Accepted Manuscript Designing, modeling and manufacturing of lightweight carbon nanotubes/polymer composite nanofibers for electromagnetic interference shielding application Komeil Nasouri, Ahmad Mousavi Shoushtari PII:
S02663538(16)319492
DOI:
10.1016/j.compscitech.2017.03.041
Reference:
CSTE 6726
To appear in:
Composites Science and Technology
Received Date: 11 December 2016 Revised Date:
28 March 2017
Accepted Date: 30 March 2017
Please cite this article as: Nasouri K, Shoushtari AM, Designing, modeling and manufacturing of lightweight carbon nanotubes/polymer composite nanofibers for electromagnetic interference shielding application, Composites Science and Technology (2017), doi: 10.1016/j.compscitech.2017.03.041. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Designing, modeling and manufacturing of lightweight carbon nanotubes/polymer composite nanofibers for electromagnetic interference shielding application Komeil Nasouri*, Ahmad Mousavi Shoushtari
RI PT
Department of Textile Engineering, Amirkabir University of Technology, Tehran 158754413, Iran ABSTRACT
SC
Lightweight conductive multiwalled carbon nanotubes (MWCNTs) / polyvinyl alcohol
M AN U
(PVA) composite nanofibers were prepared by electrospinning process with an aim to investigate the potential of such nanofibers as an effective electromagnetic interference (EMI) shielding material. The influence of MWCNTs content, thickness, and frequency on the EMI shielding of conductive MWCNTs/ PVA composite nanofiber has been investigated. These experiments were designed by response surface methodology (RSM) and quadratic model was
TE D
used to calculation of the responses. The predicted responses were in good agreement with the experimental results according to RSM model. The RSM analysis confirmed that MWCNTs content and thickness were the main significant variables affecting the absorption shielding
EP
effectiveness. Moreover, the sample thickness has no significant influence on the reflection
AC C
shielding effectiveness. The RSM model predicted the 31.5 dB value of the highest absorption with low reflection (8.8 dB) at conditions of 7.7 wt% MWCNTs content, 3 mm of the sample thickness, and 12 GHz of incident EM wave frequency. The obtained RSM results confirmed that the selected RSM model presented suitable performance for evaluating the involved variables and prediction of EMI shielding parameters.
*
Corresponding author. Tel: +98(21)64542638; Fax: +98(21)66400245. Email address:
[email protected] (K. Nasouri)
1
ACCEPTED MANUSCRIPT Keywords: A. Carbon Nanotubes; A. Fibers; A. Nano composites; B. Electrical properties, E. Electrospinning
1. Introduction
RI PT
Electrically conductive polymer composites containing carbonic fillers are new alternative candidates for highperformance electromagnetic interference (EMI) shielding applications based on their own advantages, such as their lightweight, low cost, good processability, and
SC
tunable conductivities over a wide range [14]. Graphite [5], carbon black (CB) [6], graphene [7], carbon fibers (CFs) [8], carbon nanofibers (CNFs) [9], and carbon nanotubes (CNTs) [10]
M AN U
have been popularly employed. Among these carbonic fillers, CNTs is primarily preferred for EMI shielding materials due to its portability, extraordinary electrical property, very low percolation threshold, high aspect ratio (up to 1000), and excellent thermal stability [1113]. In order to increase the electrical and thermal properties, better dispersion of CNTs in the
TE D
polymer composites is required [14]. Several methods have been developed to produce CNTs/polymer composites, such as insitu polymerization [15], melt processing [16], solution blending [17], and electrospinning [18]. Compared with other techniques, electrospinning is a
EP
promising method to disperse CNTs in polymer composites with nanofibrous structure. There are two methods for EMI shielding which consists of reflection and absorption of the
AC C
electromagnetic wave. In recent years, the EMI shielding materials with high absorption have drawn considerable attention due to their low reflection. In the fabrication of such materials, some researchers prefer to fabricate CNTs/polymer composites with high surface area and porosity [1921]. Therefore, the electrospun CNTs/polymer composite nanofibers have interesting properties such as the large surface area, high porosity, small pore sizes, and superior thermal and electrical properties, which are excellent candidates for highperformance EMI shielding application with high absorbency content. Polyvinyl alcohol
2
ACCEPTED MANUSCRIPT (PVA), an important semicrystalline polymer, is high solubility power in most organic solvents and nontoxic [2224]. PVA is a promising polymer for potential application in biological engineering materials and shielding devices [25]. PVA nanofiber structure has good tensile strength, high thermal stability, low chemical toxicity, and high specific surface
RI PT
area [18]. Due to the high solubility and good processability of PVA, and note that the excellent MWCNTs dispersion results in an electrospinning solution, PVA can be a promising candidate for the matrix of MWCNTs [26]. Therefore MWCNTs/PVA composite nanofiber is
SC
promising structure for potential application in EMI shielding device, medical devices, and biological engineering materials.
M AN U
The EMI shielding of conductive polymer composite materials depend on several variables. As described in the literature [27], the EM frequency, electrical conductivity, and thickness of the polymer composite have the most significant variables influence in the EMI shielding with higher absorption. Presumably, finding the relationships among these variables and
composite.
