Composite Structures xxx (2020) 113286

Contents lists available at ScienceDirect

Composite Structures

journal homepage: www.elsevier .com/locate /compstruct

Design and verification of simultaneously self-sensing and microwave-absorbing composite structures based on embedded SiC fiber network

https://doi.org/10.1016/j.compstruct.2020.113286Received 29 April 2020; Revised 20 October 2020; Accepted 30 October 2020Available online xxxx0263-8223/© 2020 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: cgkim@kaist.edu (C.-G. Kim).

Please cite this article in press as: Kwon H et al. Design and verification of simultaneously self-sensing and microwave-absorbing composite structures bembedded SiC fiber network. Compos Struct (2020), https://doi.org/10.1016/j.compstruct.2020.113286

Hyunseok Kwon a, Min-Su Jang a, Jong-Min Yun a, Yurim Park a, Changkyo Shin a, Jinkyu Yang b,Chun-Gon Kim a,⇑aDepartment of Aerospace Engineering, KAIST, Daejeon 34141, Republic of KoreabWilliam E. Boeing Department of Aeronautics and Astronautics, University of Washington, Seattle, WA 98195, USA

A R T I C L E I N F O A B S T R A C T

Keywords:Multi‐functional composite structureFunctional fibersSilicon carbide fibersMicrowave‐absorbing structurePiezoresistive sensor

Silicon carbide (SiC) fibers can act not only as a load‐bearing component, but also as a piezoresistive and micro-wave lossy material when embedded in fiber‐reinforced polymer composites. By leveraging these multifunc-tional characteristics, we propose simultaneous self‐sensing and microwave‐absorbing composite structuresbased on embedded SiC fiber networks. A SiC fiber network consists of SiC fiber yarns embedded in each inter-laminar of the host composite at regular intervals. The SiC fiber yarns, which do not come into contact withother SiC fiber yarns, act as independent sensing elements using their piezoresistivity. By positioning theSiC fiber yarns at regular intervals within the wavelength of the targeted microwave‐absorption bandwidth,the SiC fiber network functions as a microwave lossy material. In this study, specific focus has been placedon designing a multi‐functional composite that senses external impacts and also absorbs microwaves in theX‐band (8.2–12.4 GHz). The simultaneous self‐sensing and microwave‐absorbing performance of the fabricatedSiC‐embedded composite was experimentally verified through microwave return loss measurements anddynamic signal acquisition tests. The proposed SiC fiber network‐embedded composite structure has greatpotential for use in future aerospace structures.

1. Introduction

With the continued progress of technology, the requirement forstructures has moved beyond the conventional function of loadbear-ing. To embrace added functionality, nonstructural features such asself‐sensing, self‐healing, microwave‐absorbing, and energy storagecapabilities have been integrated into conventional load‐bearing struc-tures. Composites comprising two or more distinctive materials havedriven the development of multi‐functional structures. Since the phys-ically combined constituents of composite materials can retain theirinherent characteristics even after the fabrication process, severalresearchers have developed multifunctional composite structures byintroducing and embedding various types of materials into compositearchitectures [1].

Previous approaches to build multi‐functional composite materialscan be categorized into two groups. The first group aims to control thematrix of composite materials. One example is the addition of fillers,such as carbon nanotubes (CNTs) [2–5] and graphene [3,5–7], to thematrix to leverage the extraordinary properties of these fillers.

Matrix‐based multi‐functional composites can easily optimize multi-functional performance by controlling the content of fillers. However,the multifunctional performance of such composites can be affected bythe dispersion and alignment of the functional particles in the matrix[8]. Therefore, the dispersion and alignment of fillers for consistentand homogeneous performance of applications in practical structuresrequires further investigation.

The other method for achieving multi‐functional composites is toexploit the inherent multifunctional properties of fibers. Carbon fiberis a representative multi‐functional material. Owing to the parallelalignment of the carbon atom layers along the fiber axis, the carbonfiber exhibits excellent mechanical properties relative to its weightand volume, as well as good thermal and electric conductivities inthe fiber direction [9,10]. Its excellent electrical properties allow car-bon fiber to be a component of electrical systems for energy storage[11–13], electromagnetic interference shielding [14–20], and sensingpurposes [21–23]. Compared to matrix‐based multi‐functional com-posites, fiber‐based composites have advantages in the fabrication pro-cess because they can use the same resin system as the host composite

ased on

H. Kwon et al. Composite Structures xxx (2020) 113286

structure. However, traditionally, the performance of functional fibershas been inferior compared to functional matrix‐based composites.Consequently, efforts have been made to enhance the functional per-formance of fibers and find new functional fibers with superiorcharacteristics.

Silicon carbide (SiC) fibers have gained increasing attention asfunctional fibers in recent years, primarily due to their superior heatresistance. Most of the commercially available SiC fibers arepolymer‐derived SiC fibers [24], produced by the thermal degradationof precursor fibers made of polycarbosilane, similar to the productionprocess of polyacrylonitrile (PAN)‐based carbon fibers [24,25]. Conse-quently, in a manner similar to that of carbon fibers, the mechanicalproperties of SiC fibers have been improved by reducing the defectscontained inside and outside the filaments. In this process, the oxygencontent of the SiC fiber has been significantly reduced, and the electri-cal conductivity has been remarkably enhanced. As a result, some SiCfibers demonstrate semi‐conductive characteristics. Materials withsemi‐conductive characteristics possess excellent multi‐functional fea-tures. Recent studies have focused on the functionalization of compos-ite structures using semi‐conductive SiC fibers with excellent multi‐functional characteristics.

