Electromagnetic Interference Shielding Effectiveness of Direct-Grown-Carbon Nanotubes/Carbon and Glass Fiber-Reinforced Epoxy Matrix Composites

In this study, carbon nanotubes (CNTs) were grown under the same conditions as those of carbon fibers and glass fibers, and a comparative analysis was performed to confirm the potential of glass fibers with grown CNTs as electromagnetic interference (EMI) shielding materials. The CNTs were grown directly on the two fiber surfaces by a chemical vapor deposition process, with the aid of Ni particles loaded on them via a Ni-P plating process followed by heat treatment. The morphology and structural characteristics of the carbon and glass fibers with grown CNTs were analyzed using scanning electron microscopy–energy dispersive X-ray spectroscopy (SEM–EDS), X-ray diffraction (XRD), and X-ray photoelectron spectrometry (XPS), and the EMI shielding efficiency (EMI SE) of the directly grown CNT/carbon and glass fiber-reinforced epoxy matrix composites was determined using a vector-network analyzer. As the plating time increased, a plating layer serving as a catalyst formed on the fiber surface, confirming the growth of numerous nanowire-shaped CNTs. The average EMI SET values of the carbon fiber-reinforced plastic (CFRP) and glass fiber-reinforced plastic (GFRP) with grown CNTs maximized at approximately 81 and 40 dB, respectively. Carbon fibers with grown CNTs exhibited a significantly higher EMI SET value than the glass fiber-based sample, but the latter showed a higher EMI SET increase rate. This indicates that low-cost, high-quality EMI-shielding materials can be developed through the growth of CNTs on the surface of glass fibers.


Introduction
Emitted electromagnetic (EM) fields cause malfunctions in electronic devices and adverse human health problems such as headaches and dizziness. Thus, electromagnetic interference (EMI) generated by communication and broadcasting networks, power lines, and other electrical products has become a significant concern [1][2][3][4][5]. Due to the recent rapid development of the electronics industry, EMI shielding materials used to prevent such interference have been extensively investigated [6][7][8][9].
Metals such as copper, nickel, and stainless steel are considered the most effective EMI shielding materials [10][11][12]; however, their application is limited owing to high density and corrosion problems [13][14][15][16][17]. In recent years, research has been carried out on lightweight and anticorrosive polymer-based composites, wherein metal-plated fibers and carbon fillers are embedded in a polymer matrix, which can overcome the disadvantages of metal materials. Carbon nanotubes (CNTs) are excellent EMI shielding materials due to their low density, high surface area, and excellent electrical properties [18][19][20][21][22]. However, the utilization of CNTs is difficult because they easily aggregate owing to strong van der Waals interactions, resulting in poor dispersion [23,24]. Various physical and chemical methods

CNTs Grown on Fiber and Their Composites
For the growth of CNTs, the post-heat-treated fiber was loaded into a CVD chamber, heated up to 600 • C at 10 • C/min under a gas flow of Ar (500 cc/min) and H 2 (200 cc/min), and the temperature was maintained for 30 min. Furthermore, C 2 H 4 (200 cc/min) gas was introduced with Ar/H 2 gas for an additional 10 min; the flow of C 2 H 4 and H 2 gases was terminated, followed by cooling to room temperature in an Ar atmosphere. The Materials 2023, 16, 2604 3 of 13 composite was fabricated by uniformly stacking three plies using the hand lay-up method and vacuum-packing at 160 • C under a pressure of 10 MPa for 1 h. Sample names are listed in Table 1.