TE D
reflection and absorption will be worthwhile to obtain highperformance EMI shielding
In order to overcome this problem, the mathematical procedures have been optimized by
EP
utilizing multivariate statistic techniques. Response surface methodology (RSM) is a combination of statistical and mathematical technologies useful for the modeling and analysis
AC C
of problems in which a response of interest is affected by a number of independent variables, with the objective to optimize this response [28, 29]. In recent years, this approach is successfully employed to optimizing very properties of composite materials and synthesis processes [3032]. To the best of authors’ knowledge, there is no report on simultaneous EMI shielding controlling and structure optimization with RSM technique of a composite nanofibrous mat. Therefore during this study, we applied the RSM technique to evaluate the influence of the most effective variables including CNTs loading, thickness, and frequency on
3
ACCEPTED MANUSCRIPT the EMI shielding properties of the fabricated composite nanofibers. In this contribution for the first time, RSM was used to develop an empirical model between the CNTs loading, thickness, and frequency on reflection and absorption shielding effectiveness of the composite nanofibers. The interactions of variables were studied by multiple regression analysis and
RI PT
twoway analysis of variance.
2. Materials and methods
SC
2.1. Materials
PVA powder with average molecular weight of 72,000 g/mol (98 % hydrolyzed) was supplied
M AN U
from Merck. Multiwalled carbon nanotubes (MWCNTs) with 10–20 nm outer diameter and 10–30 µm long (purity > 95% and surface area>200 m2/g) were provided by US Research Nanomaterials, Inc. Sodium dodecyl sulfate (SDS) was obtained from SigmaAldrich. The
TE D
solvent used for dissolving PVA powder and MWCNTs/PVA dispersion was distilled water.
2.2. Experimental setup for Fabrication of composite nanofibers The MWCNTs/PVA composite solutions were prepared in three steps (Fig. 1). The
EP
experimental setup (Electroris, FNM, Iran) used for electrospinning of the MWCNTs/PVA composite solutions is shown in Fig. 2. The prepared MWCNTs/PVA composite solutions
AC C
were added to a 1mL glass syringe with a stainless steel needle (Length=35 mm, Diameter=0.7 mm, and L/D=50). The flow rate of the MWCNTs/PVA composite solutions was 0.25 ml/h, electrospinning voltage at 15 kV was applied to the needle, the distance between the needle tip and collector was set for 17 cm and then the electrospun MWCNTs/PVA nanofibers was collected on takeup drum under speed of 100 RPM. The electrospinning of composite solutions were performed at 25±5 °C and 3035 % relative
4
ACCEPTED MANUSCRIPT humidity. These prepared MWCNTs/PVA composite nanofibers are collected and then dried in oven at 40 °C for 4 h. [Insert Figure 1]
RI PT
[Insert Figure 2] 2.3. Surface morphologies of the nanofibers
The surface morphologies of the electrospun MWCNTS/PVA nanofibers were observed by a
SC
scanning electron microscope (SEM, Philips, XL30) with an accelerating voltage of 25 kV.
M AN U
Before SEM imaging, the nanofibers samples were sputtered with gold for 1 min by a highvacuum sputter coater ion sputter. The average diameter, standard deviation, and diameter distribution of the MWCNTS/PVA nanofiber was measured with the SEM images using Image J software (National Institutes of Health, USA) from 100 randomly selected composite
TE D
nanofibers.
2.4. Electrical conductivity of the nanofibers mats The electrical conductivities of all nanofibers samples (including pure PVA nanofibers and
EP
MWCNTs/PVA composite nanofibers) were measured by the fourpoint probe procedure at
AC C
room temperature. The electrical current (I) was measured with a Keithley 224 electrometer and measurements of voltage (V) were performed with a Keithley 196. The electrical conductivity (σ) was calculated using the following equation [33]: σ (S cm) = I × V
L(cm) , T (cm).W (cm)
where L is distance between two electrodes, W is the width of the sample, and T is the composite nanofibers thickness (T=0.1 cm).
2.5. Shielding effectiveness of the composite nanofibers
5
ACCEPTED MANUSCRIPT The scattering parameters (S11 and S12) have been carried out using Agilent 8510C Vector Network Analyzer (VNA) in 812 GHz microwave range. The S11 and S12 of the twoport network system represent the reflection and transmission coefficient, respectively. According to the scattering theory, transmittance (T) and reflectance (R) coefficient of the shielding
RI PT
material can be described as below [4, 7]: P 2 T= t =S 12 P i
SC
P 2 R= r = S 11 P i
(1)
(2)
M AN U
where Pi, Pt, and Pr are the power of incident, transmitted, and reflected EM waves, respectively.