One of the applications of such semi‐conductive SiC fibers is amicrowave‐absorbing composite structure [26–29]. To design amicrowave‐absorbing composite structure, the microwave loss charac-teristics of the material need to be controlled. In previous studies, elec-tromagnetic particles, such as CNTs [30,31] and carbon black [32],were added to the matrix to control the microwave loss characteristics.This method can optimally design the microwave loss characteristicsof a material by controlling the quantity of fillers. However, as previ-ously mentioned, the consistent and homogeneous performance of thematrix‐based microwave‐absorbing composite structure remains elu-sive due to the challenges in dispersing and aligning fillers in thematrix [8,31,33]. This fabrication problem can be resolved by usingsemi‐conductive SiC fibers. The performance of microwave‐absorbing composite structures using SiC fibers is comparable to thatof conventional microwave‐absorbing structures with various fillersin the matrix [26–29].

Another application field for semi‐conductive SiC fibers is that ofsensing. Semi‐conductive materials have been widely adopted as sen-

Fig. 1. Conceptual diagram of the embedded SiC fiber network-based self-sensing(ΔR/R0) measurements via piezoresistive SiC fiber sensor network. (Bottom rightelectric conductor (PEC).

2

sors—for example, high‐sensitivity strain gauges and pressure sen-sors—because of their excellent piezoresistivity [34]. Recently, thestrain‐sensing characteristics of semi‐conductive SiC have been inves-tigated [35]. This study showed that the semi‐conductive SiC fibersexhibited excellent strain sensitivity with a gauge factor (GF) of8.25. The resistance change was also highly linear up to the strainrange of 1.36%. The GF was approximately four times larger than thatof carbon fibers, and the linear resistance change range was approxi-mately 1% wider than that of carbon fibers. Furthermore, becausethe SiC fibers have the same generic shape as the reinforcing fibersof composites, degradation of the mechanical properties of the com-posites due to geometrical discontinuities, which is a typical problemof embedment of conventional sensors, could be minimized withoutforgoing the structural stiffness and strength of the composites[36–38].

While previous studies have explored the versatile characteristicsof semi‐conductive SiC fibers, their multi‐functional performancewhen embedded as a single network in composites has not yet beendemonstrated. In this study, we designed, fabricated, and tested self‐sensing and microwave‐absorbing composite structures simultane-ously by leveraging the piezoresistive and microwave loss characteris-tics of the semi‐conductive SiC fibers. Fig. 1 shows a conceptualdiagram of the proposed structure based on a network of SiC fiberyarns embedded at each interlaminar of the host composite structure.The SiC fiber yarns were embedded in parallel directions to the rein-forcing fibers of the host structure at regular intervals, thus contribut-ing to the load‐bearing performance. Because the SiC fiber yarns donot come into contact with other SiC fiber yarns, they can beemployed as independent sensing elements using their piezoresistivity(bottom left inset of Fig. 1). To add the microwave‐absorbing featureto the sensing feature, the SiC fiber yarn arrangement is crucial. Thedistance between the microwave lossy SiC fibers needs to be con-trolled within the wavelength of the targeted microwave‐absorptionbandwidth, so that the SiC fiber network‐embedded composites canabsorb targeted microwaves efficiently (bottom right inset of Fig. 1).To this end, we conducted full‐wave simulations of various SiC net-works and determined an appropriate configuration. We fabricatedthe designed multifunctional composite with the embedded SiC net-work, tested its performance under impact and radiation conditions,

and microwave-absorbing structure. (Bottom left) Electrical resistance change) Microwave absorption via dielectric lossy SiC network backed by a perfect

H. Kwon et al. Composite Structures xxx (2020) 113286

and verified its simultaneous self‐sensing and microwave‐absorbingperformance.

The structure of this paper is as follows. In Section 2, we discuss thedesign of the SiC fiber network‐embedded composite structure forsimultaneous self‐sensing and microwave‐absorption. Particular focusis placed on identifying the appropriate SiC network for targetedmicrowave absorption using numerical simulations. In Section 3.1,we explain the fabrication of the designed SiC fiber network‐embedded composite structure. In Section 3.2, we demonstrate exper-imentally the microwave‐absorbing performance of the fabricatedcomposite. In Section 3.3, we acquire sensing signals from the SiC net-work upon external impact and compare them with accelerometer sig-nals. Section 4 concludes this paper.

2. Design of SiC fiber network-embedded composite

We begin with the design of the SiC fiber network for microwaveabsorption and self‐sensing. As mentioned, previously, the SiC fibernetwork consists of SiC fiber yarns embedded at each interlaminar ofthe host composites. For SiC fiber yarns to act as sensing elements,they need to be positioned over the monitoring area of interest. Toachieve microwave‐absorbing characteristics, the dielectric loss char-

Fig. 2. Modeling of the SiC fiber n

3

acteristics of the composite structure need to be fine‐tuned by strategi-cally aligning the SiC fiber yarns. Thus, the design of the SiC fibernetwork in terms of its spatial arrangement is integral to realizingSiC fiber network‐embedded composites for microwave absorptionand self‐sensing. To achieve this, we investigated the microwave losscharacteristics of SiC fiber network‐embedded composites under vari-ous alignments of SiC fiber yarns using full‐wave simulation. Subse-quently, we designed the SiC fiber network‐embedded composite asa single‐slab microwave absorber based on the principle of the Dallen-bach layer. In this study, the modeling and simulation of SiC fibernetwork‐embedded composites were conducted using CST MicrowaveStudio (Dassault Systèmes Simulia Corp.).

