Effects of carbon nanotubes on electrical contact ...web.yonsei.ac.kr/mcmclab2/publication/triboint2016-1.pdf · Effects of carbon nanotubes on electrical contact resistance of a

Tribology International 104 (2016) 45–56

Contents lists available at ScienceDirect

Tribology International

http://d0301-67

n CorrE-m

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

Effects of carbon nanotubes on electrical contact resistance of aconductive Velcro system under low frequency vibration

Ilkwang Jang, Hyung Goun Joo, Yong Hoon Jang n

School of Mechanical Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, Republic of Korea

a r t i c l e i n f o

Article history:Received 7 May 2016Received in revised form16 August 2016Accepted 19 August 2016Available online 23 August 2016

Keywords:Conductive VelcroCarbon nanotubesElectrical contact resistanceLow frequency vibrationLubrication

x.doi.org/10.1016/j.triboint.2016.08.0199X/& 2016 Elsevier Ltd. All rights reserved.

esponding author.ail address: [email protected] (Y.H. Jang).

a b s t r a c t

Experimental measurements of electrical contact resistance are presented for a conductive Velcro con-nector with spayed-on carbon nanotubes under low frequency vibration. Several factors affecting theelectrical contact resistance of the Velcro connector, such as the type of carbon nanotube used, thedensity of the carbon nanotubes, the electrical load, the frequency, and the load amplitude of vibratingload, are analyzed. The role of carbon nanotubes in reducing electrical contact resistance is discussedfrom the view point of lubrication and damping.

& 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Carbon nanotubes (CNTs) have been acknowledged to enhancetraditional material performance because of their superior per-formance in strength, elastic modulus, and wear-resistance. Theirphysical characteristics such as chemical stability and electricaland thermal conductivity are also highly valued [1,2].

It has been reported that CNTs can be used to form metal/polymer composites, leading to the enhancement of mechanicalcharacteristics such as tensile strength and elastic modulus. Spe-cifically, a new structural composite material utilizes CNTs as re-inforcement for a polymer creating a light-weight, high strengthand super-elastic structural member. The member has been uti-lized for its bullet proof capabilities [3,4] and as an artificial muscle[5,6]. Furthermore, if CNTs are cast into a composite material, inwhich the original material was a non-conductor, the compositematerial is newly endowed with conductivity in addition to en-hanced mechanical properties [7,8]. The effects of CNTs are furtherrevealed in the fiber structure of CNTs [9–13].

A distinctive characteristic of CNTs is that their aspect ratio ismuch larger than that of any other material, leading to higherspecific surface area. This high specific surface area of CNTs en-hances the percolation for conductivity of composite materials. Forexample, the electrical percolation between the CNTs and the

polymer in composite material is a vital factor to establish theeffective electrical path between them [14–16]. The application ofCNTs to electrical connectors could solve a persistent problem inthe automotive industry known as fretting. In the electrical con-nectors of automobiles, the low-frequency and small amplitudevibrational environment can increase the possibility of frettingwear [17], establishing large electrical contact resistance, which –

in turn – can generate heat and oxidation [18,19]. These factorscontribute to the shortening of the connectors’ life and can induceunexpected short-circuits, possibly causing malfunction or explo-sions and loss of human life [20]. The adverse effects of fretting onelectrical contact resistance have long been observed. To overcomethe shortcomings of electrical connectors in terms of electricalcontact resistance, a continuous effort has been made to utilizeCNTs at the interface of the connector [21–23].

Velcro fasteners, traditionally known as fabric fasteners, haveexpanded their applicability to the entire area of engineeringjoining structures because of their easy fastening and detachableproperties [24–26]. Specifically, the application of conductiveVelcro fasteners in a dynamic loading environment such as thebattery assembly of an automotive vehicle is noticeable [27].However, the mechanical behavior of Velcro fasteners is complexbecause of the random alignment of hooks and loops. Moreover,their frictional behavior also shows non-compliance with theAmontons–Coulomb law; i.e., they exhibit nonlinear behavior offriction force against applied force, and fluctuations of the fric-tional forces in a dynamic motion, which requires more under-standing of the system [28]. Furthermore, the conductive Velcro

I. Jang et al. / Tribology International 104 (2016) 45–5646

fastener has an intrinsically large electrical contact resistance sincethe electrical conduction is dependent upon the actual micro-scopic contact spots between the fiber-like loops and hooks [29–32]. Recently, a study of contact resistance in the conductive Vel-cro system has shown that a certain range of low frequency at alow load amplitude causes higher electrical contact resistance andthe wear behavior after a long exposure to dynamic loading undera specific frequency at a low load amplitude has been confirmed[33].

