a binding energy (BE) of 284.6 eV. The spectra were deconvolved by subtracting a Shirley + linear background, and a Lorentzian-Gaussian (GL = 20%) mixed function. The surface chemical composition was calculated from the peak areas of relevant spectra.

Interfacial Shear Strength
The microdroplet test was conducted on HM410 evaluation equipment, which determines the

Fourier Transform Infrared Spectra (FTIR)
The chemical structure of PEEK before and after plasma treatment was examined based on the Nicolet iS50 Fourier transform infrared spectra (Thermo Fisher, Waltham, MA, USA), using the KBr pellet method at 500-4000 cm −1 wavelength.

Thermal Analyses
Differential scanning calorimetry measurements were performed in order to study the effect of plasma treatment on the melting point of PEEK fibers using STA449F3 synchronous thermal analyzer (NETZSCH, Selb, Germany). The fibrous PEEK, cut into small length sizes with~10 mg weight, were thermally cycled between 25 • C and 400 • C twice at a constant heating/cooling rate of 10 • C min −1 in a nitrogen atmosphere. The thermogravimetric analysis tests for PEEK fibers (~10-15 mg) were conducted in a nitrogen atmosphere from 25 • C to 800 • C at a heating rate of 10 • C min −1 to study changes in the PEEK fiber's thermal stability induced by plasma treatment.

X-Ray PhRotoelectron Spectroscopy (XPS)
XPS analyses were conducted using the ESCALAB 250Xi (Thermo Fisher, Waltham, MA, USA) to study the surface elements and groups of CFs and PEEK fibers after plasma treatment. XPS spectra were obtained using an Al Ka (hv = 1486.6 eV) monochromated X-ray source at 15 kV and 150 W. The pass energy and energy step were 20.0 eV and 0.05 eV, respectively. A nonlinear least-square curve-fitting program (XPSPEAK software 4.1, Raymund W.M.Kwork) was used to deconvolve the XPS data. To compensate for the charging effects, all spectra were calibrated with graphitic carbon as the reference at a binding energy (BE) of 284.6 eV. The spectra were deconvolved by subtracting a Shirley + linear background, and a Lorentzian-Gaussian (GL = 20%) mixed function. The surface chemical composition was calculated from the peak areas of relevant spectra.

Interfacial Shear Strength
The microdroplet test was conducted on HM410 evaluation equipment, which determines the interfacial property of composite material (Tohei Sangyo Co., Ltd, Tokyo, Japan)-to directly test the interfacial sheer strength (IFSS) between CF and PEEK. The process is illustrated in Figure 3a. In this experiment, each microdroplet was clamped by two steel blades and pulled to separate with the fiber at a constant speed of 0.1 mm/min. The microdroplet test specimens were prepared on alloy frames (30 mm × 70 mm) with the free fiber length~30 mm between the frames. Both sides of the CF single fiber were firmly fixed on the alloy frame, and some oval PEEK droplets obtained by melting the PEEK fibers were formed on a CF monofilament, as shown in Figure 3b. To obtain perfectly shaped PEEK droplets, we started by laying an alloy frame fixed with filament on a closed heating platform at 375 • C for 10 min and sequentially melted one side of the PEEK tows while holding the other side to form fine PEEK fiber to cross the monofilament. This process was repeated for a sufficient number of times. The single fiber composite specimens were then transferred to a vacuum oven at 375 • C for 5 min to reheat the composite specimens in order to obtain perfectly shaped microdroplets and wetted interfaces.
Polymers 2019, 11, 753 5 at a constant speed of 0.1 mm/min. The microdroplet test specimens were prepared on alloy frames (30 mm × 70 mm) with the free fiber length ~30 mm between the frames. Both sides of the CF single fiber were firmly fixed on the alloy frame, and some oval PEEK droplets obtained by melting the PEEK fibers were formed on a CF monofilament, as shown in Figure 3b. To obtain perfectly shaped PEEK droplets, we started by laying an alloy frame fixed with filament on a closed heating platform at 375 °C for 10 min and sequentially melted one side of the PEEK tows while holding the other side to form fine PEEK fiber to cross the monofilament. This process was repeated for a sufficient number of times. The single fiber composite specimens were then transferred to a vacuum oven at 375 °C for 5 min to reheat the composite specimens in order to obtain perfectly shaped microdroplets and wetted interfaces.