2.1. Modeling of SiC fiber network-embedded composite

Modeling of the SiC fiber network‐embedded composite structurewas conducted for its full‐wave simulation. Fig. 2 illustrates the mod-eling process of the SiC fiber network‐embedded composite. First, aSiC fiber yarn was modeled. The cross‐section of the SiC fiber yarnwas modeled as an ellipse, the major and minor axes of the ellipsebeing 1.5 and 0.1 mm, respectively. The fiber direction of the SiC fiberyarn was modeled as a straight line without undulation. In this regard,

etwork-embedded composite.

H. Kwon et al. Composite Structures xxx (2020) 113286

in a previous study, Lee et al. demonstrated that the undulation offibers was insignificant for microwave simulation [15]. For full‐wavesimulation, electromagnetic properties, represented as complex per-mittivity, are required. The relative complex permittivity of the SiCfiber yarn was 10.71–30.62 j at 10 GHz. Subsequently, the SiC fibernetwork‐embedded composite was modeled. The modeled SiC fiberyarns were embedded at each interlaminar of the host composite struc-ture. Glass/epoxy composite was selected as a host structure intowhich the SiC fiber network was embedded. The stacking sequenceof the SiC fiber network‐embedded composite model was set to thecross ply of ½0=SiC0=90=SiC90�2s. The effective thickness per ply ofthe glass/epoxy composite was 0.13 mm, which accounted for theincrease in the thickness per ply due to the embedding of SiC fibers.The total thickness of the SiC fiber network‐embedded compositemodel was 1.04 mm. The complex permittivity of the glass/epoxycomposite was 5.49–0.028 j at 10 GHz. Notably, the glass/epoxy com-posite was not modeled layer by layer in this full‐wave simulationbecause the same microwave lossy properties could be appliedthroughout the plies. The representative elementary volume (REV)of the SiC fiber network‐embedded‐composite model was used forfull‐wave simulation, as shown in the inset of Fig. 2.

In the process of designing the SiC fiber network‐embedded com-posite, there were two design parameters. The first design parameterwas the distance between the SiC fiber yarns in the same interlaminarlayers. For microwave absorption, the distance between the SiC fiberyarns needed to be controlled within the wavelength of the target fre-quency. In this study, the target frequency range was the X‐band(8.2–12.4 GHz). Thus, the distance between the SiC fibers was con-trolled to within 24.19 mm, which was the minimum wavelength inthe X‐band (see the side view of the laminate in Fig. 2).

The second design parameter was an alignment pattern betweenthe SiC fiber yarn groups in different interlaminar layers. As shownin the bottom panel of Fig. 2, depending on the alignment patterns,the SiC fiber network’s effective projected area to the normal incidentmicrowave could be varied. In this study, three alignment patternswere simulated, in which the SiC fiber network’s effective projectedareas were altered while maintaining the fraction of SiC fiber yarnsper unit area (i.e., eight SiC fibers were used in the REVs of all threepatterns, as shown in the bottom panel of Fig. 2). Alignment pattern#1 had the smallest effective projected area, and alignment pattern#3 had the largest effective projected area.

2.2. Full-wave simulation of SiC fiber network-embedded composite

A full‐wave simulation of the SiC fiber network‐embedded compos-ite model was conducted. First, the complex permittivity of the SiC

Fig. 3. a) Simulation model of the SiC fiber network-embedded composite and b)(at 10 GHz).

4

fiber network‐embedded composite was simulated by varying the dis-tance between the SiC fiber yarns in the same interlaminar layers forthe three patterns considered in this study (see the lower panel inFig. 2). Fig. 3(a) shows the simulation model for Pattern #3 with itsboundary conditions. Waveguide ports were located in the y‐axis planenormal to the composite model, and electromagnetic waves of trans-verse electromagnetic mode were applied for the incidence of electro-magnetic waves into the plate. As boundary conditions, tangentialelectric fields and normal magnetic fluxes were set to zero along thez‐axis, and the tangential magnetic fields and normal electric fluxeswere set to zero along the x‐axis. For the design parameters, the sim-ulation range of the distance between the SiC fibers in the same inter-laminar layers was from 6–15 mm for each alignment pattern.

Fig. 3(b) illustrates the simulated complex permittivity of the SiCfiber network‐embedded composite according to the design parame-ters. In Fig. 3(b), an optimum complex permittivity curve for thesingle‐slab absorber design at the target frequency of 10 GHz is repre-sented by a black line with square data points. The optimum complexpermittivity represents the complex permittivity in which the singleslab absorber given a specific design thickness does not reflect themicrowave incident at a certain frequency. It is calculated by Eq.(1), where d is the absorber thickness, ɛr is the relative complex per-mittivity, and λ is the wavelength of the microwave [39].