Motivated by the characteristics of CNTs and the issues ofconductive Velcro, we consider that the application of CNT to theinterface of the Velcro connection may be a possible way to en-hance the performance of conductive Velcro, specifically reducingthe electrical contact resistance of the fasteners under a low-fre-quency vibrating environment. For this reason, the electricalcontact resistance according to electrical loads, frequencies andamplitudes of vibration for several types of CNTs with differentaspect ratios is investigated. As a result of the analysis, tribologicalevidences affecting the electrical contact behavior of the systemwill be explained.

2. Experimental details

2.1. CNT and conductive Velcro

Carbon nanotubes (CNTs), as an additive material to conductiveVelcro systems to diminish the electrical contact resistance, areprepared by considering the ranges of the aspect ratio of the CNTs.The selection of this range of aspect ratio could be critical becausethe contact geometry at the fiber and hook interface is affected bythe CNT size and length. Table 1 shows several CNTs, includingsingle-wall and multi-wall CNT (SWNTs and MWNTs, respectively)with various aspect ratios. Typical SEM and TEM images forMWNTs are shown in Fig. 1.

The conductive Velcro system as an experimental specimenconsists of a pair of silver-coated nylon fibers and hooks as shownin Fig. 2. The Velcro system was manufactured by APLIX, and hasdimensions of 20 mm�20 mm.

2.2. Methods of CNT suspension on the Velcro

A CNT solution is obtained by suspending 5 g of CNTs in 500 mLof acetone using ultrasonicated dispersion. The ultrasonic disper-sion is carried out for two hours, followed by a thirty minute in-termission to relieve heat; this process is repeated five times, re-sulting in a total dispersion time of 10 h. Two methods of CNTapplication to conductive Velcro as shown in Fig. 3 are employed:Either the CNT suspension is sprayed onto the surfaces of thehooks and loops, or else the hook-and-loop layer of the Velcro areimmersed into the CNT suspension, respectively. The deposition ofCNT is carried out for 10 s in each case. All the samples are driedbefore continuing with the experiments.

Table 1Several types of CNTs used in the experiment.

No. Type Diameter (nm) Length (㎛) Purity (%) Aspect ratio

1 SWNT 1.4–2 5–20 90 2500–14,2852 MWNT 20 1–25 90 50–12503 MWNT 4–6 10–20 90 1666–50004 MWNT 10–20 10–20 95 500–2000

2.3. Experimental apparatus and variables for electrical contact re-sistance measurement under low frequency vibration

The experimental apparatus to measure the electrical contactresistance under fretting vibration is shown in Fig. 4. Since theapparatus was already introduced when the electrical contact re-sistance of the conductive Velcro without CNT was measured, webriefly summarize the experimental detail [33]. The low frequencyvibration is generated by the motor and the connected cam. Thefrequency and the amplitude of the vibratory motion can becontrolled up to 23 Hz and 100 mm, respectively. The Velcro spe-cimens with the loops and hooks contact each other, and the CNTsare located at the interface of the two components. The apparatushas natural frequencies of 29.75 Hz and 59.50 Hz; consequently,the resonance of the apparatus does not affect the low frequencyvibration at frequencies of 23 Hz or less.

The experimental variables are the frequency and amplitude ofthe vibration motion, and the electrical current. The frequencyrange is selected by considering the motion of transportation andthe failure environment of the electrical connector due to fretting[34]. The electrical load is chosen according to the battery capacityof automotive components [35]. Specifically, the amplitude rangesare selected as 710 mm,725 mm,750 mm,7100 mm with amaximum frequency of 23 Hz; an electrical current of 0.1A isemployed; and the variation of current is also measured accordingto the electrical voltage. The fretting motion is measured for up to3600 cycles.

2.4. Modeling of electrical contact resistance in the conductive Velcrosystem

Since the electrical conduction in the Velcro system occurs at alarge number of microscopic contact spots between the fiber-likeloops and hooks, the electrical contact resistance is difficult toanalyze. Thus, the lumped-element circuit analysis using Kirch-hoff's law and Ohm's law can give a useful solution [33]. A sum-mary of the analysis model is presented here: A model of theVelcro system to estimate the electrical contact resistance be-tween Velcro plates is shown in Fig. 5. The resistivities of theupper and lower plates are assumed to be ρ1 and ρ2, respectively.The electrical contact resistance per unit length between the fibersand hooks is Rc. The voltages on the upper and lower plates, ( )V x1and ( )V x2 , vary along the length L of the plate. The current densityper unit length generated by the voltage drop between the twoplates is ( )=( ( ) − ( ))i x V x V x R/ c1 2 , where the current density and the

total current are related by ( )∫ ξ ξ( ) =I x i dx

0. The potential drop in

the lower plate can be described as = − ρ∂ ( )∂

( )V xx

I x

A2 2

2where A2 is the

cross-sectional area of the lower plate. If the potential dropequation is differentiated and combined with the current equa-tion, then the governing equation for the electrical potential of thelower plate and the upper plate can be expressed as