Morphology of CF and PEEK Fiber
The surface morphology of CFs at different plasma treatment times are shown in Figure 4a-f. The data demonstrate the presence of clear ridges and striations on the surface of CF parallel to the fiber's axial direction. These formations are a result of the carbon fiber manufacturing process, such as stretching and sizing, and the gullies provide an excellent condition for infiltration and adhesion of polymers. Within a short plasma treatment time like 30 s (the total plasma treatment time), the patches on the surface of untreated carbon fibers almost disappeared due to the etching action and cleaning effect of the plasma treatment. Patch removal from the surface ameliorated the potential weak boundary condition in fiber-reinforced polymer composites, thus improving the connection between the fiber and the matrix. However, the modification effects within the same time varied depending on the type of gas (shown in Figure 4c,d), owing to the different energy density of the plasma. Plasma energy density is related to treatment time, plasma power, and area of the electrodes [24]. In this condition, the number and size of samples treated by air plasma with the largest plasma power was the most complicated among CF-air, CF-Ar, and CF-AA1, whereas CF-Ar was the simplest [24]. Increasing plasma treatment time was associated with increased complexity and generation of embossments on the surface. The number and size of bulges increased gradually with prolonged treatment time. This phenomenon can be attributed to the bombardment of high-energy particles or free radicals in the plasma on the carbon fiber surface. Plasma sputtering can destroy the intrinsic chains and surface groups of fiber, and some atoms would be knocked off at the same time [25][26][27]. Prolonged plasma treatment time provides a favorable condition, allowing the accumulation of these atoms and chains to form embossments.