Characterization
The surface morphologies of the carbon and glass fibers were studied before and after the growth of CNTs using scanning electron microscopy (SEM; S-4800, Hitachi, Japan). To prevent the sample from charging during SEM imaging, it was placed in a holder and coated with platinum nanoparticles at 4 mA for 150 s. The base pressure of the SEM chamber was approximately 5 × 10 −8 Pa, and the acceleration voltage was 10 kV. For further material dispersion analysis, elemental mapping of the cross-section of the glass fiber was conducted using SEM-energy-dispersive X-ray spectroscopy (EDS).
The structural changes following plating, post-heat treatment, and CNT growth on the carbon and glass fibers were analyzed through wide-angle X-ray diffraction (XRD) using an EMPYREAN XRD instrument (PANalytical, Almelo, The Netherlands) equipped with a customized auto-mount and a Cu-Kα radiation source at 40 kV and 30 mA. Diffraction patterns were collected in the 2θ range of 30 • to 90 • at a scan speed of 2 • /min. The spectral baseline was corrected, and the peaks were smoothed by connecting straight lines.
Furthermore, the surface chemical composition of the samples was analyzed through X-ray photoelectron spectroscopy (XPS, PHI 5000 Versa Probe II, ULVAC-PHI, Incorporated Company, Chigasaki, Japan). Unless otherwise specified, the X-ray anode was operated at >5 W, and the voltage was maintained at 5.0 kV. The energy resolution was fixed at 0.50 eV to ensure sufficient sensitivity. The base pressure of the analyzer chamber was 5 × 10 −8 Pa. Both the full-scan spectra (0 to 1200 eV) and narrow ones with a very high resolution for individual elements were recorded. Binding energies were calibrated with respect to the adventitious carbon peak (C 1s : 284.6 eV). The high-resolution C 1s , O 1s , and Ni 2p peaks of the samples were deconvoluted using a Shirley-type baseline and an iterative least squared optimization algorithm. Furthermore, a curve-fitting procedure was carried out using a nonlinear least square curve-fitting program with a Gaussian-Lorentzian production function.
EM parameters were measured on a vector network analyzer (E5062A/EM2107A, Agilent Technologies, Santa Clara, CA, USA) with transmission-reflection mode according to ASTM D4935-89 in the range of 30-1500 MHz at room temperature. EMI SE was evaluated by measuring the attenuation or reduction of the EM wave; it was calculated and expressed in decibels (dB) by using the following equation [48]: where P 1 is the incident power and P 2 is the transmitted power. The total SE (SE T ) can be expressed as the sum of the reflection (SE R ), absorption (SE A ), and multiple reflection (SE MR ) components, as follows [48]: If the SE T value is greater than or equal to 15 dB, the effect of SE MR can be neglected, and the equation can be simplified as [48]: Furthermore, SE R and SE A can be calculated from the power coefficients, which are expressed as [48]: where R, T, and A represent the power coefficients of reflectance, transmittance, and absorbance, respectively. The S parameters, including R, T, and A, are determined from the incident power coefficients, and they can be expressed as [48]: where S 11 , S 22 , S 12 , and S 21 correspond to the input reflection, output reflection, reverse transmission, and forward transmission, respectively. The measurement set-up consisted of a signal generator, specimen holder, and receiver. The composites were compressed under a pressure of~10 MPa at 160 • C for 1 h using a hot press. The prepared sample size was 150 mm × 150 mm × 1.5 mm.