The total EMI shielding effectiveness (SET) of a composite is expressed in terms of the ratio of transmitted to incident energy of the EM waves. The SET is the sum of the reflection (SER)
TE D
and the absorption (SEA) of electromagnetic radiation. Therefore, the SER and SEA of the shielding composite structure are correlated with reflectance and transmittance coefficient by the following equations [4, 16]: R
1 = 10 log ( ) 1− R
(4)
AC C
1− R SE = 10 log ( ) A T
(3)
EP
SE
2.6. Design of Experiment (DoE) The effects of the three independent numerical variables namely; MWCNTs content (C, wt%), thickness (T, mm) and frequency (F, GHz) on SER and SEA of MWCNTs/PVA composite nanofibers were investigated using RSM technique. The effect of the input parameters in the EMI shielding was investigated using the design of experiment where five levels were considered for each parameter as shown in Table 1. 6
ACCEPTED MANUSCRIPT [Insert Table 1] The total number of experiments (N=18) in this study was obtained from the equation: N=2k+2k+cp, where k is the number of variables (k=3) and cp is the center of the design
RI PT
(cp=4) for estimation of a pure error sum of squares. The total number of experiments and design space of the study is shown in Fig. 3. The design of experiment with actual and coded variables is listed in Table 2. The statistical DesignExpert software (StatEase, Inc,
plot the response surface graphs.
M AN U
[Insert Figure 3]
SC
Minneapolis, MN, USA) was used for the regression analysis of the experimental data, and to
[Insert Table 2]
In a system involving three significant independent variables C, T, and F the mathematical
polynomial equation:
TE D
relationship between the response and these variables can be approximated by the quadratic
Y = β 0 + β1C + β 2T + β 3 F + β12C.T + β13C.F + β 23T .F + β C2 + β T 2 + β F 2 22
33
EP
11
(5)
where Y is the predicted responses (SER or SEA), β0 is the offset term, βi is the ith linear
AC C
coefficient, βij is the ijth interaction coefficient, and βii is the ith quadratic coefficient.
3. Results and discussion
3.1. Surface morphology of the MWCNTs/PVA electrospun mats Fig. 4 illustrates the SEM photographs of pure PVA nanofibers and MWCNTs/PVA composite nanofibers with increasing the MWCNTs content. The electrospun PVA nanofibers were ultrafine and highly regular surface morphology with an average diameter of 198±34 nm, ranging from 100 to 300 nm.
7
ACCEPTED MANUSCRIPT [Insert Figure 4] Interestingly, with the increased weight of MWCNTs in the solution from 0 to 10 wt%, it shows two significant changes (1) the nanofiber surface morphology changes from smooth to
RI PT
rough surface gradually; (2) the nanofiber diameter increases from 198 ± 34 nm to 489 ± 94 nm. These variations in nanofibers morphology and diameter can be related to the MWCNTs and polymer chain entanglements in the solution produced due to the loading of MWCNTs. This enhancement of chain entanglements by MWCNTs loading can be due to the barrier
SC
effect of individualdispersed MWCNTs into the PVA matrix. Therefore, the good dispersion
when testing for electrical properties.
M AN U
of the MWCNTs within the PVA matrix is able to influence the behavior of the composite
3.2. Electrical conductivity of the MWCNTs/PVA composite nanofibers Fig. 5 shows the measured electrical conductivity at room temperature of the MWCNTs/PVA
TE D
composite nanofibers as a function of the MWCNTs weight fraction. Only a slight increase of the electrical conductivity, from 2.34×109 to 6.11×108 S/cm, was considered with the
EP
addition of 0.5 wt% MWCNTs in nanofibers structure when compared to pure PVA nanofibers. As the MWCNT’s loading increases to 0.75 wt%, the electrical conductivity
AC C
experienced a significant improvement (3.37×103 S/cm), suggesting the formation of electrons transferring channels (conducting network) throughout the insulting PVA matrix. Therefore, the percolation threshold was found to be around fc = 0.75 wt% MWCNTs. Higher electrical conductivity could be obtained with further increase of the MWCNT loadings, reaching 8.12×102 and 8.70×101 S/cm for MWCNTs/PVA composite nanofibers containing 2.5 and 10 wt% MWCNTs, respectively. Results indicate that incorporation of MWCNTs inside PVA matrix caused increase in electrical conductivity by the order of 10 as compared to only PVA nanofibers with conductivity of 2.34×109 S/cm.
8
ACCEPTED MANUSCRIPT [Insert Figure 5] The state of dispersion of the MWCNTs inside the polymeric matrix can be analyzed via the measurement of the electrical conductivity of the composite [34, 35]. According to the
with MWCNTs loading can be described by Eq. (6): σ = σ ( f − f )t 0
c
for f > f
c
RI PT
percolation theory [36], the behavior of electrical conductivity of the polymeric composites
(6)
SC
where σ is the MWCNTs/PVA composite nanofibers conductivity, σ0 is the scaling conductivity, f is the MWCNTs weight fraction, fc is the MWCNTs weight fraction of at
M AN U
percolation threshold and t is the critical exponent. The interpolation of the experimental conductivity values leads to the determination of the above mentioned parameters: fc = 0.75 wt%, σ0= 8.44 S/cm, and t = 1.10. The lower value of percolation threshold and critical exponent of MWCNTs/PVA composite nanofibers compared to those reported in the
nanofibrous structure.