Morphology of CF and PEEK Fiber
The surface morphology of CFs at different plasma treatment times are shown in Figure 4a-f. The data demonstrate the presence of clear ridges and striations on the surface of CF parallel to the fiber's axial direction. These formations are a result of the carbon fiber manufacturing process, such as stretching and sizing, and the gullies provide an excellent condition for infiltration and adhesion of polymers. Within a short plasma treatment time like 30 s (the total plasma treatment time), the patches on the surface of untreated carbon fibers almost disappeared due to the etching action and cleaning effect of the plasma treatment. Patch removal from the surface ameliorated the potential weak boundary condition in fiber-reinforced polymer composites, thus improving the connection between the fiber and the matrix. However, the modification effects within the same time varied depending on the type of gas (shown in Figure 4c,d), owing to the different energy density of the plasma. Plasma energy density is related to treatment time, plasma power, and area of the electrodes [24]. In this condition, the number and size of samples treated by air plasma with the largest plasma power was the most complicated among CF-air, CF-Ar, and CF-AA1, whereas CF-Ar was the simplest [24]. Increasing plasma treatment time was associated with increased complexity and generation of embossments on the surface. The number and size of bulges increased gradually with prolonged treatment time. This phenomenon can be attributed to the bombardment of high-energy particles or free radicals in the plasma on the carbon fiber surface. Plasma sputtering can destroy the intrinsic chains and surface groups of fiber, and some atoms would be knocked off at the same time [25][26][27]. Prolonged plasma treatment time provides a favorable condition, allowing the accumulation of these atoms and chains to form embossments. Typical SEM images of the PEEK fibers before and after plasma treatment are shown in Figure  5a-f. The images clearly demonstrate that the surface of untreated PEEK fibers was relatively smooth and uniform without major irregularities. However, several patches existed on the surface of untreated PEEK fibers, which were not very obvious based on the fact that we could only find little traces in a comprehensive SEM analysis. These could be eliminated under the etching action and the cleaning effect after 30 s of plasma treatment. After a short plasma treatment time, e.g., 2 min, the surface of PEEK remained smooth with several obvious patches resulting from the high-energy particles or free radical bombardment on the surface, while the morphology of PEEK fibers changed according to the energy density of different plasma gases. As with CF, PEEK fibers treated with air plasma were rougher and displayed more patches than PEEK-Ar and PEEK-AA1. Increasing plasma treatment time resulted in an increase in the quantity and size of patches [28].
These data show that plasma treatment etches the surface of CFs and PEEK fibers and significantly changes the surface morphology. Although the etching depth is determined by time and instrument, the etching effect closely corresponds with the type of plasma gas. Prolonging plasma treatment time causes outstanding etching, resulting in a more complex surface morphology with increased surface roughness for both CFs and PEEK fibers [29]. Certainly, long treatment times influence the intrinsic properties of fibers, such as the mechanical properties of CFs and thermal properties of PEEK fibers, which are important during the molding process and application of CF/PEEK composites. Hence, further tests were performed, and discussions focusing on these particular questions are explained in the next section of this article. The complex surface morphology and increased roughness of CFs provide a large specific area for CF and PEEK bonding and thus enhance the mechanical intercalation between CF and PEEK fiber resins, which is an important factor for improving the interfacial adhesion between fibers and resins. Typical SEM images of the PEEK fibers before and after plasma treatment are shown in Figure 5a-f. The images clearly demonstrate that the surface of untreated PEEK fibers was relatively smooth and uniform without major irregularities. However, several patches existed on the surface of untreated PEEK fibers, which were not very obvious based on the fact that we could only find little traces in a comprehensive SEM analysis. These could be eliminated under the etching action and the cleaning effect after 30 s of plasma treatment. After a short plasma treatment time, e.g., 2 min, the surface of PEEK remained smooth with several obvious patches resulting from the high-energy particles or free radical bombardment on the surface, while the morphology of PEEK fibers changed according to the energy density of different plasma gases. As with CF, PEEK fibers treated with air plasma were rougher and displayed more patches than PEEK-Ar and PEEK-AA1. Increasing plasma treatment time resulted in an increase in the quantity and size of patches [28].
These data show that plasma treatment etches the surface of CFs and PEEK fibers and significantly changes the surface morphology. Although the etching depth is determined by time and instrument, the etching effect closely corresponds with the type of plasma gas. Prolonging plasma treatment time causes outstanding etching, resulting in a more complex surface morphology with increased surface roughness for both CFs and PEEK fibers [29]. Certainly, long treatment times influence the intrinsic properties of fibers, such as the mechanical properties of CFs and thermal properties of PEEK fibers, which are important during the molding process and application of CF/PEEK composites. Hence, further tests were performed, and discussions focusing on these particular questions are explained in the next section of this article. The complex surface morphology and increased roughness of CFs provide a large specific area for CF and PEEK bonding and thus enhance the mechanical intercalation between CF and PEEK fiber resins, which is an important factor for improving the interfacial adhesion between fibers and resins.

Structure Analysis
Graphitic carbon and other sp 2 -bonded amorphous carbons possess strong Raman scatter despite their intense optical absorption. A Raman peak at ~1585 cm −1 is typical of bulk crystalline graphite, called the G band (ordered or graphitic). This peak represents the basic vibration mode of graphite crystal, and its intensity is closely related to the crystal size. The Raman peak at 1360 cm −1 originates from vibrations of the edge of graphite carbon and is called the D band, which represents disordered carbon. Thus, R-the ratio of D and G bands (ID/IG) is usually calculated and regarded as a measure of the size of graphite crystallites and the amount of amorphous carbon phase [31,32]. In our study, R values were obtained from the ratio of curve areas at ~1360 cm −1 and 1585 cm −1 , and the crystalline size La was calculated using the following relationship: where C = 44 Å. The Raman spectra and its cumulative fit peak for CF before and after different plasma treatments are shown in Figure 6. The corresponding R and La of untreated and treated carbon fibers obtained from Figure 6 are listed in Table 1. The data show that the value of R and La did not significantly change with different plasma treatments and times. Thus, it can be concluded that plasma treatment might not change the crystalline or graphitic structure of carbon fibers because plasma treatments can only influence surfaces on a nanometer range.