Surface Morphology Analysis
The morphological changes of the carbon and glass fiber samples grown with CNTs according to the plating time are shown in Figures 1 and 2. As shown in Figure 1, as the plating time increases, more Ni-P layers are formed on the carbon fiber surface; this confirms that Ni particles aggregate after post-heat treatment. The Ni-P layer thus formed served as a catalyst for the growth of CNTs. After CVD treatment, numerous nano-sized wire-like structures were formed; these nanowires formed a continuous network through curling and winding. As shown in Figure 2, as the plating time increases, a Ni-P plating layer is formed on the surface of the glass fiber, thus confirming CNT growth after CVD treatment. However, a pit-shaped surface was formed during post-heat treatment after Ni-P plating owing to volume reduction and the formation of a face-centered cubic (FCC) Ni crystal structure when the amorphous Ni-P plating layer is exposed to heat during the post-heat treatment. In the Ni-P plating layer, glass fibers were more uniformly formed than carbon fibers, and it is believed that the glass fibers with high polarity acted favorably toward the formation of Sn/Pd nuclei, resulting in a more uniform plating layer. This uniform plating layer acts as a catalyst for the growth of CNTs, and, thus, a larger amount of CNTs is formed compared to the carbon fiber surfaces. uniform plating layer acts as a catalyst for the growth of CNTs, and, thus, a larger amount of CNTs is formed compared to the carbon fiber surfaces.    uniform plating layer acts as a catalyst for the growth of CNTs, and, thus, a larger amount of CNTs is formed compared to the carbon fiber surfaces.   Figure 3 presents the EDS mapping images of the cross-section of the Ni-P plated glass fiber before and after the growth of CNTs. After Ni-P plating, Ni was detected in the  Figure 3 presents the EDS mapping images of the cross-section of the Ni-P plated glass fiber before and after the growth of CNTs. After Ni-P plating, Ni was detected in the EDS image, which revealed the formation of a plating layer on the surface of the glass fiber. Furthermore, after the growth of CNTs, carbon, which was not observed previously, EDS image, which revealed the formation of a plating layer on the surface of the glass fiber. Furthermore, after the growth of CNTs, carbon, which was not observed previously, was detected. These results indicate the successful formation of the plating layer and subsequent growth of the CNTs on the glass fiber surface.  Figure 4 shows the XRD patterns of the untreated, plated, post-heat treated, and CNTs-grown samples of the carbon and glass fibers. As shown in Figure 4a, the untreated carbon fiber has a C (002) peak at 2θ ≈ 26° and a small Ni-P amorphous peak at 2θ ≈ 45° after plating. After the post-heat treatment, the samples showed various peaks between 2θ ≈ 41° and 55° corresponding to the metallic Ni structure. This is because Ni is crystallized and precipitated after the post-heat treatment, thereby forming Ni3P, whose peak is observed in the XRD pattern. However, after CNT growth, the Ni peaks broadened again in comparison with those of the post-heat-treated carbon fibers. In general, as carbon fibers interact weakly with metals, when CNTs are grown using CVD, first, the hydrocarbon decomposes into carbon and hydrogen species upon contacting with hot metal nanoparticles on the fiber surface. Then, carbon diffuses under the metal and, finally, CNT forms beneath the metal particle, dislodging the entire metal particle off the carbon fiber surface [49]. Thus, the Ni peaks become prominent due to heat treatment after plating and also under the heat supplied during the process of growing CNTs. However, as CNTs grow and the Ni structure gradually collapses, the Ni peak broadens again. Additionally, the intensity of the C (002) peak decreased after plating and then increased again after CNT growth. This result confirms that the C (002) peak increased due to the growth of CNTs on the carbon fiber surface through the tip-growth mechanism. Similarly, in Figure  4b, a small Ni-P amorphous peak was observed at 2θ ≈ 45° after the glass fiber was plated. However, unlike carbon fibers, in glass fibers, sharp FCC Ni crystals were formed after the post-heat treatment. After the post-heat treatment, the thermodynamically unstable Ni-P plating layer forms a stable structure of FCC Ni crystals and body-centered tetragonal (BCT) Ni3P compounds. In an alloy with low P content, Ni precipitates first followed by Ni3P [50,51]. Therefore, in the case of the glass fiber, the sharp FCC Ni crystals were formed in the evenly and thickly coated Ni layer. Moreover, the intensity of the nickel peaks increased, and the formation of the C (002) peak after the CNTs grew. In the case of the glass fiber, highly polar interaction with the metal becomes the driving force, and the anchoring of metal particles to the glass fiber surface causes CNTs to grow on the metal surface, and not on the glass fiber surface. When CNTs are grown on glass fibers by CVD, first, the hydrocarbon decomposes around the metal surface; then, the dissolved carbon diffuses up to the metal surface; finally, the CNT grows on the metal particle. This basegrowth model verifies that the intensity of the C (002) peak increases due to the growth of CNTs on the glass fiber surface. Additionally, during the growth of CNTs on the glass  Figure 4 shows the XRD patterns of the untreated, plated, post-heat treated, and CNTsgrown samples of the carbon and glass fibers. As shown in Figure 4a, the untreated carbon fiber has a C (002) peak at 2θ ≈ 26 • and a small Ni-P amorphous peak at 2θ ≈ 45 • after plating. After the post-heat treatment, the samples showed various peaks between 2θ ≈ 41 • and 55 • corresponding to the metallic Ni structure. This is because Ni is crystallized and precipitated after the post-heat treatment, thereby forming Ni 3 P, whose peak is observed in the XRD pattern. However, after CNT growth, the Ni peaks broadened again in comparison with those of the post-heat-treated carbon fibers. In general, as carbon fibers interact weakly with metals, when CNTs are grown using CVD, first, the hydrocarbon decomposes into carbon and hydrogen species upon contacting with hot metal nanoparticles on the fiber surface. Then, carbon diffuses under the metal and, finally, CNT forms beneath the metal particle, dislodging the entire metal particle off the carbon fiber surface [49]. Thus, the Ni peaks become prominent due to heat treatment after plating and also under the heat supplied during the process of growing CNTs. However, as CNTs grow and the Ni structure gradually collapses, the Ni peak broadens again. Additionally, the intensity of the C (002) peak decreased after plating and then increased again after CNT growth. This result confirms that the C (002) peak increased due to the growth of CNTs on the carbon fiber surface through the tip-growth mechanism. Similarly, in Figure 4b, a small Ni-P amorphous peak was observed at 2θ ≈ 45 • after the glass fiber was plated. However, unlike carbon fibers, in glass fibers, sharp FCC Ni crystals were formed after the post-heat treatment. After the post-heat treatment, the thermodynamically unstable Ni-P plating layer forms a stable structure of FCC Ni crystals and body-centered tetragonal (BCT) Ni 3 P compounds. In an alloy with low P content, Ni precipitates first followed by Ni 3 P [50,51]. Therefore, in the case of the glass fiber, the sharp FCC Ni crystals were formed in the evenly and thickly coated Ni layer. Moreover, the intensity of the nickel peaks increased, and the formation of the C (002) peak after the CNTs grew. In the case of the glass fiber, highly polar interaction with the metal becomes the driving force, and the anchoring of metal particles to the glass fiber surface causes CNTs to grow on the metal surface, and not on the glass fiber surface. When CNTs are grown on glass fibers by CVD, first, the hydrocarbon decomposes around the metal surface; then, the dissolved carbon diffuses up to the metal surface; finally, the CNT grows on the metal particle. This base-growth model verifies that the intensity of the C (002) peak increases due to the growth of CNTs on the glass fiber surface. Additionally, during the growth of CNTs on the glass fiber surface, additional heat was supplied to the plating layer; this caused an increase in the intensity of the Ni peaks. The CNT growth mechanism is illustrated in Figure 5 [49]. fiber surface, additional heat was supplied to the plating layer; this caused an increase i the intensity of the Ni peaks. The CNT growth mechanism is illustrated in Figure 5