1 ¼ 1ffiffiffiffiɛr

p tanh j2πdλ

ffiffiffiffiɛr

p� �ð1Þ

As shown in Fig. 3(b), as the distance between the SiC fiber yarns inthe same interlaminar layers decreases, the complex permittivity of theSiC fiber network‐embedded composite increases. The decrease in thedistance between the SiC fiber yarns implies an increased fraction ofthe yarns per unit area. This makes sense as the semi‐conductive SiCfibers possess a higher complex permittivity compared to the glass/epoxy composite. Therefore, when the fraction of the SiC fibersincreases, the complex permittivity of the SiC fiber network‐embedded composite increases. As the distance between the SiC fiberyarns in the same interlaminar layers is maintained for the three align-ment patterns, the complex permittivity is changed due to the align-ment patterns with different effective projected areas. However, thereal part of the complex permittivity does not demonstrate a consistenttrend with the effective projected area. Conversely, the imaginary partof the complex permittivity increases with the effective projected area,and the magnitude of the change is larger than that of the real part. Bychoosing an appropriate distance between the SiC fiber yarns, giventhe alignment pattern, we can obtain the complex permittivity of theSiC fiber network‐embedded composite close to the optimum complexpermittivity. Furthermore, the simulation results reveal that the com-

simulated complex permittivity of the SiC fiber network-embedded composite

H. Kwon et al. Composite Structures xxx (2020) 113286

plex permittivity of the SiC fiber network‐embedded composite can beadjusted to a desirable complex permittivity by choosing an appropri-ate stacking pattern of the SiC fiber network (i.e., by varying the effec-tive projected area of the SiC fiber network) when the fraction of SiCfibers per unit area is limited.

As a subsequent step, a single‐slab microwave absorber based onthe principle of the Dallenbach layer was designed. The Dallenbachlayer comprises an absorber layer with dielectric loss characteristicsand a perfect electric conductor (PEC) that completely reflects themicrowave at the back of the absorber layer. A major microwave‐absorption principle of the Dallenbach layer is the destructive interfer-ence of the reflected microwaves. Fig. 4 illustrates the destructive

Fig. 4. Destructive interference between microwaves reflected from thesingle-slab microwave absorber based on the principle of the Dallenbach layer.

Fig. 5. (a) Simulation model of the designed SiC fiber network-embedded compnetwork-embedded composite.

5

interference between microwaves reflected from the single‐slab micro-wave absorber based on the principle of the Dallenbach layer. When amicrowave is incident on the absorber, the incident microwave is par-tially reflected from the first interface between the air and the absor-ber (surface‐reflected microwave), and the transmitted microwave iscompletely reflected from the second interface between the absorberand the PEC (secondary reflected microwave). In order to inducedestructive interference between the reflected microwaves, the sec-ondary reflected microwave is guided to be 180° out of phase to thesurface‐reflected microwave. This is achieved by designing the thick-ness of the absorber layer to be a quarter wavelength of the transmit-ted microwave. For dielectric materials with real and imaginarypermeability values of 1 and 0, the wavelength of the microwavetransmitted into the absorber is determined by the complex permittiv-ity of the microwave absorber layer. Therefore, using the complex per-mittivity of the microwave absorber layer, the thickness of theabsorber layer can be determined.

Returning to Fig. 3(b), for the alignment patterns #1, #2, and #3,the complex permittivity is close to the optimum permittivity curve asthe SiC fiber yarns are aligned at a distance of approximately 7, 8, and9 mm, respectively. As mentioned previously, the optimum complexpermittivity represents the complex permittivity in which the single‐slab absorber—given a specific design thickness—does not reflectthe microwave incident at a certain frequency. In this study, the stack-ing sequence of the host composite was set to the cross ply, and twoperpendicular layers of 0° and 90° were set to one‐unit layers. Thus,the SiC fiber network‐embedded composite could be designed in mul-tiples of the unit layer. As shown in Fig. 3(b), the complex permittivityof the SiC fiber network‐embedded composite was proximal to theoptimum complex permittivity in the thickness range of approximately3.0–3.1 mm. In this thickness range, the designable thickness glass/epoxy composite was 3.12 mm, which was composed of 24 glass/epoxy plies. The complex permittivity of the SiC fiber network‐embedded composite with alignment pattern #3 and the distancebetween SiC fiber yarns of 9 mm was closest to the optimum complexpermittivity at 3.12 mm. Furthermore, this combination approachedthe optimum complex permittivity with a relatively small amount ofSiC fiber per unit area compared to the other combinations. Therefore,as design parameters for the SiC fiber network‐embedded microwavecomposite using semi‐conductive SiC fibers, the alignment pattern#3 and SiC fiber yarn distance of 9 mm were selected for furtherinvestigation.

The microwave‐absorbing performance of the SiC fiber network‐embedded composite with the selected design parameters was simu-

osite and (b) simulated microwave reflection loss of the designed SiC fiber

H. Kwon et al. Composite Structures xxx (2020) 113286

lated. Fig. 5(a) shows the simulation model. The SiC fiber network‐embedded composite with a thickness of 3.12 mm was modeled witha layer of PEC that completely reflected microwaves at the back of theabsorber layer. Fig. 5(b) shows the simulated electromagnetic wavereflection loss of the designed SiC fiber network‐embedded composite.The frequency range of more than 90% of the electromagnetic waveenergy absorption ranged from 8.25–11.83 GHz, and the −10 dBbandwidth was 3.58 GHz. The minimum return loss was −39.00 dBat 9.87 GHz. Based on the numerical simulations, the SiC network‐embedded composite plate designed here successfully exhibitedmicrowave‐absorbing performance that covered most of the X‐band.