Structure Analysis
Graphitic carbon and other sp 2 -bonded amorphous carbons possess strong Raman scatter despite their intense optical absorption. A Raman peak at~1585 cm −1 is typical of bulk crystalline graphite, called the G band (ordered or graphitic). This peak represents the basic vibration mode of graphite crystal, and its intensity is closely related to the crystal size. The Raman peak at 1360 cm −1 originates from vibrations of the edge of graphite carbon and is called the D band, which represents disordered carbon. Thus, R-the ratio of D and G bands (I D /I G ) is usually calculated and regarded as a measure of the size of graphite crystallites and the amount of amorphous carbon phase [30][31][32]. In our study, R values were obtained from the ratio of curve areas at~1360 cm −1 and 1585 cm −1 , and the crystalline size L a was calculated using the following relationship: where C = 44 Å. The Raman spectra and its cumulative fit peak for CF before and after different plasma treatments are shown in Figure 6. The corresponding R and L a of untreated and treated carbon fibers obtained from Figure 6 are listed in Table 1. The data show that the value of R and L a did not significantly change with different plasma treatments and times. Thus, it can be concluded that plasma treatment might not change the crystalline or graphitic structure of carbon fibers because plasma treatments can only influence surfaces on a nanometer range.

Tensile Strength of Carbon Fibers
As the etching effect depends on the plasma treatment time and equipment, we focused on the tensile strength of CFs treated for different times. The Weibull plots and fitted straight lines for singlefiber tensile tests are presented in Figure 7. The parameters required to obtain a fitted straight line, including scale parameter of unit fiber length ratio (σ0), Gamma function (Γ(σ0)), slope B, and intercept A of the straight line, are listed in Table 2. Based on the test results, the single-fiber tensile strength of untreated CFs was 3.25 GPa, and the value changed slightly during the tested plasma treatment times. This result indicates that a 10 min plasma treatment time merely affects the CF surface and causes little damage to their essential structures. However, as time extends, the etching depth increases, and damage to the intrinsic quality of CFs is inevitable [32]. Therefore, when treating CFs using plasma, its effect on the strength of bulk fibers should be considered and studied.

Tensile Strength of Carbon Fibers
As the etching effect depends on the plasma treatment time and equipment, we focused on the tensile strength of CFs treated for different times. The Weibull plots and fitted straight lines for single-fiber tensile tests are presented in Figure 7. The parameters required to obtain a fitted straight line, including scale parameter of unit fiber length ratio (σ 0 ), Gamma function (Γ(σ 0 )), slope B, and intercept A of the straight line, are listed in Table 2. Based on the test results, the single-fiber tensile strength of untreated CFs was 3.25 GPa, and the value changed slightly during the tested plasma treatment times. This result indicates that a 10 min plasma treatment time merely affects the CF surface and causes little damage to their essential structures. However, as time extends, the etching depth increases, and damage to the intrinsic quality of CFs is inevitable [32]. Therefore, when treating CFs using plasma, its effect on the strength of bulk fibers should be considered and studied.

Chemical Structure of PEEK Fiber
We used FTIR spectroscopy to investigate PEEK fibers before and after plasma treatment. The wide spectra from 4000 to 500 cm −1 for PEEK and different plasma-treated PEEK are shown in Figure  8a. In addition, PEEK and PEEK-AA5 with peak positions [32][33][34] are separately compared for detailed analyses in Figure 8b, with peak assignments provided in Table 3. Wavenumber values are shown for both spectra, and they were identified by reference to the absorbance bands shown in Table  3. FTIR measurements showed subtle distinctions between PEEK fibers under specific plasma conditions, even though significant peak data were presented. This could possibly be attributed to the technique's 1-2 μm testing depth. Furthermore, the low sensitivity of FTIR impedes its application for determining trace components. As reported by the producer, the rate of surface etching is 20 nm/min at "Hi" for this plasma treatment equipment. In our test, the maximum treatment time was 10 min, and thus the etching depth was ~200 nm, which is far from the testing depth of 1-2 μm. This was still regarded as an effective result as it suggests that plasma treatment is aimed at surface modification, and the effects are therefore exclusively surface effects.