Analysis of the Surface Characteristics
The chemical composition and surface properties of the post-heat-treated and CNT grown glass fiber surfaces were investigated by XPS. The XPS profiles of the untreated post-heat-treated, and CNT-grown glass fiber samples are shown in Figure 6. The spectr clearly show the peaks for carbon (C1s), oxygen (O1s), nickel (Ni2p), and silicon (Si2p) in th samples. The untreated glass fibers mainly show peaks for O1s (binding energy (BE) = 531 eV), C1s (BE = 284.3 eV), and Si2p (BE = 101.9 eV). In the case of post-heat-treated sample Ni2p (BE = 855.8 eV) and Ni3p (BE = 68.4 eV) peaks appeared, and although a plating laye was formed, the O1s peak intensity did not decrease. We infer that the O1s peak did no decrease due to the influence of the nickel oxide of the Ni plating layer, which was thickl formed on the surface of the glass fiber. In the XPS profile of the CNT-grown sample, th Ni2p, Ni3p peaks, and O1s peaks almost disappeared, and the intensity of the C1s peak in 6, x FOR PEER REVIEW 7 of 13 fiber surface, additional heat was supplied to the plating layer; this caused an increase in the intensity of the Ni peaks. The CNT growth mechanism is illustrated in Figure 5 [49]. 10