3. Fabrication and performance verification of SiC fiber network-embedded composite structure

The designed SiC fiber network‐embedded composite structure wasfabricated and its self‐sensing and microwave‐absorbing performance

Fig. 6. The fabrication process of the SiC fibe

6

were experimentally verified. To verify the microwave‐absorbing per-formance of the designed SiC fiber network‐embedded composite, themicrowave reflection loss of the fabricated composite structure wasmeasured. Subsequently, to show that the SiC fiber yarns function assensors, dynamic signal acquisition using the embedded SiC fiber sen-sors was performed.

3.1. Fabrication of SiC fiber network-embedded composite structure

Fig. 6 illustrates the fabrication process of the SiC fiber network‐embedded glass/epoxy composite structure. The SiC fiber network‐embedded glass/epoxy composite structure was fabricated withdimensions of 300 (length) × 300 mm (width). As shown in Fig. 6(a), a total of 24 unidirectional glass/epoxy prepreg plies (UGN 160,Muhan Composite Co., Ltd) were stacked, and the SiC fiber yarns wereembedded at every interlaminar following pattern #3. In this study,commercially available SiC fibers (Tyranno S, Ube Industries, Ltd)

r network-embedded composite structure.

Fig. 7. Microscopic image of a cross-section of the fabricated SiC fibernetwork-embedded composite structure.

Fig. 8. Scanning electron microscope image of the interfacial bond betweenfibers and epoxy.

H. Kwon et al. Composite Structures xxx (2020) 113286

were used. The stacking sequence of the SiC fiber network‐embeddedglass/epoxy composite structure was ½0=SiC0=90=SiC90�6s. The SiC fiberyarns were embedded at each interlaminar of the glass/epoxy compos-ite structure in parallel directions of reinforcing fibers of glass/epoxylayers at a distance of 9 mm.

In this study, four SiC fiber yarns were selected as independent sen-sors among the embedded SiC fiber yarns. Two SiC fiber yarns embed-ded between the 22nd ply and 23rd ply and another two SiC fiberyarns embedded between the 23rd ply and 24th ply were selected assensors. These outer layers were selected as the SiC sensor positionsdue to their relatively large strains under plate bending. For the egressof the SiC fiber sensors, holes were made in the glass/epoxy prepregsat the end of the selected SiC fiber sensors and were filled with conduc-tive silver paste [40]. Fig. 6(b) shows the vacuum bagging of the lay-ered SiC fiber network‐embedded glass/epoxy prepregs. They were

Fig. 9. (a) Free-space measurement system and (b) measured reflectio

7

cured using an autoclave in the curing cycle, as shown in Fig. 6(c).Fig. 6(d) shows the fabricated SiC fiber network‐embedded compositestructure. The thickness of the cured specimen was 3.132 mm. Fig. 7shows a microscopic image of a cross‐section of the fabricated SiC fibernetwork‐embedded composite structure. In this figure, a resin‐richregion—a problem that might occur when different materials areembedded— was not observed because the SiC fibers had the samegeneric shape as the reinforcing fibers of the composites [36–38].The bottom panel of Fig. 7 shows a magnified image of the whiteboxed area of the top panel. In addition, Fig. 8 is a scanning electronmicroscope image showing the interfacial bond between fibers andepoxy. As shown in these figures, extra resin of the glass/epoxy pre-pregs penetrated well into the SiC fibers.

3.2. Microwave-absorption performance of the fabricated compositestructure

The microwave return loss of the fabricated SiC fiber network‐embedded composite structure was experimentally measured using afree‐space measurement system [41]. Fig. 9(a) illustrates the free‐space measurement system consisting of two X‐band horn antennasand a network analyzer (N5222A, Keysight Technologies Inc.). Coppertape was attached to the back of the plate to serve as a PEC material.Fig. 9(b) shows the measured microwave wave reflection loss of the

n loss of the fabricated SiC fiber-embedded composite structure.

Fig. 10. Experimental setup for impact testing of the SiC fiber network-embedded composite structure.

H. Kwon et al. Composite Structures xxx (2020) 113286

fabricated SiC fiber network‐embedded composite. The frequencyrange of more than 90% of the microwave energy absorption was from8.22–11.73 GHz, and the −10 dB bandwidth was 3.51 GHz. The min-imum return loss was−38.70 dB at 9.73 GHz. The resonant frequency(9.73 GHz) at which the minimum return loss occurs was shiftedslightly to the lower frequency region than the designed resonant fre-quency (9.87 GHz). As mentioned previously, the major microwave‐absorption principle of the Dallenbach absorber is the destructiveinterference of the waves reflected from the first interface (betweenthe air and the composite absorber) and the second interface (betweenthe absorber and the PEC). Thus, as the thickness of the absorberincreases, the minimum return loss occurs at a longer wavelength(lower frequency). Because the thickness of the fabricated SiC fibernetwork‐embedded composite (3.132 mm) was slightly thicker thanthe target thickness (3.12 mm), the minimum return loss might haveoccurred at a much longer wavelength (lower frequency) than the min-imum return loss frequency of the designed SiC fiber network‐embedded composite. However, the overall microwave reflection lossof the fabricated SiC fiber network‐embedded composite fitted wellwith that of the designed SiC fiber network‐embedded composite. Con-sequently, the fabricated SiC fiber network‐embedded compositestructure successfully demonstrated the predicted microwave‐absorption performance.