Chemical Structure of PEEK Fiber
We used FTIR spectroscopy to investigate PEEK fibers before and after plasma treatment. The wide spectra from 4000 to 500 cm −1 for PEEK and different plasma-treated PEEK are shown in Figure 8a. In addition, PEEK and PEEK-AA5 with peak positions [32][33][34] are separately compared for detailed analyses in Figure 8b, with peak assignments provided in Table 3. Wavenumber values are shown for both spectra, and they were identified by reference to the absorbance bands shown in Table 3. FTIR measurements showed subtle distinctions between PEEK fibers under specific plasma conditions, even though significant peak data were presented. This could possibly be attributed to the technique's 1-2 µm testing depth. Furthermore, the low sensitivity of FTIR impedes its application for determining trace components. As reported by the producer, the rate of surface etching is 20 nm/min at "Hi" for this plasma treatment equipment. In our test, the maximum treatment time was 10 min, and thus the etching depth was~200 nm, which is far from the testing depth of 1-2 µm. This was still regarded as an effective result as it suggests that plasma treatment is aimed at surface modification, and the effects are therefore exclusively surface effects.  Aromatic out-of-plane bending 825, 757 C-H out-of-plane bending substitution patterns

Thermal Property Analysis
To study the effect of plasma treatments on the thermal properties of PEEK fibers, DSC/TGA was applied to test the samples' glass transition temperature (Tg) and thermal stability. DSC and TG curves of PEEK fibers treated with plasma for different lengths of time are shown in Figure 9. The data show that the glass transition temperature Tg of PEEK at 343 °C did not change after different durations of plasma treatment. Figure 9b shows that PEEK fibers and modified PEEK fibers began to lose weight at 525 °C, and the mass loss rate reached the maximum at 600 °C. Meanwhile, the thermogravimetric curves of PEEK fibers changed slightly, indicating that PEEK fiber modification using plasma treatments do not affect the thermal properties of PEEK. In this study, we introduced the etching and surface modification effects, which are both devoted to nanoscale depths of the surface material, making it difficult to influence the permanent thermal properties of PEEK fibers [35].  Skeletal in-plane vibration of aromatic ring 1474, 1407 Aromatic rotations 1300-1050 Diphenyl ether group, C-O-C rotation and stretch 923 Aromatic out-of-plane bending 825, 757 C-H out-of-plane bending substitution patterns

Thermal Property Analysis
To study the effect of plasma treatments on the thermal properties of PEEK fibers, DSC/TGA was applied to test the samples' glass transition temperature (T g ) and thermal stability. DSC and TG curves of PEEK fibers treated with plasma for different lengths of time are shown in Figure 9. The data show that the glass transition temperature T g of PEEK at 343 • C did not change after different durations of plasma treatment. Figure 9b shows that PEEK fibers and modified PEEK fibers began to lose weight at 525 • C, and the mass loss rate reached the maximum at 600 • C. Meanwhile, the thermogravimetric curves of PEEK fibers changed slightly, indicating that PEEK fiber modification using plasma treatments do not affect the thermal properties of PEEK. In this study, we introduced the etching and surface modification effects, which are both devoted to nanoscale depths of the surface material, making it difficult to influence the permanent thermal properties of PEEK fibers [35].