Analysis of the Surface Characteristics
The chemical composition and surface properties of the post-heat-treated and CNTgrown glass fiber surfaces were investigated by XPS. The XPS profiles of the untreated, post-heat-treated, and CNT-grown glass fiber samples are shown in Figure 6. The spectra clearly show the peaks for carbon (C1s), oxygen (O1s), nickel (Ni2p), and silicon (Si2p) in the samples. The untreated glass fibers mainly show peaks for O1s (binding energy (BE) = 531.3 eV), C1s (BE = 284.3 eV), and Si2p (BE = 101.9 eV). In the case of post-heat-treated samples, Ni2p (BE = 855.8 eV) and Ni3p (BE = 68.4 eV) peaks appeared, and although a plating layer was formed, the O1s peak intensity did not decrease. We infer that the O1s peak did not decrease due to the influence of the nickel oxide of the Ni plating layer, which was thickly formed on the surface of the glass fiber. In the XPS profile of the CNT-grown sample, the Ni2p, Ni3p peaks, and O1s peaks almost disappeared, and the intensity of the C1s peak increased. This indicates that numerous CNTs were successfully grown on the plating layer.

Analysis of the Surface Characteristics
The chemical composition and surface properties of the post-heat-treated and CNTgrown glass fiber surfaces were investigated by XPS. The XPS profiles of the untreated, post-heat-treated, and CNT-grown glass fiber samples are shown in Figure 6. The spectra clearly show the peaks for carbon (C 1s ), oxygen (O 1s ), nickel (Ni 2p ), and silicon (Si 2p ) in the samples. The untreated glass fibers mainly show peaks for O 1s (binding energy (BE) = 531.3 eV), C 1s (BE = 284.3 eV), and Si 2p (BE = 101.9 eV). In the case of post-heattreated samples, Ni 2p (BE = 855.8 eV) and Ni 3p (BE = 68.4 eV) peaks appeared, and although a plating layer was formed, the O 1s peak intensity did not decrease. We infer that the O 1s peak did not decrease due to the influence of the nickel oxide of the Ni plating layer, which was thickly formed on the surface of the glass fiber. In the XPS profile of the CNT-grown sample, the Ni 2p , Ni 3p peaks, and O 1s peaks almost disappeared, and the intensity of the Materials 2023, 16, 2604 8 of 13 C 1s peak increased. This indicates that numerous CNTs were successfully grown on the plating layer.
The high-resolution XPS O 1s profiles of untreated and post-heat-treated glass fiber samples are shown in Figure 7. For the case of untreated glass fibers, the O 1s peak consists of Si-O and Si-O-Si peaks. However, NiO and Ni 2 O 3 peaks also contributed to the O 1s peak after post-heat treatment following plating. This is due to the oxidation of Ni particles during the post-heat treatment after plating. The so-formed plating layer serves as a catalyst for the growth of CNTs, and the XPS O 1s peak almost disappeared after the growth of CNTs. Si 2p Ni 3p Figure 6. Survey-scan X-ray photoelectron spectra of the glass fibers modified by different treatments.
The high-resolution XPS O1s profiles of untreated and post-heat-treated glass fiber samples are shown in Figure 7. For the case of untreated glass fibers, the O1s peak consists of Si-O and Si-O-Si peaks. However, NiO and Ni2O3 peaks also contributed to the O1s peak after post-heat treatment following plating. This is due to the oxidation of Ni particles during the post-heat treatment after plating. The so-formed plating layer serves as a catalyst for the growth of CNTs, and the XPS O1s peak almost disappeared after the growth of CNTs.