3.3. Dynamic signal acquisition using embedded SiC fiber sensors

Dynamic signal acquisition using the embedded SiC fiber sensorswas performed by impact testing on the fabricated composite struc-ture. Fig. 10 shows the experimental setup for the impact testing.The SiC fiber network‐embedded composite structure was fixed usinga steel fix zig. The resistance changes of the embedded SiC fiber sen-sors were measured using a Wheatstone bridge‐based electric circuit[40,42], the output of the circuit being the voltage. The embedded

8

SiC fiber sensors were connected to the circuit using wires throughan egress made during the fabrication process. To acquire referencesignals for comparative verification, an accelerometer (208C01, PCBPiezotronics Inc.) was attached to the specimen surface. The positionsof the embedded SiC fiber sensors and the accelerometer are shown inFig. 10. Impact excitation was applied using an impact hammer(086C03, PCB Piezotronics Inc.), and signals were acquired at a sam-pling rate of 10 kHz.

Fig. 11(a) shows the acquired impact signals in the time domain.The SiC fiber sensors #1 and #2, which were embedded relativelyfar from the neutral axis of the specimen, exhibit relatively high ampli-tudes. The signal noise levels varied for each SiC fiber sensor. This waspartly because the noise level was highly dependent on the alignmentstate of filaments in the yarn [43]. Therefore, for more consistent sen-sor performance, attention to the filament alignment is required.

Based on the temporal signals, the frequency response function(FRF) was estimated by applying a fast Fourier transform to both inputand response signals to obtain the natural frequencies of the SiC fibernetwork‐embedded composite specimen. The eight sets of frequency‐transformed signals were averaged for accurate natural frequency esti-mation. Fig. 11(b) shows the frequency response functions of the SiCfiber sensors and the accelerometer. The FRFs of the embedded SiCfiber sensors not only showed similar characteristics, but also exhib-ited characteristics comparable to the FRF of the accelerometer. Thenatural frequencies up to the 3rd mode were sufficiently distinguish-able. The natural frequencies measured using the accelerometer were242.3, 482.0, and 516.1 Hz for the 1st, 2nd, and 3rd modes, respec-tively. The natural frequencies measured using the embedded SiC fibersensors were 242.7, 476.2, and 514.5 Hz—these were in excellentagreement, within a 1.22% error—implying that the signals acquiredby the SiC sensors exhibited an appropriate structural response andthus demonstrated a sensing performance comparable to that of a com-mercial accelerometer.

Fig. 11. (a) Acquired impact signals from the embedded SiC fiber sensors and (b) frequency response functions of the embedded SiC fiber sensors andaccelerometer.

H. Kwon et al. Composite Structures xxx (2020) 113286

4. Conclusions

In this study, an embedded SiC fiber network‐based self‐sensingand microwave‐absorbing structure was proposed. The SiC fiber net-

9

work comprised SiC fiber yarns embedded at each interlaminar ofthe host composites at regular intervals. The SiC fiber yarns, whichdid not come into contact with other SiC fiber yarns, acted as indepen-dent sensing elements using their piezoresistivity. By means of the

H. Kwon et al. Composite Structures xxx (2020) 113286

strategic patterning of the SiC fiber yarns, the SiC fiber network‐embedded composites also acted as microwave lossy materials,thereby absorbing microwaves at a target frequency within the X‐band. To aid the design of the microwave‐absorbing functionality,full‐wave simulations were conducted. The simulation results verifiedthe effect of the SiC network design parameters, specifically in terms offiber yarn spacing and stacking patterns, on the complex permittivityof the composite. This, in turn, influenced the microwave loss charac-teristics of the SiC network‐embedded composite at a target frequency.The designed SiC fiber network‐embedded composite structure wasfabricated, and its self‐sensing and microwave‐absorbing perfor-mances were tested. The microwave return loss of the fabricated SiCfiber network‐embedded composite structure was measured using afree‐space measurement system. The fabricated SiC fiber network‐embedded composite structure covered most of the X‐band, in agree-ment with the designed microwave‐absorption performance. Dynamicsignal acquisition using the embedded SiC fiber sensors was also per-formed using an impact hammer test. Impact signals were successfullyacquired by the embedded SiC fiber sensors—they were comparable tothe signals acquired using a commercial accelerometer. The resultsrevealed that the SiC fiber sensors captured the structural responseaccurately. Therefore, this study successfully demonstrated themulti‐functional performance of the SiC network‐embedded compositefor both microwave‐absorbing and self‐sensing purposes. The SiCnetwork‐based composite structure proposed in this study has greatpotential for use in advanced aerospace structures that require simul-taneous self‐sensing and microwave‐absorbing capabilities.

CRediT authorship contribution statement

Hyunseok Kwon: Conceptualization, Methodology, Software,Investigation, Writing ‐ original draft. Min‐Su Jang: Methodology,Investigation. Jong‐Min Yun: Methodology, Investigation. YurimPark: Resources. Changkyo Shin: Formal analysis. Jinkyu Yang: Val-idation, Writing ‐ review & editing. Chun‐Gon Kim: Conceptualiza-tion, Supervision, Validation, Writing ‐ review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financialinterests or personal relationships that could have appeared to influ-ence the work reported in this paper.

Acknowledgments

This research was supported by the MSIT (Ministry of Science,ICT), Korea, under the High‐Potential Individuals Global Training Pro-gram (2019‐0‐01598) supervised by the IITP (Institute for Information& Communications Technology Planning & Evaluation). J. Y. has beensupported partly by Brain Pool Program through the National ResearchFoundation of Korea (NRF) funded by the Ministry of Science and ICT(2020H1D3A2A01061559).

References

[1] Gibson RF. A review of recent research on mechanics of multifunctional compositematerials and structures. Compos Struct 2010;92:2793–810. https://doi.org/10.1016/j.compstruct.2010.05.003.