TE D
literature [1, 2, 13, 35, 36] can be attributed to the better dispersion of MWCNTs in the
EP
3.3. Response model for reflection
The design of experiment with coded variables accompanied with reflection and absorption is
AC C
shown in Table 3. As observed from the results at each design point, the reflection shielding effectiveness quantities of composite nanofibers were in the ranges of 1.1–13.0 dB, depending on the variable conditions. The analysis of variance (ANOVA) for the reflection shielding effectiveness of composite nanofibers has been summarized in Table 4. When the Pvalue is less than 0.05, the selected variable has significant effect on the output response at a confidence level at 95% (α=0.05) [28]. Otherwise (Pvalue>0.05), the variable has not significant effect on the response. The ANOVA analysis of the optimization study indicated that the model terms, x1, x3, and x1x3 were significant (P<0.05) while x2, x1x2, x2x3, x12, x22, and 9
ACCEPTED MANUSCRIPT x32 were nonsignificant (P>0.05). Therefore, the final RSM model in terms of actual variables for reflection shielding effectiveness of MWCNTs/PVA composite nanofibers is given in Eq. (7):
SE = 3.77 + 1.36C − 0.19 F − 0.03 C. F
(7)
R
RI PT
It is observed that this model is statistically significant because the Pvalue is < 0.0001. The lack of fit was used to evaluate the validity of the model. The lack of fit Fvalue of 8.71 implies the significance of the proposed equation. Fig. 6 shows a comparison between
SC
experimental reflections of MWCNTs/PVA composite nanofibers values and predicted values
M AN U
using the RSM model. [Insert Table 3] [Insert Table 4]
TE D
[Insert Figure 6]
The plot indicates a good agreement between experimental data and the ones obtained from model. The efficiency of correlated model was determined by the coefficient of determination
EP
(R2), which was calculated to be 0.9874, indicating >98% of the reflection variability in the observations. In general, the closer the R2 value to 1.00, the effective the RSM model is and
AC C
the superior it calculates the reflection shielding effectiveness. The adjusted coefficient of determination (Radj2) indicates the amount of variation in the RSM model. The value of the Radj2 is also very high (98.47%), which indicates a high significance of the RSM model.
3.4. Effect of variables on reflection From statistical results (see Table 4), thickness of the composite nanofibers sample has no significant effect on the reflection. Predicting the expected model equation for reflection can
10
ACCEPTED MANUSCRIPT be obtained by the contour plot. In the contour plot, lines of the constant reflection are drawn in the plane of the selected variables. This plot gives very useful information about the fitted model. Fig. 7 shows the contour surface plots of the reflection shielding effectiveness of the composite nanofibers at different MWCNTs contents and frequency. From the Fig. 7 it can be
RI PT
seen that the reflection shielding effectiveness decreases with increase in frequency for all levels of MWCNTs content and thickness. Theoretically, the reflection shielding effectiveness of composite materials can be predicted using the following derived model [10, 16, 37]:
σ ) 2π f .µ
SC
SER = 39.5 + 10 log (
(8)
M AN U
where σ is the electrical conductivity of the composite (S/m), f is the frequency of radiation (Hz), µ is the magnetic permeability of the sample, µ=µ0µr, µ0= 4 π ×107 H/m is the permeability of free space.
TE D
[Insert Figure 7]
The Eq. (8) indicates that reflection is inversely proportional to frequency of the incident electromagnetic wave (reflection ∝ 1/frequency). Moreover, increasing reflection was found
EP
to decrease with frequency as experimentally observed in other researchers for MWCNTs/polymer composite [16, 37]. It can be seen from Fig. 6, increasing MWCNTs
AC C
content will result in higher electrical conductivity enabling the larger reflection of electromagnetic wave. According to the Eq. (8), the reflection shielding effectiveness is directly proportional to electrical conductivity of composite nanofibers. Our finding is consistent with the trend experimentally observed by other researchers [13, 16, 19].