Electromagnetic Interference Shielding Efficiency Behavior
Generally, the frequency range of 0.3-1.5 GHz is associated with radio frequency (RF). RF is widely used in wireless communications and in other technologies, such as  Si 2p Ni 3p Figure 6. Survey-scan X-ray photoelectron spectra of the glass fibers modified by different treatments.
The high-resolution XPS O1s profiles of untreated and post-heat-treated glass fiber samples are shown in Figure 7. For the case of untreated glass fibers, the O1s peak consists of Si-O and Si-O-Si peaks. However, NiO and Ni2O3 peaks also contributed to the O1s peak after post-heat treatment following plating. This is due to the oxidation of Ni particles during the post-heat treatment after plating. The so-formed plating layer serves as a catalyst for the growth of CNTs, and the XPS O1s peak almost disappeared after the growth of CNTs.

Electromagnetic Interference Shielding Efficiency Behavior
Generally, the frequency range of 0.3-1.5 GHz is associated with radio frequency (RF). RF is widely used in wireless communications and in other technologies, such as mobile communications, wireless optical communications, Bluetooth, Wi-Fi, and GPS. They are also used extensively in medical, industrial, and military applications. These frequency ranges are extremely important for research and applications aimed at addressing various technical problems. Figure 8a,b shows the power coefficients of A, R, and T

Electromagnetic Interference Shielding Efficiency Behavior
Generally, the frequency range of 0.3-1.5 GHz is associated with radio frequency (RF). RF is widely used in wireless communications and in other technologies, such as mobile communications, wireless optical communications, Bluetooth, Wi-Fi, and GPS. They are also used extensively in medical, industrial, and military applications. These frequency ranges are extremely important for research and applications aimed at addressing various technical problems. Figure 8a,b shows the power coefficients of A, R, and T calculated from S parameters, such as S 11 and S 21 collected under various conditions to determine the actual shielding ability of the carbon and glass fibers with surface-grown CNTs. In the case of carbon fibers, the R value of all samples after CNT growth is significantly higher than that of A. This result indicates that the amount of energy blocked by reflection is larger than that blocked by absorption. For glass fibers, there was little shielding effect up to HNCF-CNTs-30, because T was higher than A and R. However, as T decreased from HNCF-CNTs-60, it was confirmed that the shielding effect was improved, and R was rapidly improved as in the case of the carbon fibers. According to the EMI shielding theory [52], the interaction of the mobile charge carriers of a material leads to the reflection of EM waves, while the absorption of EM waves depends on the dielectric loss and/or magnetic loss of the material. Therefore, as CNTs were grown on the surface of the fiber, the electron mobility and electron density on the material surface were increased, which improved the interaction between the incident EM waves and electrons; consequently, the incident EM waves were significantly reflected. A, R, and T are quantitative parameters, while EMI SE is a relative parameter that is independent of the absolute power coefficient. Parameter A represents the ratio of the attenuated power of the EMI shielding material to the total incident power, but SE A indicates the ability of the shielding material to disperse the EM waves. Therefore, the shielding mechanism is evaluated based on the contribution rates of SE R and SE A to SE T . Figure 8c,d show the EMI SE T over the frequency range of 0.3-1.5 GHz of the carbon fiber-reinforced plastic (CFRP) and glass fiber-reinforced plastic (GFRP) with CNTs grown on the fiber surface as a function of the plating time. In Figure 8c, the CFRP with grown CNTs shows high EMI SE of >50 dB. As the plating time increases, many CNTs grow on the carbon fiber surface, improving the shielding behavior in the low-frequency region (30~600 MHz). However, a slight improvement in the shielding behavior was confirmed in the high-frequency region (600~1500 MHz) in relation to that in the low-frequency region. In Figure 8d, in the case of the GFRP with grown CNTs, GF-CNTs-10 and GF-CNTs-30 show less than 10 dB over all frequency ranges. However, as the plating time increases, the number of CNTs on the glass fiber surface increases, confirming that EMI SE improved up to ∼43 dB in the high-frequency region. These results indicate the possibility that glass fiber with grown CNTs can be used to replace carbon fiber as an EMI shielding material by changing the growth conditions of CNTs. Figure 8e,f shows the SE T , SE A , and SE R values of the CFRP and GFRP with grown CNTs. The SE T , SE A , and SE R values of both the carbon fiber and glass fiber composites tended to increase with increasing Ni plating time, and the SE A value of all composites was higher than SE R , indicating that the shielding effect due to absorption dominated the SE T . This EMI shielding mechanism is schematically illustrated in Figure 9. First, as EM waves are incident on the surface of a shielding material, some of the EM waves are immediately reflected due to the impedance mismatch between the intrinsic impedance of the shielding material and the impedance of the propagation medium. This results in reflection losses, which occur owing to energy transmission through the boundary and the resulting energy loss. Thereafter, most of the EM waves that enter the shielding material are converted into heat owing to multiple reflections, resulting in absorption loss; that is, energy dissipation. As multi-layer composites have two or more layers, EMI shielding is achieved by the repetition of the EM wave absorption/reflection and EM wave penetration/re-absorption processes several times. The study findings suggest that the CNTs grown using a Ni-P layer, formed on the fiber surface via Ni-P plating, as the catalyst improve the EMI SE performance of the composite material.