[2] Obitayo W, Liu T. A review: Carbon nanotube-based piezoresistive strain sensors. JSensors 2012;2012. https://doi.org/10.1155/2012/652438.

[3] Chan KY, Jia B, Lin H, Hameed N, Lee JH, Lau KT. A critical review onmultifunctional composites as structural capacitors for energy storage.Compos Struct 2018;188:126–42. https://doi.org/10.1016/j.compstruct.2017.12.072.

[4] Yu J, Liu S, Duan G, Fang H, Hou H. Dense and thin coating of gel polymerelectrolyte on sulfur cathode toward high performance Li-sulfur battery. ComposCommun 2020;19:239–45. https://doi.org/10.1016/j.coco.2020.04.015.

10

[5] Liu S, Yang H, Sui L, Jiang S, Hou H. Self-adhesive polyimide (PI)@ReducedGraphene Oxide (RGO)/PI@Carbon Nanotube (CNT) hierarchically porouselectrodes: maximizing the utilization of electroactive materials for organic Li-Ion batteries. Energy Technol 2020;8:1–10. https://doi.org/10.1002/ente.202000397.

[6] Lightcap IV, Kamat PV. Graphitic design: Prospects of graphene-basednanocomposites for solar energy conversion, storage, and sensing. Acc Chem Res2013;46:2235–43. https://doi.org/10.1021/ar300248f.

[7] Yang H, Liu S, Cao L, Jiang S, Hou H. Superlithiation of non-conductive polyimidetoward high-performance lithium-ion batteries. J Mater Chem A2018;6:21216–24. https://doi.org/10.1039/C8TA05109G.

[8] Song YS, Youn JR. Influence of dispersion states of carbon nanotubes on physicalproperties of epoxy nanocomposites. Carbon N Y 2005;43:1378–85. https://doi.org/10.1016/j.carbon.2005.01.007.

[9] Huang X. Fabrication and properties of carbon fibers. Materials (Basel)2009;2:2369–403. https://doi.org/10.3390/ma2042369.

[10] Duan G, Fang H, Huang C, Jiang S, Hou H. Microstructures and mechanicalproperties of aligned electrospun carbon nanofibers from binary composites ofpolyacrylonitrile and polyamic acid. J Mater Sci 2018;53:15096–106. https://doi.org/10.1007/s10853-018-2700-y.

[11] Yu Y, Zhang B, Feng M, Qi G, Tian F, Feng Q, et al. Multifunctional structurallithium ion batteries based on carbon fiber reinforced plastic composites. ComposSci Technol 2017;147:62–70. https://doi.org/10.1016/j.compscitech.2017.04.031.

[12] Johannisson W, Ihrner N, Zenkert D, Johansson M, Carlstedt D, Asp LE, et al.Multifunctional performance of a carbon fiber UD lamina electrode for structuralbatteries. Compos Sci Technol 2018;168:81–7. https://doi.org/10.1016/j.compscitech.2018.08.044.

[13] Carlstedt D, Asp LE. Thermal and diffusion induced stresses in a structural batteryunder galvanostatic cycling. Compos Sci Technol 2019;179:69–78. https://doi.org/10.1016/j.compscitech.2019.04.024.

[14] Luo X, Chung DDL. Electromagnetic interference shielding using continuouscarbon-fiber carbon-matrix and polymer-matrix composites. Compos Part B Eng1999;30:227–31. https://doi.org/10.1016/S1359-8368(98)00065-1.

[15] Lee S-E, Park K-Y, Oh K-S, Kim C-G. The use of carbon/dielectric fiber wovenfabrics as filters for electromagnetic radiation. Carbon N Y 2009;47:1896–904.https://doi.org/10.1016/j.carbon.2009.02.013.

[16] Lee SE, Lee WJ, Oh KS, Kim CG. Broadband all fiber-reinforced composite radarabsorbing structure integrated by inductive frequency selective carbon fiber fabricand carbon-nanotube-loaded glass fabrics. Carbon N Y 2016;107:564–72. https://doi.org/10.1016/j.carbon.2016.06.005.

[17] Luo H, Chen W, Zhou W, Long L, Deng L, Xiao P, et al. Carbon fiber/Si3N4composites with SiC nanofiber interphase for enhanced microwave absorptionproperties. Ceram Int 2017;43:12328–32. https://doi.org/10.1016/j.ceramint.2017.06.096.

[18] Zhou W, Yin R ming, Long L, Luo H, Hu W da, Ding Y hong, et al. Enhanced high-temperature dielectric properties and microwave absorption of SiC nanofibersmodified Si3N4 ceramics within the gigahertz range. Ceram Int 2018;44:12301–7.https://doi.org/10.1016/j.ceramint.2018.04.017.

[19] Zhou W, Long L, Li Y. Mechanical and electromagnetic wave absorption propertiesof Cf-Si3N4 ceramics with PyC/SiC interphases. J Mater Sci Technol2019;35:2809–13. https://doi.org/10.1016/j.jmst.2019.07.002.

[20] Chung DDL. Materials for electromagnetic interference shielding. Mater ChemPhys 2020;255:. https://doi.org/10.1016/j.matchemphys.2020.123587123587.

[21] Wang X, Fu X, Chung DDL. Strain sensing using carbon fiber. J Mater Res1999;14:790–802. https://doi.org/10.1557/JMR.1999.0105.