3.5. RSM model fitting for absorption Table 5 shows the analysis of variance results of the established RSM model for EMI shielding by absorption mechanism. The ANOVA analyses of the quadratic polynomial
11
ACCEPTED MANUSCRIPT model indicated that the selected RSM model was highly efficacious, as the Fvalue for the RSM model was 330.03 (Pvalue< 0.0001). From the Pvalues listed in Table 5, the terms x2x3, x22, and x32 have nonsignificant impact on absorption. But the Pvalues for the terms x1, x2, x3, x1x2, x1x3, and x12 are less than 0.05. Meanwhile, the lack of fit Pvalue for the RSM
RI PT
model is greater than 0.05 (Fvalue=8.34) which is nonsignificant. Insignificant lack of fit is a prerequisite for a successful predictive model. Therefore, regression analysis of experimental data was performed and the RSM model in terms of actual variables is presented
SC
in Eq. (9): SE = 0.67 − 1.38C + 0.45T + 0.10 F + 0.52 C.T + 0.18 C. F + 0.17 C 2
(9)
M AN U
A
Fig. 8 shows a comparison between experimental absorption shielding effectiveness values and predicted values using the RSM model. In this case, the value of the R2=0.9962 indicates that only 0.38% of the total variations cannot be explained by the RSM model. The value of
TE D
the Radj2=0.9942 is also high to advocate a high significance of the RSM model. [Insert Table 5]
EP
[Insert Figure 8]
AC C
3.6. Variables interaction effects on absorption The interactive effects of two independent parameters are shown in contour plots (Fig. 9). As it is indicated in Fig. 9a, with an increase in the MWCNTs content (up to 10 wt%) and thickness, the EMI absorption efficiency raised proportionally until it reached the optimum value. The MWCNTs content in the composite nanofibers is more important factor than thickness in the case of EMI shielding by absorption mechanism. Fig. 9b illustrates the absorption capacity dependence on MWCNTs content and frequency for constant thickness. It is observed that the absorption shielding effectiveness increased with an increase in both
12
ACCEPTED MANUSCRIPT factors in the MWCNTs content and frequency. The contour plots indicate that the resultant absorption shielding effectiveness is very responsive to the changes in MWCNTs content which is quite similar to earlier studies [19, 20]. According EMI shielding theory, a mathematical relation of absorption shielding effectiveness with thickness (T), frequency (f),
RI PT
and electrical conductivity (σ) of composite material are given as [16, 38]: SE = 8.7 T π . f .µ.σ A
(10)
Therefore, the absorption shielding effectiveness is directly proportional to thickness,
SC
frequency, and electrical conductivity of composite nanofibers. The results are in good
M AN U
agreement with those reported in literatures for absorption shielding effectiveness of MWCNTs/polymer composites [1, 3, 13, 16].
[Insert Figure 9]
TE D
3.7. Optimizing of responses using RSM
A key advantage of using RSM is it enables one to optimize responses by controlling the independent variables. The objective of RSM optimization is to find a desirable location in
EP
the design space. The numerical optimization finds a point that maximizes the desirability function [31]. In this section, using Design Expert software, the optimizing of influential
AC C
factors was done, in order to achieve the optimum values of SEA and SER (higher absorption and lower reflection). The conditions for lower reflection and higher absorption shielding effectiveness estimated by the RSM equations were MWCNTs content=7.7 wt%, thickness=3 mm, and frequency = 12 GHz. The reflection and absorption shielding effectiveness under the above conditions was SER=8.8 dB and SEA=31.5 dB. The absorption shielding effectiveness of 30 dB means that 99.9% of the EM waves have been attenuated. For optimized composite
13
ACCEPTED MANUSCRIPT nanofibers, the SEA shows more than 99.99 % absorption shielding effectiveness of EM energy over the entire frequency range of measurement.
4. Conclusions
RI PT
The electrically conductive MWCNTs/PVA composite nanofibers were successfully fabricated by electrospinning approach. The average diameter and electrical conductivity of composite nanofibers were found to be dependent on MWCNTs content and showed an
SC
increase with increasing MWCNTs loading. RSM technique was used to modeling and optimizing of the reflection and absorption shielding effectiveness of composite nanofibers.
M AN U
The variables include MWCNTs content, sample thickness, and frequency. The RSM analysis confirmed that MWCNTs contents were the main significant variables affecting the EMI shielding. High regression coefficient between the independent variables and the reflection (R2=0.9874) and absorption (R2=0.9962) indicates excellent evaluation of experimental data
TE D
by quadratic model. Based on the RSM function, the finest value for each process variable was also determined for MWCNTs content (7.7 wt%), thickness (3 mm), and frequency (12 GHz). The reflection and absorption shielding effectiveness under the optimized settings was
EP
reflection=8.8 dB and absorption=31.5 dB.
AC C
Acknowledgments
The authors would like to thank to the National Elites Foundation of Iran (NEF) for financial support and encouragement of this research.
References [1]
T. Kuang, L. Chang, F. Chen, Y. Sheng, D. Fu, X. Peng, Facile preparation of lightweight highstrength biodegradable polymer/multiwalled carbon nanotubes
14
ACCEPTED MANUSCRIPT nanocomposite foams for electromagnetic interference shielding, Carbon 105 (2016) 305–313. [2]
R. Dou, Y. Shao, S. Li, B. Yin, M. Yang, Structuring tricontinuous structure multiphase composites with ultralow conductive percolation threshold and excellent
RI PT
electromagnetic shielding effectiveness using simple melt mixing, Polymer 83 (2016) 34–39. [3]
I.W. Nam, H.K. Lee, J.H. Jang, Electromagnetic interference shielding/absorbing
SC
characteristics of CNTembedded epoxy composites, Compos. Part A 42 (2011) 1110– 1118.
S. Kashi, R.K. Gupta, T. Baum, N. Kao, S.N. Bhattachary, Morphology,
M AN U
[4]
electromagnetic properties and electromagnetic interference shielding performance of poly lactide/ graphene nanoplatelet nanocomposites, Mater. Des. 95 (2016) 119–126. [5]
R.K. Goyal, Costefficient high performance polyetheretherketone/expanded graphite
TE D
nanocomposites with high conductivity for EMI shielding application, Mater. Chem. Phys. 142 (2013) 195–198. [6]
J.S. Im, J.G. Kim, Y.S. Lee, Fluorination effects of carbon black additives for electrical
EP
properties and EMI shielding efficiency by improved dispersion and adhesion, Carbon 47 (2009) 2640–2647.