Conclusions
In this study, CNTs were grown through the CVD process under the same conditions as carbon fibers (a conductor) and glass fibers (an insulator). A comparative analysis was performed to investigate the potential of glass fiber as an EMI shielding material. It was confirmed that a large amount of CNTs in the form of nanowires was formed on carbon and glass fibers as the plating treatment time increased. Glass fibers grew more CNTs than carbon fibers owing to the higher number of plating layers, which served as catalysts for CNT growth, formed on glass fibers with relatively high polarity. In terms of the EMI SET, HNCF-CNTs-120 and HNGF-CNTs-120, which were prepared using the longest plating time, exhibited the highest EMI SET value in the frequency range of 0.3 GHz to 1.5 GHz. In the case of the carbon fiber, which is a conductor, the average EMI SET value over all frequency ranges tended to increase after the growth of CNTs, and all samples presented a high average EMI SET value of ≥70 dB. In the case of the glass fiber, which is an insulator, the HNGF-CNTs-30 sample exhibited a low EMI SET value; however, as the plating time increased, the average EMI SET value maximized to ~40 dB. Both carbon and glass fibers improved EMI SET after CNT growth; however, glass fiber showed a higher EMI SET increase rate compared to carbon fiber. These results suggest that low-cost and high-quality electromagnetic shielding materials can be developed via the growth of CNTs on the surface of glass fibers.

Conclusions
In this study, CNTs were grown through the CVD process under the same conditions as carbon fibers (a conductor) and glass fibers (an insulator). A comparative analysis was performed to investigate the potential of glass fiber as an EMI shielding material. It was confirmed that a large amount of CNTs in the form of nanowires was formed on carbon and glass fibers as the plating treatment time increased. Glass fibers grew more CNTs than carbon fibers owing to the higher number of plating layers, which served as catalysts for CNT growth, formed on glass fibers with relatively high polarity. In terms of the EMI SE T , HNCF-CNTs-120 and HNGF-CNTs-120, which were prepared using the longest plating time, exhibited the highest EMI SE T value in the frequency range of 0.3 GHz to 1.5 GHz. In the case of the carbon fiber, which is a conductor, the average EMI SE T value over all frequency ranges tended to increase after the growth of CNTs, and all samples presented a high average EMI SE T value of ≥70 dB. In the case of the glass fiber, which is an insulator, the HNGF-CNTs-30 sample exhibited a low EMI SE T value; however, as the plating time increased, the average EMI SE T value maximized to~40 dB. Both carbon and glass fibers improved EMI SE T after CNT growth; however, glass fiber showed a higher EMI SE T increase rate compared to carbon fiber. These results suggest that low-cost and high-quality electromagnetic shielding materials can be developed via the growth of CNTs on the surface of glass fibers.