ρ ρ∂ ( )∂

=( ) − ( ) ∂ ( )

∂= −

( ) − ( )( )

V x

x AV x V x

RV x

x AV x V x

R,

1c c

22

22

2

1 22

12

1

1

2 1

where A1 is the cross-sectional area of the upper plate. Theboundary conditions for the two governing equations are

∂∂

( ) =∂∂

( ) = ( ) − ( ) = Δ ( )Vx

Vx

V L V L V0 0, 0 0, 21 2

1 2

Four boundary conditions are needed in order to solve the twodifferential Eq. (1), but the equations give a solutionwith one arbitraryunknown constant if we use three conditions (2). The arbitrary un-known constant can be determined if we choose a datum potential atthe end of the plate, =x L. Thus the resulting solutions for voltage are

Fig. 1. Multi-wall carbon nanotubes (a) SEM image (b) TEM image.

Fig. 2. Conductive Velcro system consisting of (a) hook (b) loop.

Fig. 3. CNT deposition through (a) spray (b) immersion.

I. Jang et al. / Tribology International 104 (2016) 45–56 47

Fig. 4. (a) Apparatus for fretting experiment; (b) Schematic of Velcro interfaces with CNTs.

Fig. 5. Model for electrical contact resistance.

I. Jang et al. / Tribology International 104 (2016) 45–5648

( ) − = Δ+

( − + ( + ))

( + )

( ) − = − Δ+

( + ( + ))

( + ) ( )

V x C Vk

k k

k k x

k k L

V x C Vk

k k

k k k k x

k k L

1 cosh

cosh,

cosh

cosh 3

11

1 2

1 2

1 2

21

1 2

1 2 1 2

1 2

Fig. 6. SEM images of the hook surface; (a) original surface (b) by the immersion medispersion.

where ρ= ( )k A R/ c1 1 1 , ρ= ( )k A R/ c2 2 2 , and C is an arbitrary voltage. Theconstants ρ A/1 1and ρ A/2 2 are measured as 3.2Ω/m, 2.1Ω/m,respectively.

3. Results and discussion

The characteristics of the electrical contact resistance for con-ductive Velcro fasteners under low frequency vibrational loadingare reported and discussed here. Firstly, CNT formation on con-ductive Velcro interfaces according to the CNT deposition methodsis investigated. Using the outcomes from the proposed depositionmethods, the most appropriate method is suggested. Secondly,electrical contact resistances of conductive Velcro systems withdifferent CNT aspect ratios are determined according to the ex-perimental input variables such as frequency, amplitude of motion,and electrical loads. Several effects of CNTs, such as lubrication anddamping, which affect the conductive Velcro system under lowfrequency vibrating loads are discussed.

thod without ultrasonic dispersion and (c) by immersion method with ultrasonic

Fig. 7. A high-resolution SEM photograph of the hook surface by the spray method in the case of ultrasonically dispersed CNT.

I. Jang et al. / Tribology International 104 (2016) 45–56 49

3.1. CNT formation on conductive Velcro interface according to CNTdeposition

It has been recognized that the deposition method may affectthe functionality of CNTs [36]. The CNT formation on the surfacesof hooks and loops after dispersion can influence the contactdistribution of Velcro connections, leading to variations in elec-trical contact resistance. Thus we present a detailed description ofthe deposition. For the deposition methods, we use spraying orimmersion methods. Fig. 6 shows SEM pictures of the hook surfaceby the immersion method with and without ultrasonic dispersion,compared to the original surface before deposition. When the ul-trasonic dispersion step is neglected, CNTs are clustered in a largemass. After deposition, because of the flow of the solvent andgravity, CNTs drip to the base section of the hook part of theVelcro. Note that the CNTs are no longer attached to the hooksrendering the Velcro useless for experimentation. On the otherhand, the ultrasonic dispersion causes the CNTs to form into smalllumps, which then become well attached to the surface of thehook. A detailed configuration of CNT formation in the hook sur-face by the spray method can be seen in Fig. 7. A higher-magni-fication SEM image shows that most CNTs are well attached to thesurface of the hook while CNT clusters are formed at some pro-tuberance areas. It can also be seen that the CNTs are well attachedat the surface of the loop fibers, as shown in Fig. 8.