J. Chen, J. Wu, H. Ge, D. Zhao, C. Liu, X. Hong, Reduced graphene oxide deposited
AC C
[7]
carbon fiber reinforced polymer composites for electromagnetic interference shielding, Compos. Part A 82 (2016) 141–150. [8]
S.S. Hwang, Tensile, electrical conductivity and EMI shielding properties of solid and foamed PBT/carbon fiber composites, Compos. Part B 98 (2016) 1–8.
15
ACCEPTED MANUSCRIPT [9]
A. Das, H.T. Hayvaci, M.K. Tiwari, I.S. Bayer, D. Erricolo, C.M. Megaridis, Superhydrophobic and conductive carbon nanofiber/PTFE composite coatings for EMI shielding, J. Colloid Interface Sci. 353 (2011) 311–315.
[10] M.H. Saleh, Electrical, EMI shielding and tensile properties of PP/PE blends filled with
RI PT
GNP: CNT hybrid nanofiller, Synth. Met. 217 (2016) 322–330.
[11] D. Micheli, A. Vricella, R. Pastore, A. Delfini, A. Giusti, M. Albano, et al., Ballistic and electromagnetic shielding behaviour of multifunctional Kevlar fiber reinforced epoxy
SC
composites modified by carbon nanotubes, Carbon 104 (2016) 141–156.
[12] K. Nasouri, A.M. Shoushtari, A. Kaflou, H. Bahrambeygi, A. Rabbi, Singlewall carbon
M AN U
nanotubes dispersion behavior and Its effects on the morphological and mechanical properties of the electrospun nanofibers, Polym. Comp. 33 (2012) 1951–1959. [13] Z. Zeng, M. Chen, H. Jin, W. Li, X. Xue, L. Zhou, et al., Thin and flexible multiwalled carbon
nanotube/waterborne
polyurethane
composites
with
highperformance
TE D
electromagnetic interference shielding, Carbon 96 (2016) 768–777. [14] X. Gao, A.I. Isayev, C. Yi, Ultrasonic treatment of polycarbonate/carbon nanotubes composites, Polymer 84 (2016) 209–222.
EP
[15] H.J. Lee, S.J. Oh, J.Y. Choi, J.W. Kim, J. Han, L.S. Tan, et al., In situ synthesis of poly (ethylene terephthalate) (PET) in ethylene glycol containing terephthalic acid and
AC C
functionalized multiwalled carbon nanotubes (MWNTs) as an approach to MWNT/PET nanocomposites, Chem. Mater. 17 (2005) 5057–5064. [16] M.H. Saleh, U. Sundararaj, Electromagnetic interference shielding mechanisms of CNT/polymer composites, Carbon 47 (2009) 1738–1746. [17] S. Kaida, J. Matsui, T. Sagae, Y. Hoshikawa, T. Kyotani, T. Miyashita, The production of large scale ultrathin aligned CNT films by combining AC electric field with liquid flow, Carbon 59 (2013) 503–511.
16
ACCEPTED MANUSCRIPT [18] G. Salimbeygi, K. Nasouri, A.M. Shoushtari, R. Malek, F. Mazaheri, Fabrication of polyvinyl alcohol/multiwalled carbon nanotubes composite electrospun nanofibres and their application as microwave absorbing material, Micro Nano Lett. 8 (2013) 455–459. [19] Q. Li, L. Chen, J. Ding, J. Zhang, X. Li, K. Zheng, X. Zhang, X. Tian, Opencell
RI PT
phenolic carbon foam and electromagnetic interference shielding properties, Carbon 104 (2016) 90–105.
[20] H. Wang, K. Zheng, X. Zhang, X. Ding, Z. Zhang, C. Bao, et al., 3D network porous
SC
polymeric composites with outstanding electromagnetic interference shielding, Compos. Sci. Technol. 125 (2016) 22–29.
M AN U
[21] Y. Li, B. Shen, X. Pei, Y. Zhang, D. Yi, W. Zhai, et al., Ultrathin carbon foams for effective electromagnetic interference shielding, Carbon 100 (2016) 375–385. [22] C. Shao, H.Y. Kim, J. Gong, B. Ding, D.R. Lee, S.J. Park, Fiber mats of poly (vinyl alcohol)/ silica composite via electrospinning, Mater. Lett. 57 (2003) 1579–1584.
TE D
[23] G. Salimbeygi, K. Nasouri, A.M. Shoushtari, Fabrication of homogeneous multiwalled carbon nanotube/ poly (vinyl alcohol) composite films using for microwave absorption application, Fiber. Polym. 15 (2014) 583–588.
EP
[24] M. Krumova, D. Lopez, R. Benavente, C. Mijangos, J.M. Perena, Effect of crosslinking on the mechanical and thermal properties of poly (vinyl alcohol), Polymer 41 (2000)
AC C
9265–9272.