[22] Horoschenkoff A, Mueller T, Kroell A. On the characterization of thepiezoresistivity of embedded carbon fibres. Proc 17th int conf compos materICCM-17, 2009.

[23] Chung DDL. A critical review of piezoresistivity and its application in electrical-resistance-based strain sensing. J Mater Sci 2020;55:15367–96. https://doi.org/10.1007/s10853-020-05099-z.

[24] Ishikawa T, Oda H. Structural control aiming for high-performance SiCpolycrystalline fiber. J Korean Ceram Soc 2016;53:615–21. https://doi.org/10.4191/kcers.2016.53.6.615.

[25] Yusof N, Ismail AF. Post spinning and pyrolysis processes of polyacrylonitrile(PAN)-based carbon fiber and activated carbon fiber: a review. J Anal ApplPyrolysis 2012;93:1–13. https://doi.org/10.1016/j.jaap.2011.10.001.

[26] Ding D, Zhou W, Zhang B, Luo F, Zhu D. Complex permittivity and microwaveabsorbing properties of SiC fiber woven fabrics. J Mater Sci 2011;46:2709–14.https://doi.org/10.1007/s10853-010-5140-x.

[27] Ye F, Zhang L, Yin X, Liu Y, Cheng L. Dielectric and electromagnetic waveabsorbing properties of two types of SiC fibres with different compositions. J MaterSci Technol 2013;29:55–8. https://doi.org/10.1016/j.jmst.2012.11.006.

[28] Liu H, Cheng H, Tian H. Design, preparation and microwave absorbing propertiesof resin matrix composites reinforced by SiC fibers with different electricalproperties. Mater Sci Eng B Solid-State Mater Adv Technol 2014;179:17–24.https://doi.org/10.1016/j.mseb.2013.09.019.

[29] Choi JH, Nam YW, Jang MS, Kim CG. Characteristics of silicon carbide fiber-reinforced composite for microwave absorbing structures. Compos Struct2018;202:290–5. https://doi.org/10.1016/j.compstruct.2018.01.081.

[30] Micheli D, Pastore R, Apollo C, Marchetti M, Gradoni G, Primiani VM, et al.Broadband electromagnetic absorbers using carbon nanostructure-basedcomposites. IEEE Trans Microw Theory Tech 2011;59:2633–46. https://doi.org/10.1109/TMTT.2011.2160198.

H. Kwon et al. Composite Structures xxx (2020) 113286

[31] Shin JH, Jang HK, Choi WH, Song TH, Kim CG, Lee WY. Design and verification ofa single slab RAS through mass production of glass/MWNT added epoxy compositeprepreg. J Appl Polym Sci 2015;132:42019. https://doi.org/10.1002/app.42019.

[32] Oh JH, Oh KS, Kim CG, Hong CS. Design of radar absorbing structures using glass/epoxy composite containing carbon black in X-band frequency ranges. ComposPart B Eng 2004;35:49–56. https://doi.org/10.1016/j.compositesb.2003.08.011.

[33] Shin J-H, Jang H-K, Choi W-H, Song T-H, Kim C-G, Lee W-Y. Numerical analysisof the complex permittivity of MWNT added epoxy depending on agglomerationsize. Compos Res 2014;27:190–5. https://doi.org/10.7234/composres.2014.27.5.190.

[34] Barlian AA, Park W-T, Mallon JR, Rastegar AJ, Pruitt BL. Review: semiconductorpiezoresistance for microsystems. Proc IEEE 2009;97:513–52. https://doi.org/10.1109/JPROC.2009.2013612.

[35] Kwon H, Park Y, Kim C-G. Strain sensing characteristics using piezoresistivity ofsemi-conductive silicon carbide fibers. Smart Mater Struct 2019;28:. https://doi.org/10.1088/1361-665x/ab3b2f105035.

[36] Shivakumar K, Emmanwori L. Mechanics of failure of composite laminates with anembedded fiber optic sensor. J Compos Mater 2004;38:669–80. https://doi.org/10.1177/0021998304042393.

11

[37] Jensen DW, Pascual J, August JA. Performance of graphite/bismaleimidelaminates with embedded optical fibers II: uniaxial compression. Smart MaterStruct 1992;1:31–5.

[38] Warkentin DJ, Crawley EF. Embedded electronics for intelligent structures. In:32nd Struct struct dyn mater conf. https://doi.org/10.2514/6.1991-1084.

[39] Vinoy KJ, Jha RM. Radar absorbing materials from theory to design andcharacterization. Boston: Kluwer Academic Publishers; 1996.

[40] Kwon H, Park Y, Shin C, Kim J-H, Kim C-G. Embedded silicon carbide fibersensor network based low-velocity impact localization of composite structures.Smart Mater Struct 2020;29:. https://doi.org/10.1088/1361-665x/ab7946055030.

[41] Knoff EF, Shaeffer JF, Tuley MT. Radar cross section: its prediction, measurementand reduction. Massachusetts: Artech House; 1985.

[42] Kwon H, Park Y, Shrestha P, Kim CG. Application of silicon carbide fibers as asensor for low-velocity impact detection and localization. Struct Heal Monit2019;18:1372–82. https://doi.org/10.1177/1475921718810398.

[43] Saifeldeen MA, Fouad N, Huang H, Wu ZS. Stabilization of electrical sensingproperties of carbon fiber sensors using pre-tensioning approach. Smart MaterStruct 2017;26. https://doi.org/10.1088/1361-665X/26/1/015012.