In order to validate the effect of the CNT deposition methods,spray and immersion, we measure the electrical contact resistanceunder a vibrating loading condition of 10Hz frequency and

710mm amplitude. Note that after the immersion, the hook-and-loop samples are dried by turning over the specimen, which in-cludes both hooks and loops, over. The evolution of electricalcontact resistance for each deposition method is presented inFig. 9. The addition of CNTs by either one of the two depositionmethods commonly reduces the overall electrical contact re-sistance, compared to the case of no CNT additions. The reasonmay be that the CNTs play a role to increase contact area, en-hancing the conduction. As contacting indenters, the fibers mi-croscopically make very small contact area because of the smallcurvature of the surface. The CNTs near the edge of contact areacan create additional pathway for electrical conduction, resultingin an extended electrical contact area. This explanation can befurther understood if we determine the type or aspect ratio ofCNTs that affect electrical contact resistance, which may con-tribute to the formation of the shape of the passage at the edge ofthe contact. On the other hand, the variation in the range of theelectrical contact resistance from spray deposition is less than thevariation from the immersion method. The immersion depositioncauses abrupt stepwise fluctuations in electrical contact resistanceand exhibits a slowly increasing trend. However, the spray de-position method mostly shows stable behavior in the variation ofelectrical contact resistance after comparatively small fluctuationsduring the initial stage. It is conjectured that these behaviors arerelated to the distribution of CNTs on the surface of the Velcro. Asshown in Fig. 6–8, the spray deposition distributes the CNTs moreuniformly on the surface of the Velcro than does the immersionmethod. The CNT distribution due to the spray deposition can

Fig. 8. SEM images of a loop surface by the spray method in the case of ultrasonically dispersed CNT.

Fig. 9. Evolution of electrical contact resistance according to deposition methods of (a) spray dispersion and (b) immersion.

I. Jang et al. / Tribology International 104 (2016) 45–5650

serve well as a solid lubrication film, which provides a stablecontact area along the sliding region, leading to a reduction ofwear under long-term vibrational loading. All the issues that havebeen raised at this point will be examined thoroughly in the fol-lowing sections.

The most effective application method for CNT deposition inthis experiment is chosen as the spray deposition of the ultra-sonically dispersed CNT suspension. In order to find the optimumnumber of injections, the density of CNTs is analyzed as a functionof the number of spray injection. The density of CNT by spray

Fig. 10. Distribution of CNTs on the hook surface in a conductive Velcro system according to the number of spray injection; (a) 1 spray (b) 2 sprays (c) 3 sprays.

Table 2Density of CNTs and electrical contact resistance of conductive Velcro system ac-cording to the number of spray injection.

Number of CNT injections Density of CNTs (%) Electrical contact resistance(Ω)

1 3.79 0.712 15.82 0.283 28.50 0.114 38.21 0.11

Fig. 11. Electrical contact resistance for a Velcro system with various types of CNTas a function of frequency, with an amplitude of 710 mm and an electrical currentof 0.1 A.

Fig. 12. Dynamic characteristics of a conductive Velcro system with SWNTs andMWNTs.

I. Jang et al. / Tribology International 104 (2016) 45–56 51

injection, defined as the ratio of the CNT-coated regions to thetotal area, is calculated by the image processing of a photo wherethe CNTs are deposited onto white paper. The image is convertedinto a 255-level grayscale, identifying the black regions in theupper half of the gray-scale as CNT objects. Fig. 10 shows opticalimages (upper row) of the CNT density on white paper accordingto the number of injection and SEM images (lower row) of thecorresponding conductive Velcro surface. Each Velcro sample istested under the vibrating loading condition of 10 Hz frequencyand 710 mm amplitude. Table 2 shows the number of CNT injec-tions, the density of CNTs, and the corresponding electrical contactresistance. These results show that the density of CNT increases asthe number of injections increases but the electrical contact

Fig. 13. Variation of electrical contact resistance with frequency for the Velcrosystem without [33] and with CNTs for several values of the vibrational amplitude,and at an electrical current of 0.1A. The shaded region in (a) represents the fluc-tuation range of contact resistance without using CNTs. The contact resistances inthe samples containing CNTs are enlarged in (b).

I. Jang et al. / Tribology International 104 (2016) 45–5652

resistance in the conductive Velcro sample approaches 0.11Ω,once the density of approximately 30% has been reached.

The guidelines mentioned above for the spray deposition ofCNT and the number of injections are followed to prepare theexperimental samples. Specifically, the number of injections ischosen, based on the criterion that the CNT density should equalor exceed 30%, because the number of injections may vary ac-cording to the spray device, method, and experimentenvironments.

Fig. 14. CNTs on Velcro fiber surface (a) CNT distribution in a region where thesilver has been removed (b) formation of lubrication film (c) CNT carpet moved byfretting action.

3.2. Effect of CNT types and aspect ratios on electrical contactresistance

The different types of CNTs, such as SWNTs or MWNTs, in-trinsically have different electrical resistivities, and the CNT layerwith different aspect ratios can change the stiffness of the interfaciallayer. These characteristics of the CNTs may affect the electrical

Fig. 15. SEM images of the hooks and EDX mapping for carbon on the hooks for (a) 5 Hz; (b) 10 Hz; (c) 15 Hz. (For interpretation of the references to color in this figure, thereader is referred to the web version of this article.)