Surface Chemical Analyses
Wide-scan spectra with binding energy ranging from 0 to 1300 eV was obtained to identify surface elements, and the detailed narrow-scan spectra C1s and O1s were quantitatively analyzed to identify the chemical groups. C, O, and N contents on CFs and PEEK fibers before and after plasma

Surface Chemical Analyses
Wide-scan spectra with binding energy ranging from 0 to 1300 eV was obtained to identify surface elements, and the detailed narrow-scan spectra C1s and O1s were quantitatively analyzed to identify the chemical groups. C, O, and N contents on CFs and PEEK fibers before and after plasma modifications are shown in Table 4. The major constituents on CF and PEEK were carbon and oxygen. The appearance of nitrogen was due to the plasma modifications including air, Ar, and Ar-air. The proportion of oxygen increased after plasma treatment, both in CFs and PEEK fibers. In addition, the ratio of oxygen increased with prolonged plasma treatment time of 10 min. Plasma treatments by Ar-air increased the oxygen content compared with air and Ar treatments performed for the same amount of time (2 min). This can be explained by the reaction between the grafted active sites provided by Ar and air plasma [36]. The bonding energies corresponding to the groups that deconvolve from C1s of CFs and PEEK fibers and their concentrations are listed in Table 5 [36,37]. The proportion of C1 and C2 peaks for CF decreased after plasma treatment. For most samples, the proportion of C3, C4, and C5 increased by plasma modification to different extents depending on the types and times, except the C3 peak of CF-AA5, which showed an opposite trend. These changes indicated that the sp 2 and sp 3 carbon atoms were activated, damaged, and replaced with new oxygen on the surface during plasma treatment. Therefore, various oxidative reactions induced by plasma were predicted to occur. As reported, oxygen and argon plasma were able to create free radicals that coupled with active species in the plasma environment [38]. Due to the active C=C bond, it was more susceptible to plasma attack, and radicals were generated on the dissociated C=C bond for further reaction with active oxygen atoms. This was the reason for the decrease in the proportion of C=C bond after plasma treatment. Through this process, C-O bond was created, as illustrated in Figure 10a, while the new C=O bond was formed from oxygen radicals on the C-C bond, as shown in Figure 10b. The formation of O-C=O bonds was generated by C=O bond through combining radicals on the C=O bond with active oxygen atoms, as displayed in Figure 10c. However, the three presented reactions occurred throughout the plasma treatment process, and different bonds were activated or generated at the same time along with the changes in the element ratio. Hence, it is hard to describe the effect by the change of one group. Thus, the polar ratio combined with elemental proportion were introduced to study the effects caused by type and time of plasma treatment. Compared with air and Ar plasma treatments, CFs treated sequentially by Ar and air gained the largest number of polar groups. Proportions of polar peaks for CF-air, CF-Ar and CF-AA1 were close to each other, but the oxygen content of AA1 was larger than the other two for the fast graft effect after activation by Ar, indicating the better activation capacity of Ar-air. Prolonged treatment times increased oxygen content but not polar group content. Ratios in Table 5 were defined as the product of polar/nonpolar and effective element contents, including oxygen and nitrogen. Hence, we can conclude that samples treated with Ar-air gain the most effective group sites, which provide the most favorable conditions for interfacial adhesion. In addition, these methods are designed for hybrid CF/PEEK pre-prepreg; thus, CFs and PEEK fibers should be modified simultaneously. The CF results highlight the optimization conditions. The XPS study for PEEK fibers focused on the samples treated by Ar-air. For PEEK fibers, group concentration changes were more obvious. C3, C4, and C5 increased, while C1 and C2 decreased. However, the CFs and PEEK fiber group contents showed different degrees of increase or decrease. It can still be concluded that the polar oxygen groups were grafted onto CF and PEEK fiber surfaces by plasma treatment, but they varied with gas type and time. Furthermore, increasing polar contents on CFs and PEEK fibers probably supplied reactive sites for interfacial interaction between CFs and PEEK fibers.
type and time of plasma treatment. Compared with air and Ar plasma treatments, CFs treated sequentially by Ar and air gained the largest number of polar groups. Proportions of polar peaks for CF-air, CF-Ar and CF-AA1 were close to each other, but the oxygen content of AA1 was larger than the other two for the fast graft effect after activation by Ar, indicating the better activation capacity of Ar-air. Prolonged treatment times increased oxygen content but not polar group content. Ratios in Table 5 were defined as the product of polar/nonpolar and effective element contents, including oxygen and nitrogen. Hence, we can conclude that samples treated with Ar-air gain the most effective group sites, which provide the most favorable conditions for interfacial adhesion. In addition, these methods are designed for hybrid CF/PEEK pre-prepreg; thus, CFs and PEEK fibers should be modified simultaneously. The CF results highlight the optimization conditions. The XPS study for PEEK fibers focused on the samples treated by Ar-air. For PEEK fibers, group concentration changes were more obvious. C3, C4, and C5 increased, while C1 and C2 decreased. However, the CFs and PEEK fiber group contents showed different degrees of increase or decrease. It can still be concluded that the polar oxygen groups were grafted onto CF and PEEK fiber surfaces by plasma treatment, but they varied with gas type and time. Furthermore, increasing polar contents on CFs and PEEK fibers probably supplied reactive sites for interfacial interaction between CFs and PEEK fibers.

Interface Shear Strength
IFSS before and after plasma treatment was determined using the microdroplet test, and the destruction are shown in Figure 11. For untreated CF/PEEK system, there was not much retained matrix, but after plasma treatments, the retained matrix increased. The IFSS results are listed in Table 6. The IFSS of untreated CF/PEEK composites was 42.36 ± 4.23 MPa and increased to a maximum of 59.73 ± 3.88 MPa after plasma treatment with an increment of 41.01%. Compared with untreated CF/PEEK (Figure 11a) with AA1 plasma treatment samples (Figure 11i), the retained matrix of AA1 was greater, and there was PEEK resin on the surface of the CF. The increase in IFSS can be attributed to the increased mechanical interlocking and chemical bonding. Overall, the trend of IFSS was consistent with the polar/element ratio for strong chemical interaction. However, the increment was not proportional to the polar group contents for the effect of roughness and wettability between CFs and PEEK fibers after plasma treatment. It can be concluded that plasma treatment by Ar-air can significantly increase the interfacial strength between CFs and PEEK fibers and can be applied for the modification of CF/PEEK woven fabrics.
6. The IFSS of untreated CF/PEEK composites was 42.36 ± 4.23 MPa and increased to a maximum of 59.73 ± 3.88 MPa after plasma treatment with an increment of 41.01%. Compared with untreated CF/PEEK (Figure 11a) with AA1 plasma treatment samples (Figure 11i), the retained matrix of AA1 was greater, and there was PEEK resin on the surface of the CF. The increase in IFSS can be attributed to the increased mechanical interlocking and chemical bonding. Overall, the trend of IFSS was consistent with the polar/element ratio for strong chemical interaction. However, the increment was not proportional to the polar group contents for the effect of roughness and wettability between CFs and PEEK fibers after plasma treatment. It can be concluded that plasma treatment by Ar-air can significantly increase the interfacial strength between CFs and PEEK fibers and can be applied for the modification of CF/PEEK woven fabrics.

Conclusions
In this work, the physical and chemical structures of CFs and PEEK fibers before and after plasma treatments were investigated by both surface morphology and surface chemical structure analyses using SEM, AFM, Raman spectra, FTIR, and XPS. In addition, IFSS was tested using the microdroplet test.
SEM results demonstrated the surface morphology of fibers, thus providing a brief explanation of the physical combination and interlock between CFs and PEEK fibers. Improving CF roughness is propitious to the enhancement of its IFSS. The significant properties of CFs and PEEK fibers were studied to examine the effect of plasma over different testing times. In conclusion, plasma treatments were unable to affect the tensile strength of carbon fibers and the thermal properties of PEEK fibers within 10 min in our study. XPS analyses indicated the chemical state of the surface. Plasma treatment was shown to be an effective method for increasing oxygen content and active groups on CFs and PEEK fibers.
Active groups partly revealed the interfacial interaction and adhesion between CF and PEEK during the molding process. To make the work more satisfactory, IFSS values after different plasma treatments were tested, and composites treated with Ar for 1 min and air for 1 min in sequence had