[25] H.R. Kim, K. Fujimori, B.S. Kim, I.S. Kim, Lightweight nanofibrous EMI shielding nanowebs prepared by electrospinning and metallization, Compos. Sci. Technol. 72 (2012) 1233–1239. [26] J. Yun, J.S. Im, Y.S. Lee, H.I. Kim, Effect of oxyfluorination on electromagnetic interference shielding behavior of MWCNT/ PVA/ PAAc composite microcapsules, Eur. Polym. J. 46 (2010) 900–909.
17
ACCEPTED MANUSCRIPT [27] J. Huo, L. Wang, H. Yu, Polymeric nanocomposites for electromagnetic wave absorption, J. Mater. Sci. 44 (2009) 3917–3927. [28] S.O. Gonen, M.E. Taygun, S. Kucukbayrak, Evaluation of the factors influencing the
Behnken design, Mater. Sci. Eng. C 58 (2016) 709–723.
RI PT
resultant diameter of the electrospun gelatin/sodium alginate nanofibers via Box–
[29] Y. Zheng, Q. Tang, T. Wang, J. Wang, Molecular size distribution in synthesis of polyoxymethylene dimethyl ethers and process optimization using response surface
SC
methodology, Chem. Eng. J. 278 (2015) 183–189.
[30] B.P. Chang, H.M. Akil, M.G. Affendy, A. Khan, R.B.M. Nasir, Comparative study of
M AN U
wear performance of particulate and fiberreinforced nanoZnO/ultrahigh molecular weight polyethylene hybrid composites using response surface methodology, Mater. Des. 63 (2014) 805–819.
[31] S. Daneshpayeh, F.A. Ghasemi, I. Ghasemi, M. Ayaz, Predicting of mechanical
TE D
properties of PP/LLDPE/TiO2 nanocomposites by response surface methodology, Compos. Part B 84 (2016) 109–120.
[32] K. Kalantari, M.B. Ahmad, H.R.F. Masoumi, K. Shameli, M. Basri, R. Khandanlou,
EP
Rapid and high capacity adsorption of heavy metals by Fe3O4/ montmorillonite nanocomposite using response surface methodology: Preparation, characterization,
AC C
optimization, equilibrium isotherms, and adsorption kinetics study, J. Taiwan. Inst. Chem. Eng. 49 (2015) 192–198. [33] S.S. Qavamnia, K. Nasouri, Conductive polyacrylonitrile/ polyaniline nanofibers prepared by electrospinning process, Polym. Sci. Ser. A 57 (2015) 343–349. [34] N. Wang, Y. Wang, Z. Yu, G. Li, In situ preparation of reinforced polyimide nanocomposites with the noncovalently dispersed and matrix compatible MWCNTs, Compos. Part A 78 (2015) 341–349.
18
ACCEPTED MANUSCRIPT [35] H. Yazdani, B.E. Smith, K. Hatami, Multiwalled carbon nanotubefilled polyvinyl chloride composites: influence of processing method on dispersion quality, electrical conductivity and mechanical properties, Compos. Part A 82 (2016) 65–77. [36] M. Fogel, P. Parlevliet, M. Geistbeck, P. Olivier, E. Dantras, Thermal, rheological and
RI PT
electrical analysis of MWCNTs/epoxy matrices, Compos. Sci. Technol. 110 (2015) 118–125.
[37] L. Nayak, D. Khastgir, T.K. Chaki, A mechanistic study on electromagnetic shielding
SC
effectiveness of polysulfone/carbon nanofibers nanocomposites, J. Mater. Sci. 48 (2013) 1492–1502.
M AN U
[38] M.H. Saleh, U. Sundararaj, Xband EMI shielding mechanisms and shielding effectiveness of high structure carbon black/polypropylene composites, J. Phys. D:
AC C
EP
TE D
Appl. Phys. 46 (2013) 304310.
19
ACCEPTED MANUSCRIPT Figure captions: Fig. 1. Flow chart for the synthesis of MWCNTs/PVA composite solutions
RI PT
Fig. 2. Schematic of electrospinning apparatus Fig. 3. Schematic of design space (Dots in cubic indicate the location of runs in experiment) Fig. 4. Surface morphology of electrospun MWCNTs/PVA composite nanofibers for various
SC
MWCNTs contents: (a) 0 (pure PVA nanofibers), (b) 2.5, (c) 5, (d) 7.5, and (e) 10 wt%
MWCNTs/ PVA composite nanofibers.
M AN U
Fig. 5. Dependence of the electrical conductivity on the MWCNTs weight fraction of
Fig. 6. The experimental versus RSM predicted plot for reflection shielding effectiveness
TE D
(dB) of composite nanofibers.
Fig. 7. The contour plots of the reflection shielding effectiveness of composite nanofibers as a function of MWCNTs content and frequency.
EP
Fig. 8. Correlation plot between RSM predicted and experimental values of the absorption
AC C
shielding effectiveness.
Fig. 9. Response surface plots showing effect of two independent variables on absorption shielding effectiveness: (a) MWCNTs content and thickness, and (b) MWCNTs content and frequency.