I. Jang et al. / Tribology International 104 (2016) 45–56 53

contact resistance under low-frequency vibrating loads. Thus, wemeasure the electrical contact resistance under low-frequency vi-brating conditions according to the type and aspect ratio of the CNTs.Four types of CNTs are shown in Table 1, having various aspect ratiosranging from 50 to around 15,000. The vibrating loads, at frequenciesof 2, 5, 10, 15, and 20 Hz with an amplitude of 710 mm, are appliedduring 3600 cycles of fretting motion. An electrical current of 0.1 A isalso applied. Fig. 11 shows the electrical contact resistance as afunction of frequency for several types of CNTs.

The distinctive features of the results show that the response ofelectrical contact resistance for the Velcro with SWNTs variessignificantly according to frequency, reaching its maximum valueat frequencies of 5 and 10 Hz, while the Velcro with MWNTs hasnearly constant electrical contact resistance within the frequencyrange from 2 to 20 Hz, having a smaller value than the maximumvalue of the SWNT case. This difference is primarily caused by thetype of CNT, which have relatively different electrical resistivity.The SWNTs in this experiment, however, have a wider range ofaspect ratio up to 15,000, which is than that of the MWNTs. Sincethe structure of CNTs with large aspect ratio is vulnerable to thevibrating condition, the vibrational characteristics of the systemswith SWNTs and MWCNTs are estimated by using an LMS tester

with an accelerometer. The acceleration amplitude versus fre-quency is obtained as shown in Fig. 12, indicating that the vibra-tion amplitude of the Velcro connector with SWNTs is significantlylarge between a frequency of 5 and 10 Hz, while the vibrationamplitude of the Velcro connector with MWNTs lacks a noticeablevariation. This acceleration amplitude response with frequencyparallels the behavior of electrical contact resistance for the Velcrosystem. For the Velcro system with the SWNTs at a certain re-sonant frequency, the fibers of the loops in the Velcro structureendure greater movement, resulting in considerable change anddecrease of contact area, and corresponding increases in theelectrical contact resistance. It is speculated that the resonantfrequency within the low frequency vibration range is the struc-tural frequency of the conductive Velcro connector, not of the in-dividual CNTs. Note that CNTs are known to possess a spectrum ofresonant frequencies ranging from several kHz to GHz dependingtheir aspect ratios [37]. Thus, this dynamic characteristic of in-dividual CNTs is not revealed in this low frequency range. For theVelcro system with the MWNTs, the damping effects are sig-nificant, showing that the acceleration amplitude between5 and10 Hz is finite and has a gentle slope. It is more evident that ifwe compare the acceleration amplitude of Velcro with and

Fig. 16. SEM images of the loops and EDX mapping for carbon on the loop for (a) 5 Hz; (b) 10 Hz; (c) 15 Hz. (For interpretation of the references to color in this figure, thereader is referred to the web version of this article.)

Table 3Damping constant and stiffness for Velcro with MWNTs and without CNTs.

C (N/m/s) k (N/m)

Velcro 0.002 0.403VelcroþMWNT 0.021 1.708

I. Jang et al. / Tribology International 104 (2016) 45–5654

without CNTs, the magnitude of amplitude for the Velcro systemwith SWNTs is reduced by a factor of ten, as shown in Fig. 5 in [33].A detailed analysis for the system with MWNTs is presented in thenext section.

3.3. Electrical contact resistance for the Velcro system with andwithout MWNTs

The electrical contact resistance of the Velcro system withMWNTs is independent of frequency, especially in the low fre-quency range less than 20 Hz, as stated in the previous section. Inorder to highlight the outcome, the electrical contact resistance of

the Velcro system with CNTs is compared to the case withoutCNTs. The results for the response of electrical contact resistancewithout CNTs have already been published [33]. We select theMWNT denoted as No. 2 in Table 1. An electrical current of 0.1 Aand a fixed vibrational amplitude are applied during 3600 loadingcycles.

Fig. 13 shows plots of the electrical contact resistance as afunction of frequency for the conductive Velcro with and withoutCNTs. For the case without CNTs, the responses of the electricalcontact resistance vary significantly in the frequency range of 7–10 Hz, particularly under the load amplitude of 710 mm and725 mm, compared with the case of higher amplitudes of750 mm and 7100 mm. The electrical contact resistance withCNTs, however, is almost constant regardless of frequency and loadamplitude, maintaining values between 0.04Ω and 0.06Ω, whichare much lower than those in the case without CNTs. The relativereduction is almost 97% for the maximum electrical contact re-sistance at 10 Hz. As shown in Fig. 13(a), the Velcro fiber interac-tion that determines electrical contact resistance is affected by theresonant frequency and the load amplitude if CNTs are not

Fig. 17. Current–voltage curves for the conductive Velcro without CNTs [33](shown as a shaded region) and with CNTs (shown as red marks) at vibrationalfrequencies of 5, 10 and 15 Hz. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)

I. Jang et al. / Tribology International 104 (2016) 45–56 55

involved. However, the Velcro systems with CNTs are not affectedby frequency or load amplitude. Thus, it may be conjectured thatsome properties of the CNTs may constrain the movement of fibersat the resonant frequency and at certain load amplitudes. In thefollowing paragraphs, we introduce several mechanisms that mayaffect the behavior of electrical contact resistance in the con-ductive Velcro connector under low frequency vibration.

After long exposure to vibrational loading, wear at the contactarea is inevitable, leading to an increase of the electrical contactresistance. In conductive Velcro fibers, the silver-coating layer isvulnerable to wear and can be easily removed. If CNTs are added tothe silver-coated fibers, the CNTs can form electrical pathway evenafter the removal of silver has occurred. Evidence of this is dis-played in Fig. 14(a), which shows the CNT distribution on the re-gion of the silver coating layer where the silver has been removedduring a fretting experiment. It is probable that the CNTs havemoved into the silver-depleted region and adhere to the nylonregion. Thus, the CNTs act as an additive to create electricalpathway and enlarge the contact area, which reduces the electricalcontact resistance.

Besides their role in reducing the electrical contact resistance inthe conductive Velcro connector, another capability of CNTs islubrication that leads to the electrical contact resistance beingmaintained constant without fluctuation. Fig. 14(b) shows that theCNTs are smeared, formed as agglomerated patches, and mostlyfound at the edge of wear patches. Fig. 14(c) also shows that theCNT carpet moves in the sliding direction by fretting action. Si-milar features have been reported in other studies [23,38],showing that the CNT films established by wear reduce the coef-ficient of friction and the frictional fluctuation. This evidencesuggests that the CNTs provide solid lubrication, since the varia-tion in the range of the electrical contact resistance is dramaticallyreduced. In this specific circumstance, the CNTs shed by wearmotion can transfer to the silver-depleted region, in addition tothe establishment of a solid lubrication layer.

The fretting wear in the Velcro connector can damage the lu-brication films of CNTs, possibly leading to elimination of the CNTsfrom the surface. In order to identify the loss of CNTs after longexposure to vibrating load, a load amplitude of 7100 mm at fre-quencies of 5, 10, and 15 Hz is applied to a hook-and-loop as-sembly for 24 h, and then the surfaces are checked using SEM

images and EDX mapping as shown in Figs. 15 and 16. Each figurerepresents the best way to verify the removal of CNTs. The SEMimages at different frequencies show a similar worn surface butthe EDX mapping shows that there is no difference of residualcarbon between them, meaning that CNTs are not eliminated onthe Velcro surface under the low frequency vibrating load and canmaintain a useful role as lubricants during a considerable serviceperiod. Note that the EDX mapping displays carbon as yellow incolor.

As pointed out in the previous section, the damping effects ofCNTs are important. A comparison is performed to identify thedamping effect for the Velcro with and without CNTs; the dampingconstant and stiffness are calculated by estimating the dampingratio and frequency of the damped vibrations, as shown in Table 3.The damping constant and the stiffness of the Velcro with CNTsare 10 and 4 times larger than those of the Velcro without CNTs.These vibrational characteristics of CNTs in the Velcro connectorshow that the deformation of fibers in Velcro under the low fre-quency vibration load is significantly reduced and surface damageat the contact region is decreased.

3.4. Effect of electrical load on electrical contact resistance

To design a connector, the basic relationship between the cur-rent through the Velcro connector with CNTs and the voltageacross its terminals must be characterized by a current–voltage (I–V) curve. Fig. 17 shows the I–V plots for the conductive Velcro withand without MWNTs at vibrational frequencies of 5, 10, and 15 Hz.The slope of the I–V curve represents the conductance of theconductive Velcro system. The I–V curve for the Velcro withoutCNTs was previously reported in [33]. It can be summarized bysaying that a linear I–V relation occurs at the frequencies of 5 and15 Hz, but for 10 Hz, the slope at low voltage is different from theslope at high voltage. It is believed that in the low voltage range, alarger movement of Velcro occurs at the frequency of 10 Hz andthis peculiar vibration combined with the low current acceleratesthe rate of oxidation, leading to an increase of the electrical con-tact resistance, but for the high voltage range, the large electricalheating enlarges the contact area because of the softening of thematerial.

Compared to the I–V characteristic for the conductive Velcrowithout CNTs, the I–V relation for the Velcro with CNTs follows alinear trend, regardless of frequency. The slope of the I–V curve issix times as large as that of the Velcro without CNTs, meaning thatthe conductance is also six times as high. When we investigatedthe behavior of electrical contact resistance for the Velcro with theMWNTs (Fig. 13(a)), the response was not frequency-dependent;and this observation also applies to the I–V characteristics. In ad-dition, the occurrence of oxidation in the low current state couldbe hampered by the existence of CNTs. This conjecture is sup-ported by the experimental evidence given in [39]. For the highcurrent state, heat due to Joule heating could be lessened by theCNT lubrication characteristic, leading to no significant change incontact area.

Like CNT, graphene is a promising material to increase electricalconductivity greatly by the addition to the surface of a material[40], which is anticipated to reduce the electrical contact re-sistance under fretting. The questions regarding the electricalcontact resistance raised in this analysis can be applied to gra-phene material and are the subject of ongoing research.

4. Conclusion

The effect of CNTs on the electrical contact resistance of con-ductive Velcro connectors, especially under low frequency

I. Jang et al. / Tribology International 104 (2016) 45–5656

vibration was experimentally investigated. Among the varioustypes of CNTs, the conductive Velcro connector with the MWNTshaving a small aspect ratio maintained a constant electrical con-tact resistance regardless of frequency and load amplitude, andhad a considerably smaller electrical contact resistance than theVelcro without CNTs. A nearly linear I–V curve, independent offrequency, was obtained. The behavior of the electrical contactresistance parallels the dynamic response of the Velcro connector.Supporting evidence for the behavior of the electrical contact re-sistance was inferred from the characteristics of the CNT-coatedsamples, such as solid lubrication and the damping coefficient. Theapplication of CNTs using the spray deposition of ultrasonicallydispersed CNT suspension becomes effective when the areal den-sity of CNTs is greater than 30% of the total deposition area. Ex-perimental results show that CNTs are effective as an additivesubstance to reduce the electrical contact resistance under a low-frequency vibrating load condition, indicating that they can beapplied to the design of electrical connectors composed of fiberstructures.

References

[1] Sugano K, Kurata M, Kawada H. Evaluation of mechanical properties of un-twisted carbon nanotube yarn for application to composite materials. Carbon2014;78:356–65.

[2] Nam TH, Goto K, Yamaguchi Y, Premalal EVA, Shimamura Y, Inoue Y, et al.Effects of CNT diameter on mechanical properties of aligned CNT sheets andcomposites. Compos A: Appl Sci Manuf 2015;76:289–98.

[3] Jagtap P, Kumar A, Kumar P. Effect of electric field on creep and stress-re-laxation behavior of carbon nanotube forests. RSC Adv 2016;6:67685–92.

[4] Aldajah S, Haik Y, Moustafa K, Alomari A. High and low speed impact char-acteristics of nanocomposites. Adv Mater Res 2015;1105:62–6.

[5] Fraczek-Szczypta A, Menaszek E, Blazewicz S, Adu J, Shevchenko R, Syeda TB,et al. Influence of different types of carbon nanotubes on muscle cell re-sponse. Mat Sci Engin C 2015;46:218–25.

[6] Lima MD, Hussain MW, Spinks GM, Naficy S, Hagenasr D, Bykova JS, et al.Efficient, absorption-powered artificial muscles based on carbon nanotubehybrid yarns. Small 2015;11:3113–8.

[7] Liu ZY, Xiao BL, Wang WG, Ma ZY. Tensile strength and electrical conductivityof carbon nanotube reinforced aluminum matrix composites fabricated bypowder metallurgy combined with friction stir processing. J Mater Sci Technol2014;30:649–55.

[8] Jang J-S, Varischetti J, Suhr J. Strain dependent energy dissipation in multi-scale carbon fiber composites containing carbon nanofibers. Carbon2014;50:4277–283..

[9] Abot JL, Alosh T, Belay K. Strain dependence of electrical resistance in carbonnanotube yarns. Carbon 2014;70:95–102.

[10] Guignier C, Bueno MA, Camillieri B, Le Huu T, Oulanti H, Durand B. Tribologicalbehaviour and adhesion of carbon nanotubes grafted on carbon fibres. TribolInt 2015;100:104–15.

[11] Li W, Xu F, Wang Z, Wu J, Liu W, Qiu Y. Effect of thermal treatments onstructures and mechanical properties of aerogel-spun carbon nanotube fibers.Mater Lett 2016;183:117–21.

[12] Yu J, Wang L, Lai X, Pei S, Zhuang Z, Meng L, et al. A durability study of carbonnanotube fiber based stretchable electronic devices under cyclic deformation.Carbon 2015;94:352–61.

[13] Hatipoglu G, Kartal M, Uysal M, Cetinkaya T, Akbulut H. The effect of slidingspeed on the wear behavior of pulse electro Co-deposited Ni/MWCNT nano-composite coatings. Tribol Int 2016;98:59–73.

[14] Ata S, Yoon H, Subramaniam C, Mizuno T, Nishizawa A, Hata K. Scalable,

solvent-less de-bundling of single-wall carbon nanotube into elastomers forhigh conductive functionality. Polymer 2014;55:5276–83.

[15] Thostenson ET, Ren ZF, Chou TW. Advances in the science and technology ofcarbon nanotubes and their composites: a review. Compos Sci Technol2001;61:1899–912.

[16] Bauhofer W, Kovacs JZ. A review and analysis of electrical percolation in car-bon nanotube polymer composites. Compos Sci Technol 2009;69:1486–98.

[17] Park YW, Sankara Narayanan TSN, Lee KY. Effect of fretting amplitude andfrequency on the fretting corrosion behaviour of tin plated contacts. Surf CoatTechnol 2006;201:2181–92.

[18] Fu R, Chae SY, Choe, Jackson RL, Flowers FT, Kim D. Modeling and analysis ofvibration-induced changes in connector resistance of high power electricalconnectors for hybrid vehicles. Mech Based Design Struct Machines2012;40:349–65.

[19] Xie F, Flowers GT, Chen C, Bozack M, Suhling J, Rickett BI. Analysis and pre-diction of vibration-induced fretting motion in a blade/receptacle connectorpair. IEEE Trans Comp Pack Technol 2009;32:585–92.

[20] Bouzera A, Carvou E, Abdi RE, Benjemâa N, Tristani L, Zindine EM. Study ofminimum fretting for connectors used in automotive applications. Electr Eng2013;95:201–8.

[21] Guoxin X, Dan G, Jianbin L. Lubrication under charged condition. Tribol Int2015;84:22–35.

[22] Lee J, Kwon H, Seo J, Shin S, Koo JH, Pang C, et al. Conductive fiber-basedultrasensitive textile pressure sensor for wearable electronics. Adv Mater2015;27:2433–9.

[23] Kaushik V, Lee J, Hong J, Lee S, Lee S, Seo S, et al. Textile-based electroniccomponents for energy applications: principles, problems, and perspective.Nanomatr 2015;5:1493–531.

[24] Damstra RJ, Partsch H. Prospective, randomized, controlled trial comparingthe effectiveness of adjustable compression Velcro wraps versus inelasticmulticomponent compression bandages in the initial treatment of leg lym-phedema. J Vasc Surg 2013;1:13–9.

[25] Avila AG, Hinestroza JP. Tough cotton. Nat Nanotechnol 2008;3:458–9.[26] Panhuis M, Wu J, Asharf AS, Wallace GG. Conducting textiles from single-

walled carbon nanotubes. Synth Met 2007;157:358–62.[27] Stanforda T, Barvosa-Cartera W. Active material reversible attachments: shape

memory polymer based. Proc SPIE 2008;6930 693003-1.[28] Mariani LM, Esposito CM, Angiolillo PJ. Observations of stick-slip friction in

Velcro. Tribol Lett 2014;56:189–96.[29] Jang YH, Barber JR. Effect of contact statistics on electrical contact resistance. J

Appl Phys 2003;94:7215–21.[30] Lee S, Jang YH, Kim W. Effects of nanosized contact spots on thermal contact

resistance. J Appl Phys 2008;103(74308):1–8.[31] Lee S, Cho H, Jang YH. Multiscale electrical contact resistance in clustered

contact distribution. J Phys D: Appl Phys 2009;42(165302):1–7.[32] Yoo S, Jang YH. Multiscale electrical contact resistance between gas diffusion

layer and bipolar plate in proton exchange membrane fuel cells. J Fuel Cell SciTechnol 2012;9(031003):1–7.

[33] Joo HG, Jang YH, Choi HS. Electrical contact resistance for a conductive Velcrosystem. Tribol Int 2014;80:115–21.

[34] Yang H, Flowers G. Fretting in electrical connectors induced by axial vibration.IEEE Trans Compon Packag Manuf Technol 2015;5:328–36.

[35] Köhler U. Applications—transportation/hybrid electric vehicles: batteries. In:Garche J, editor. Encyclopedia of electrochemical power sources. Amsterdam:Elsevier; 2009. p. 269–85.

[36] Rider AN, An Q, Brack N, Thostenson ET. A comparison of mechanical andelectrical properties in hierarchical composites prepared using electrophoreticor chemical vapor deposition of carbon nanotubes. MRS Adv 2016;1:785–90.

[37] Yoon J, Ru CQ, Mioduchowski A. Terahertz vibration of short carbon nanotubesmodeled as Timoshenko beams. J Appl Mech 2005;72:1017.

[38] Feng Y, Zhang M, Xu Y. Effect of the electric current on the friction and wearproperties of the CNT–Ag–G composites. Carbon 2005;43:2685–92.

[39] Qiufeng W, Hongling W, Yunxia W, Fengyuan Y. The influences of severalcarbon additions on the fretting wear behaviors of UHMWPE composites.Tribol Int 2016;93:390–8.

[40] Guoxin X, Mattias F, Jinshan P. Direct electrochemical synthesis of reducedgraphene oxide (rGO)/copper composite films and their electrical/electro-active properties. ACS Appl Mater Interfaces 2014;6:7444–55.