20
ACCEPTED MANUSCRIPT Table 1 Actual values of the variables and their levels Level2
Level3
Level4
Level5
MWCNTs content (wt%)
0
2.5
5
7.5
10
Thickness (mm)
1
1.5
2
Frequency (GHz)
8
9
10
RI PT
Level1
2.5
3
11
12
AC C
EP
TE D
M AN U
SC
Input parameters
21
ACCEPTED MANUSCRIPT
Table 2 RSM design and corresponding process variable values for each experiment
Run
MWCNTs content
Thickness
Coded values of the variables Frequency
MWCNT.
(T, mm)
(F, GHZ)
(x1)
1
0
1
12
1
2
5
2
10
0
3
2.5
2
10
4
0
3
8
5
0
1
6
10
1
7
5
2
8
7.5
2
9
5
10
10
11
5
12
5
13
5
14
5
16 17 18
Freq. (x3)
1
+1
0
0
0
0
1
+1
1
M AN U
0.5
1
1
1
12
+1
1
+1
10
0
0
0
10
+0.5
0
0
1.5
10
0
0.5
0
3
8
+1
+1
1
2
11
0
0
+0.5
2
10
0
0
0
2
10
0
0
0
2
9
0
0
0.5
0
3
12
1
+1
+1
10
1
8
+1
1
1
10
3
12
+1
+1
+1
5
2.5
10
0
+0.5
+1
EP
TE D
8
AC C
15
Thick. (x2)
SC
(C, wt%)
RI PT
Input variables
22
ACCEPTED MANUSCRIPT
Table 3 The experimental design for the three independent variables and the responses at different variable levels SER (dB)
SEA (dB)
RI PT
Independent variables Run x2
x3
Actual
Predicted
Actual
Predicted
1
1
1
+1
1.5
1.5
2.4
2.3
2
0
0
0
7.0
7.2
13.7
14.1
3
0.5
0
0
4.3
4.5
7.1
7.3
4
1
+1
1
1.9
2.3
2.9
2.8
5
1
1
1
6
+1
1
+1
7
0
0
0
8
+0.5
0
0
9
0
0.5
10
+1
+1
11
0
0
12
0
0
13
0
14
0
15
M AN U
SC
x1
2.3
2.2
1.9
10.5
11.5
31.4
32.3
7.2
7.2
14.5
14.1
10.6
9.8
24.1
23.1
0
7.2
7.2
12.1
12.6
1
12.4
13.5
35.2
36.0
+0.5
6.6
6.8
15.9
15.1
0
7.4
7.2
13.7
14.1
0
0
7.1
7.2
14.1
14.1
0
0.5
7.7
7.5
12.2
13.1
1
+1
+1
1.1
1.5
3.1
3.2
16
+1
1
1
13.0
13.5
26.3
24.7
17
+1
+1
+1
10.8
11.5
44.7
43.6
18
0
+0.5
+1
7.3
7.2
16.7
15.6
AC C
EP
TE D
2.1
23
ACCEPTED MANUSCRIPT Table 4 Analysis of variance for quadratic model of SER Sum of
Degree of
Mean of
squares
freedom
squares
226.98
9
25.22
170.90
x1MWCNTs content 220.07
1
220.07
1491.26
x2Thickness
0.08
1
0.08
0.58
x3Frequency
4.31
1
4.31
29.18
x 1x 2
0.01
1
0.01
x 1x 3
0.91
1
x 2x 3
0.06
1
x 12
0.04
1
x 22
0.02
1
x 32
0.10
1
Residual
1.18
Lack of fit
1.09
Pvalue
Status
< 0.0001
Significant
< 0.0001
Significant 
0.0006
Significant
0.08
0.7895

0.91
6.18
0.0378
Significant
0.06
0.42
0.5374

0.04
0.28
0.6140

0.02
0.14
0.7176

0.10
0.68
0.4336

TE D
M AN U
SC
0.4697
8
0.15


5
0.22
7.50
0.0640
AC C
EP
Model
Fvalue
RI PT
Source
24
Not significant
ACCEPTED MANUSCRIPT Table 5 ANOVA results for the absorption (SEA, dB) of MWCNTs/PVA composite nanofibers Degree of
Mean of
squares
freedom
squares
Model
2432.83
9
x1MWCNTs content
2160.03
x2Thickness
Fvalue
Pvalue
270.31
330.03
1
2160.03
2637.20
78.92
1
78.92
96.35
x3Frequency
33.40
1
33.40
40.78
RI PT
Sum of Source
Status
x 1x 2
54.08
1
54.08
x 1x 3
25.21
1
x 2x 3
2.42
1
x 12
4.88
1
x 22
0.02
1
x 32
0.22
1
Residual
6.55
Lack of fit
6.11
Significant
< 0.0001
Significant
< 0.0001
Significant
0.0002
Significant
66.03
< 0.0001
Significant
25.21
30.77
0.0005
Significant
2.42
2.95
0.1240

4.88
5.96
0.0405
Significant
0.02
0.02
0.8839

0.22
0.27
0.6189

TE D
M AN U
SC
< 0.0001
0.82


5
1.22
8.34
0.0556
AC C
EP
8
25
Not significant
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT