Recycled carbon fibre reinforced polymer composite for electromagnetic interference shielding

Recycled carbon fibre reinforced polymer composite for electromagnetic interference shielding

Composites: Part A 41 (2010) 693–702 Contents lists available at ScienceDirect Composites: Part A journal homepage: www.elsevier.com/locate/composit...

1MB Sizes 0 Downloads 46 Views

Composites: Part A 41 (2010) 693–702

Contents lists available at ScienceDirect

Composites: Part A journal homepage: www.elsevier.com/locate/compositesa

Recycled carbon fibre reinforced polymer composite for electromagnetic interference shielding K.H. Wong *, S.J. Pickering, C.D. Rudd School of Mechanical, Materials and Manufacturing Engineering, The University of Nottingham, University Park, Nottingham, NG7 2RD, UK

a r t i c l e

i n f o

Article history: Received 20 August 2009 Received in revised form 19 January 2010 Accepted 22 January 2010

Keywords: A. Carbon fibres A. Recycling A. Thermosetting resin B. Electrical properties

a b s t r a c t This paper describes the development of an electromagnetic interference shielding material using recycled carbon fibre. Fibre recycled from a fluidised bed process was transformed into a non-woven veil and was moulded into a glass-fibre reinforced polymer plaque to provide shielding. Factors affecting shielding performance were established using a virgin fibre and the result was compared with veil made of the recycled fibre. Shielding performance increased with veil areal densities. The influence of fibre length on shielding seemed insignificant provided the fibre was distributed evenly. Sandwiching the plaque between fibre veils enhanced the shielding performance. A shielding value of 40 dB was attained from a layer of 80 g/m2 recycled fibre veil and it was increased to 70 dB when the plaque was sandwiched between two layers of 30 g/m2 recycled fibre veil. The correlation between veil formation and shielding effectiveness was established and found that shielding effectiveness increased with the degree of fibre dispersion. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Research into the recycling polymer composite scrap is the subject of international activity and a variety of recycling technologies ranging from mechanical size reduction [1,2] to the use of supercritical fluids [3,4] have been investigated. Ways to reuse the recyclate are also under consideration and this must be proven viable to close the recycling route in addition to making the recycling system economically sustainable. Carbon fibre is a valuable commodity that has been used increasingly in polymeric composites. In 2006, the price of carbon fibre lay between €15 and €19 per kg, which was typically ten times more expensive than commonly used glass fibre [5]. The global demand for PAN-based carbon fibre was estimated to be about 27,000 tonnes in year 2006, and the demand was expected to grow at 15% per year in the future [6]. Thus, a financial opportunity is available for exploitation. Routes for reusing the recycled composites have focused mainly on the reinforcing potential for structural applications. Comminuted recyclate and recycled fibre from other means have been investigated either as a filler or a reinforcement in thermosetting bulk moulding compound (BMC) or sheet moulding compound (SMC) [7], injection-moulded thermoplastic composites [8–10] and construction [11–13]. Exploitation of other feasible applications is lacking but is necessary to expand the reuse rate. Carbon fibre has a low electrical resistivity and its incorporation into a polymeric matrix offers application for electromagnetic interference (EMI) shielding. This is a low-risk and cost-effective approach to reuse the recycled * Corresponding author. Tel.: +44 115 951 3797; fax: +44 115 951 3800. E-mail address: [email protected] (K.H. Wong). 1359-835X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.compositesa.2010.01.012

carbon fibre, as retention of fibre strength is not as critical as in other engineering applications. A key limitation in the use of conventional polymers in electrical and electronic applications is transparency to electromagnetic waves, which has inhibited electromagnetic compliance (EMC). EMC is a key requirement in ensuring good functionality for electronic components in a common electromagnetic environment. Enhancing the electrical conductivity of the polymer improves its electromagnetic shielding ability and this can be achieved by inclusion of recycled carbon fibre to form a conductive network within the polymer matrix [14]. The ease of forming the network increases with fibre length and so the shielding ability will be enhanced if longer fibres are used. This was supported by the investigation performed by Luo and Chung [15], who suggested that composites with continuous carbon fibres provided higher shielding than those with discontinuous fibres. Similarly, comparing short carbon fibre and carbon black powder, at the same concentration, the former, with a higher aspect ratio, provided a higher shielding performance than the latter [16]. In the present work, carbon fibre was recycled from thermoset composite scrap using a fluidised bed process [17] and was then processed via a papermaking method into a non-woven veil with different areal densities. Unlike a twin-screw compounding method [18], the degradation of fibre length in the papermaking process is lessened, and the fibres are dispersed randomly and physically interconnected, thus a conductive network is readily formed. Other benefits of the papermaking process are low cost and the capability for high volume production and the carbon fibre paper produced has wide industrial application, such as for antistatic products, EMI shielding, resistive heating, chemical resistance and also structural

694

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

purposes [19]. The quality of fibre dispersion in the veil is generally known as formation, which is defined as the degree of variation of the micro-scale base weight. Good formation means the fibre is homogenously distributed whilst a veil with poor formation includes bundles and strings [20]. In this study, the main objective was to investigate the feasibility of developing an EMI screening material using recycled carbon fibre. The veils were moulded onto a glass fibre reinforced polyester plaque surface to provide shielding capability. The shielding performance of the composite is quantified in term of shielding effectiveness (SE), which is the logarithmic ratio of incident to transmitted wave power and is expressed in decibels (dB). The effects of fibre length, loading and veil configuration on the shielding performance were established using a virgin carbon fibre, and the results were then benchmarked against a veil made from the recycled fibre. The influence of veil formation on shielding performance was also evaluated. 2. Experimental 2.1. Materials Virgin epoxy-sized Toray T600SC-24000-60E carbon fibre was supplied by the Advanced Composites Group, Heanor, UK. The fibre has a diameter of 7 lm, a specific density of 1.8, a tensile strength and a Young’s modulus of 4.1 GPa and 230 GPa respectively. A pneumatic chopper gun was used to cut the fibre into staple lengths of 6.4 mm, 9.6 mm, 14.4 mm, 19.2 mm and 28.8 mm. Scrap prepreg roll with a width of 0.6 m, comprising an MTM28-2 epoxy resin reinforced with the similar grade of unidirectional carbon fibre, was also provided by the Advanced Composites Group. The prepreg was cut across the fibre axis into sheets with dimensions approximately 0.6 m by 1 m and then cured at 100 °C for 5 h inside an oven. After removing the non-stick silicon paper, the cured prepreg was cut into 30 mm-wide ribbons (as measured along the fibre axis) using a paper guillotine. The ribbons were torn manually along the fibre axis into pieces of width 30 mm and length 60 mm to 100 mm. A Retsch SM2000 miller with a 20 mm-by-20 mm square aperture sieve was used to shred the prepreg into irregular small strips in order to mimic an industrial size reduction process. The dimension of the strips was not measured but visually consisted of width ranging from 0.5 to 30 mm and length 2 to 15 mm. 2.2. Fluidised bed recycling process Fig. 1 shows a schematic representation of the fluidised bed recycling rig. Two electric heaters preheated the air to 550 °C and

the shredded prepreg was then fed at 2 g/min into a bubbling sand bed. The average sand particle diameter was 0.85 mm and the fluidising velocity was 1.0 m/s. The bubbling condition ensured a rapid and uniform heating of the shredded prepregs. Once at high temperature, the epoxy resin was burnt off and fibres were released and carried upward out of the fluidised bed by the hot air. A high efficiency cyclone was then used to disengage the recycled fibre from the outgoing exhaust. The fibre was collected inside a bin connected beneath the cyclone. Exhaust was directed to an afterburner to oxidise all of the volatile gases at 850 °C before being discharged to atmosphere. 2.3. Manufacture of carbon fibre veil A wet papermaking method was used to convert both virgin and recycled carbon fibres into circular discs of non-woven veil (190 mm in diameter) with areal densities ranging from 20 g/m2 to 100 g/m2 (gsm), in increments of 20 gsm. The method comprised three stages; fibre dispersion, fibre filtration and veil drying. For virgin fibre, weighed samples of chopped fibre were added to a water solution and stirred using a top-entry high-shear radial impeller at a speed of 800 rpm for 2–5 min. The water solution contained 4.5 g/l of hydroxyethyl cellulose binding agent from Dow Chemical Company Ltd., Middlesex, UK, 0.6 g/l of MerpolÒ OJ non-ionic surfactant from Sigma–Aldrich Company Ltd., Dorset, UK and 0.3 g/l of Bevaloid 6686 W anti-foaming agent from Rhodia UK Ltd., Watford, UK. The process was undertaken using a 10 l, 4sided, baffled tank to ensure an efficient disintegration and dispersion of the fibre tows. The fibre suspension was then poured into a filtration unit, where the water solution was drained through a stainless steel wire mesh. Excessive water solution from the wet veil was vacuumed off before removing the veil from the wire mesh. The veil was then dried in an oven at 100 °C for 15 min. The recycled carbon fibre was fluffy and, when wetted by the water solution, formed a dense fibre clump. The clump was partially disintegrated using a top-entry axial impeller rotating at a speed of 800 rpm in a smooth 10 l stainless steel container for 10 min. The clumps were then further disintegrated using the baffled tank and radial impeller. Similar fibre filtration and veil drying processes were then followed. 2.4. Manufacture of composite Composite specimens were manufacture in a resin transfer moulding (RTM) technique using an unsaturated isophthalic polyester resin, Crystic 701PA from Scott Bader, Wellingborough, UK, Eglass continuous filament mats (CFM), i.e. U750/450 m2-6% binder

Fig. 1. A schematic representation of the fluidised bed recycling process.

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

and U751/375/m2 from Saint-Gobain Vetrotex, Herzogenrath, Germany as well as the virgin and recycled carbon fibre veils. The resin was pre-treated with 1.4 wt.% of NL-49P cobalt accelerator and 2 wt.% of MEKP catalyst, Butanox M50, to give a gel time of approximately 13 min at room temperature. Both accelerator and catalyst were supplied by K & C Mouldings Ltd., Norfolk, UK. Preform was first fabricated by pressing and consolidation at 70 °C under a pressure of 30 bar for 10 min. Three types of preforms were fabricated in order to investigate the effects of veil position and quantity on the shielding effect for the composite plaque. First, as shown in Fig. 2, multiple circular carbon fibre veils were located on the top surface of three layers of 450 gsm CFM (U750/450/m26% binder), followed by a layer of 375 gsm CFM (U751/375/m2) and a layer of 450 gsm CFM (U750/450/m2-6% binder). The second type of preform had an extra layer of a similar grade of carbon fibre veil located immediately underneath the CFM stacks to create a sandwich structure and this layout is referred to as a double configuration. The third type comprised only the CFM from which a plain glass-fibre reinforced polymer (GRP) plate was fabricated. The preform was placed within a 2.8 mm steel frame and then pressed between two flat platens. This was followed with air removal via vacuum suction and resin impregnation at an injection pressure of 7 bars. When the impregnation completed, the platens were then heated from room temperature to 60 °C for curing. 2.5. Recycled fibre length measurement The length distribution of the recycled fibre was measured by an image analysis technique developed in an earlier project [21]. The recycled fibres were dispersed onto a glass petri dish. The focal length and aperture of a CCTV camera lens was adjusted until a sharp picture was shown on a computer monitor. The image was captured and then analysed by Aphelion image processing software from ADCIS, Herouville-Saint-Clair, France. The software possessed a range of built-in commands to process and analyse the captured images. Instead of manually repeating similar commands for measuring the length of each fibre, the process was automated by a macro. The software was able to deal with crossing fibres and fibres with slight curvature. A total of 750 fibres were measured and the number average and weight average length of the distribution were then determined [22]. The former, which is also known as an arithmetic mean, provides an image of the physical length distribution of the fibres and it assumes each fibre contributing equally to the final average value. However, this might not be true, particularly for the degree of fibre dispersion, as longer fibres tend

695

to be more easily tangled up and resulted in a poorer dispersion. The use of weight average value allows the longer fibres to contribute more and thus a more meaningful and effective measure of mean length is obtained.

P Ni Li Number average length; LN ¼ P Ni

ð1Þ

P Ni L2i Weight average length; LW ¼ P Ni Li

ð2Þ

where Ni is the number of fibres of length Li. 2.6. Single fibre electrical resistivity measurement A four-point probe single fibre resistivity measurement rig was set up to measure the electrical resistivity of virgin and recycled carbon fibres. The four-point probe setup is schematically presented in Fig. 3. A stripboard with a 2.54 mm pitch and 1 mmdiameter hole was used. The outer two wires (about 0.5 mm diameter), 10.16 mm apart, were connected in series with: (1) a variable resistor with resistance ranging from 0 to 500 kX, (2) a precision linear power supply from TTI (model: TS3022S) providing 3 V to the circuit and (3) a multimeter (model 1705f from TTI) to read the current. The inner two wires, 5.08 mm apart, were connected to another multimeter (Fluke 45) to record the potential difference. A single carbon fibre was placed on the four wires, and silver paint was used to make fibre-to-wire connection. The variable resistor was adjusted to provide a constant current of 10 lA to prevent the fibre from overheating. The current direction was then reversed, and the average of the potential drops across the two inner wires was then taken. At least 15 fibres were measured from each type of fibres. The resistivity of the fibre, expressed in X m, is calculated using Eq. (3) below:



 

pd2 V 4l

I

ð3Þ

where d = fibre diameter, m; l = fibre length between the two inner wires, m; V = voltage, V; A = current, A. The fibre diameter was measured using a scanning electron microscope, JSM-6400. The fibre was placed on an adhesive and silver-coated at 15 mA for two minutes. The fibre diameter was determined by measuring the width of at least 20 filaments at a magnification of 3500.

Carbon fibre veils

A A A B A

A = one layer of 450gsm CFM with binder B = one layer of 375gsm CFM Fig. 2. Preform for RTM.

696

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

V Voltmeter Stripboard

Carbon fibre

A Ammeter Variable resistor

Power supply

Fig. 3. A schematic representation of the four-point probe for single fibre resistivity measurement.

2.7. Veil formation measurement

one hand, same result trends were obtained from analysing the dispersion of a bleached Eucalypt Kraft pulp using the two indices. On the other hand, with increasing fibre loadings, only the NFI correctly recognising changes in dispersion and provided a meaningful result. Both indices are incapable of distinguishing textures and patterns within veils of similar formation index [24]. All the veils in the current investigation are manufactured using a papermaking method, which means there is only one texture available, i.e. randomly oriented. Hence, the limitation of first-order statistics in not recognising texture and pattern should not be an issue. The application of the COV to formation measurements has been found in areas not involving texture recognition. For example, Nazhad et al. [25] used it to study the effect of formation on paper tensile strength, and Ichiura et al. [26] studied the relationship between acetaldehyde absorptivity and the formation of a zeolite sheet. Both experimental specimens were fabricated using a papermaking process and possessed a randomly-oriented texture. 2.8. EMI shielding effectiveness measurement

An in-house formation measurement system was developed using a light transmittance method. The entire system was located inside a dark room. A Nikon D1 digital camera was attached to a tripod. The height between the Calumet fluorescent light box (model 603-117P) and the camera lens was set to 74 cm. A composite specimen, with the carbon fibre veil facing toward the camera lens, was placed on a fixed location on the light box to ensure a constant illumination was attained. The images taken were analysed using image analysis software, KS 400 (Version 3), from Carl Zeiss Vision GmbH. The software identified the centre of the circular specimen and then determined the distribution of greyscale from black (0) to white (255) within a region of interest (ROI). The ROI was specified by an annulus with internal and outer diameters of 33.0 mm and 76.2 mm respectively. The centre of the annulus coincided with the centre of the circular specimen identified earlier. These dimensions and positions were selected because electromagnetic waves only propagated through the ROI in the EMI test. The greyscale level of each pixel within the ROI was then exported to an Excel spreadsheet for calculation of average, W and standard deviation, r Formation was then expressed in terms of coefficient ofpvariation, COV ðr=WÞ and normalised formation inffiffiffiffiffiffi dex, NFI ðr= W Þ. COV and NFI involve no sophisticated mathematical algorithms and are convenient and simple for the description of formation or degree of fibre dispersion of the veil. Both were used in this study due to their different capability in distinguishing dispersion quality. As reported by Helmer et al. [23], on

The EMI shielding effectiveness was measured according to the coaxial transmission line method based on ASTM D4935, 1999. An HP8752A network analyser with an input power of 5dBm was used to generate a plane-wave, far-field electromagnetic wave with a frequency ranging from 30 MHz to 1.5 GHz. The composite plaque was cut using an abrasive water-jet cutting machine into reference and load specimens. The reference specimen comprised a disc of diameter 33 mm and a circular ring of inner and outer diameters of 76.2 mm and 133 mm respectively. The load specimen had a diameter of 133 mm. One reference sample and at least two load samples were tested for each veil formulation. The reference specimen was first clamped in a nickel-coated brass specimen holder and the received power (P1) was then measured by the network analyser. The reference was then replaced with the load specimen and the received power (P2) was measure. The shielding effectiveness (SE) of the material was calculated according to Eq. (4).

SE ¼ 10logðP1 =P2 Þ

The entire system was calibrated by measuring standard specimens with known attenuation, which were supplied by Electro-Metrics, New York, US. The manufacturer part numbers of the standard specimens were X-K710-1, X-K710-2 and X-K711. The calibration was performed at frequencies ranging from 200 MHz to 1500 MHz and this frequency range was used for the present work for shielding measurement. On average, it was found that with our system, the

9

Number of fibre in percentage, %

ð4Þ

8 7 6 5 4 3 2 1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Fibre length, mm Fig. 4. Number average length distribution of recycled carbon fibre.

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

697

7

Weight fraction, %

6

5

4

3

2

1

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Fibre length, mm Fig. 5. Weight average length distribution of recycled carbon fibre.

measured shielding values of the standard specimens were 1.5% lower than the given manufacturer data, and the highest measurable shielding value using the existing equipment was about 80 dB.

on a petri dish for length distribution measurement. The results are shown in Figs. 4 and 5. The number and weight average length was 10.8 mm and 16.4 mm respectively. Table 1 shows the electrical resistivity of the virgin and recycled fibres. By performing a two-tailed hypothesis test based on a t-distribution, even at a high significant level of 10%, there is no statistically significant difference in resistivity between them. This suggests that heat-treatment during the fluidised bed process did not affect the fibre electrical resistivity.

3. Results and discussion

3.2. Veil formation

3.1. Characterisation of the fibre

Veil formation was quantified using a light-transmission setup and was expressed in terms of COV and NFI. The setup used was capable of measuring formation for veils up to 60 gsm as veils with heavier areal density were too dense for light to pass through. Fig. 6 shows the images for the composites containing either a

Table 1 Physical properties of virgin and recycled carbon fibre. Type of fibre

Diameter, lm

Electrical resistivity,  103 X cm

Virgin Recycled

7.50 ± 0.28 7.56 ± 0.21

1.83 ± 0.10 1.86 ± 0.10

Carbon fibre recycled from the fluidised bed process is fluffy and with some degree of agglomeration due to the swirling motion within the cyclone. The fibres were carefully separated and placed

Fig. 6. Images of veils.

698

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

(a)

3.5 3.0

6.4mm 19.2mm

9.6mm 28.8mm

14.4mm Recycled fibre

NFI

2.5 2.0

3.3. Effects of fibre loading on EMI shielding performance

1.5 1.0 0.5 0.0 20

40

60

Veil areal density, gsm

(b)

0.35 0.30

6.4mm 19.2mm

9.6mm 28.8mm

14.4mm Recycled fibre

COV

0.25 0.20 0.15 0.10 0.05 0.00 20

40

60

Veil areal density, gsm Fig. 7. Formation indices for virgin and recycled fibre veils.

layer of virgin fibre (14 mm long) or a recycled fibre veil at different areal densities. The latter, with the presence of fibre strings and bundles, is shown to have a poorer formation. Images of other specimens are not given here but their corresponding COV and NFI are plotted in Fig. 7. Overall, the results can be interpreted by making two types of comparison: intra-grammage and intergrammage. For the intra-grammage comparison, it can be seen that the two formation indices generally suggest that veils made of longer fibre have a poorer degree of fibre dispersion, with the poorest formation associated with the recycled fibre veil. This is in agreement with the veil-making experiment in that the difficulty in making a good fibre dispersion increases with fibre length. The recycled fibre veil contains a fraction of long fibres, as shown in Figs. 4 and 5, and these make dispersion more difficult. For the inter-grammage comparison, for each individual fibre length, the NFI value decreases with increasing fibre loading. This does not agree with observation as shown in Fig. 6; for example, where the formation of the recycled fibre veil degrades with higher loadings. The observed decrease in the formation of recycled fibre veils is in

Shielding effectiveness, dB

agreement with COV values for areal densities between 20 gsm and 40 gsm only. Above that, a similar error persists. It is thus concluded that the application of COV and NFI for formation analysis for the present study is limited to intra-grammage comparison only.

0.3

Glass fibre and polyester resin are electrical insulators and a composite made from them typically has an electrical resistivity between 1010 X m and 1012 X m [27]. Such a high resistivity means the GRP has negligible EMI shielding ability, as shielding performance is directly proportional to the material’s electrical conductivity. This agrees with the experimental results shown in Fig. 8. Consistent results were obtained from the three plain GRP specimens and the average shielding effectiveness over the whole frequency range was 0.12 dB. According to percolation theory, an electrically conductive network is formed when the volume fraction of carbon fibre is higher than a percolation threshold [14]. Below the percolation threshold, the composite behaves as an insulator or a semiconductor. In this investigation, fibres were filtered out from suspension by draining through a stainless-steel mesh to form a non-woven veil. The fibres in the veil were randomly oriented and physically interconnected. The Pike and Seager percolation model [28] was used to calculate the percolation threshold for a two-dimensional ‘‘stick” system with a random lattice configuration. The model was modified for this study with the threshold being expressed as the minimum veil areal density, b, at which a network of infinite size is first formed. The value of b was estimated using Eq. (5), where L was the carbon fibre length, q was the density of the carbon fibre and V was the physical volume of the fibre. Eq. (5) was used with assumptions that the fibres were straight, fully dispersed into single filaments and uniformly and randomly distributed across the veil. The result is tabulated in Table 2. The percolation threshold decreases with increasing fibre aspect ratio. The 6.4 mm-long fibre is estimated to require the heaviest veil density to form a conductive network, i.e. 0.062 gsm. As the lightest veil density in the present work was 20 gsm, which was much higher than the estimated 0.062 gsm, the EMI shielding ability of the GRP specimen would be enhanced by incorporating a layer of the electrically conducting carbon fibre veil.

Minimum veil areal density; b ¼

 2 1 4:236 ðqVÞ L p

ð5Þ

Fig. 9 shows the shielding results across the entire frequency range for GRP specimens containing a layer of 14.4 mm-long virgin carbon fibre veil with different fibre loadings. The shielding performance increases with fibre loading but is independent of frequency. Electromagnetic waves are attenuated by a shield in three different ways: reflection loss, absorption loss and multiple internal reflections [29]. The last factor can be neglected if the absorption loss is greater than 10 dB [30]. Reflection loss occurs when the wave passes through two different media and the degree of reflection is related to the difference in the characteristic impedance between each of them. Each veil had a fibre loading well

0.2

Table 2 The minimum veil areal density, with various fibre lengths, at which a conducting network of infinite size is first formed.

0.1

0 0

200

400

600

800

1000

1200

1400

Frequency, MHz Fig. 8. Shielding effectiveness of three plain GRP specimens.

1600

Carbon fibre length, mm

Aspect ratio

Minimum veil areal density, b, gsm

6.4 9.6 14.4 19.2 28.8

914 371 2057 2742 4114

0.062 0.041 0.027 0.021 0.014

699

60 50 40 30 20 10 20gsm

40gsm

60gsm

80gsm

100gsm

0 0

200

400

600

800

1000

1200

1400

1600

Frequency, MHz

Shielding effectiveness, dB

Fig. 9. EMI SE for GRP plates incorporating a layer of virgin veil (fibre length = 14.4 mm) of different areal density. 50 40 30 20 10 20gsm

40gsm

60gsm

80gsm

600

800

1000

1200

100gsm

0 0

200

400

1400

1600

Frequency, MHz

Average shielding effectiveness, dB

Fig. 10. EMI shielding results for GRP plates incorporating a layer of recycled fibre veil of different areal density.

60 50 40 30 20 Virgin fibre (14mm)

10

Recovered fibre

value, which is determined by averaging the SE values over the entire frequency range. Fig. 11 shows the average SE for the virgin (14 mm long) and recycled fibre veils. Veils with virgin fibre of this length were chosen as the length was closest to the weight average length of the recycled fibre. The relationship between fibre loading and SE is generally linear for the virgin fibre. However, the linear trend only applies for recycled fibre veil with areal densities up to 60 gsm, beyond which it levels off. At 20 gsm, the shielding performance of the recycled fibre veil was 13.7% lower than the 14 mm-long virgin fibre and the reduction reached 17.2% at 100 gsm. On average, the shielding performance of a layer of recycled fibre veil was about 11.9% lower than that of 14.4 mm-long virgin fibre. The lower shielding performance was not due to any change in the electrical conductivity of the recycled fibre and therefore may result from a reduction in fibre dispersion. From Fig. 7, it can be seen that for each fibre loading, the formation indices of the recycled fibre veil are higher and the differences in these indices between the recycled and virgin fibre increase with fibre loadings. Das et al. [16] suggested that a conductive filler should be closely packed to minimise void content that might degrade shielding performance. Similarly, Bushko et al. [32] also claimed that more electromagnetic waves would propagate through the resin-rich areas and thus degrade the overall veil shielding performance. The correlation between formation and shielding effectiveness will be discussed in the subsequent section. For practical applications, the SE value must be above 20 dB [32,33] and for most industrial and consumer applications, 30 dB is required [34]. In order to satisfy the FCC Class B requirement for commercial applications, the material used for making electronic housings should have an attenuation of at least 40 dB [35]. The same level of shielding effectiveness was also suggested by Jou et al. [36] for electronic appliances. However, for military applications, it was anticipated that the required SE should be between 80 dB and 100 dB [35]. The maximum SE obtained from the recycled fibre was about 40 dB from a layer of 80 gsm veil and thus the minimum requirement for commercial applications has been met with this material. However, it should be noted that there are always discontinuities in any practical shielding enclosure; for example, openings to accommodate ventilation, wiring and even personnel access as well as joints and seams between two sections. The eventual shielding effectiveness of an enclosure is thus affected not only by its constituent materials but also by the presence of these discontinuities [37,38].

0 0

20

40

60

80

100

120

Veil areal density, gsm

Fig. 11. Comparison of average EMI shielding values (single veil configuration) between virgin (fibre length = 14.4 mm) and recycled fibre veils.

above the electrical conductivity percolation threshold, hence their electrical conductivity should be of similar order of magnitude. The veil thickness increased with fibre loading, and the thickness of the veil is an important factor in determining the absorption loss. The amplitude of an electromagnetic wave decreases exponentially when passing through a conductive medium because of ohmic losses by induced currents in the medium [31]. With the absorption loss mechanism, the shielding effectiveness was increased with increasing fibre loading. Fig. 10 shows the shielding effectiveness in a GRP specimen made with a layer of recycled carbon veil of different areal density. Again, the SE is independent of frequency. By adding a 20 gsm veil, an increase in shielding effectiveness from 0 to around 25 dB can be seen from Figs. 8 and 10. The highest value attained was about 40 dB with a layer of a 80 gsm veil. Given that the results largely show frequency independence, the composite shielding performance is represented by an average

3.4. Effect of fibre length on EMI shielding performance Fig. 12 displays the effect of virgin carbon fibre lengths on shielding performance. Generally, all short fibres (between

Average shielding effectiveness, dB

Shielding effectiveness, dB

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

50 45 40 35 6.4mm 9.6mm 14.4mm 19.2mm 28.8mm

30 25 20 0

20

40

60

80

100

120

Veil areal density, gsm

Fig. 12. Average SE of GRP composite incorporating a layer of virgin fibre veil with various fibre lengths.

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702 5.0

0.30

4.5 0.25

4.0 3.5 3.0 2.5

0.15

NFI

2.0 0.10

1.5

0.05

1.0

COV NFI

0.5

90 80 70 60 50 40 30 20 10

D-20gsm S-20gsm

0 0

200

400

0.0

0.00 25

30

35

40

600

D-60gsm S-60gsm

800

1000

D-100gsm S-100gsm

1200

1400

1600

Frequency, MHz

45

Shielding effectivenss, dB Fig. 13. Scatter plot showing the relationship between SE, COV and NFI for veils with 40 gsm areal density.

6.4 mm and 14.4 mm) produced similar shielding performances. However, the 19.6 mm- and 28.8 mm-long fibres show lower shielding values. The conductivity of a composite or shielding effectiveness increases with increasing aspect ratio of the fibres [18,39,40]. However, when the fibre volume fraction was well above the percolation threshold, the influence of fibre length on the composite conductive properties is less significant [41]. This was supported by Shinagawa et al. [42], who found that EMI shielding did not depend on fibre length above a minimum threshold. As the veils used in this investigation had areal densities much heavier than the estimated percolation threshold, changes in fibre length between 6.4 mm and 14.4 mm did not seem to affect the shielding effectiveness. The reduction in effectiveness with longer fibres may thus result from poor fibre dispersion. 3.5. Veil formation and shielding performance The correlation between formation index and SE was performed using SPSS (Release 14) statistical software. A scatter plot for 40gsm veil is shown in Fig. 13. Direction and strength of the correlation was indicated by a Pearson’s product-moment correlation coefficient (expressed in terms of R). The p-value of the R-value was evaluated according to a two-tailed t-distribution inference test. Table 3 lists the corresponding R and p-value for areal densities ranging from 20 gsm to 60 gsm. Both the NFI and COV provide good correlations with shielding value. As the minimum magnitude of correlation coefficient is found to be greater than 0.5, a strong correlation between shielding performance and formation index is suggested [43]. The negative sign of the correlation indicates that shielding effectiveness decreases with increasing formation index. The correlation is also strongly supported by the associated p-value, which is <0.05 for all the cases. 3.6. Effects of veil configuration on shielding performance It was previously shown that a higher shielding performance could be achieved by increasing the fibre loading. However, this would result in higher material costs. An alternative way of increasing shielding performance could be to use two layers of veil,

Fig. 14. Comparison of EMI shielding results for single and double layer virgin veil configurations (S = single, D = double).

Shielding effectiveness, dB

COV

0.20

Shielding effectiveness, dB

700

80 70

Double layers

60 50

Single layer configuration

40 30 20 10 400MHz

800MHz

1200MHz

0 0

20

40

60

80

100

120

Net veil areal density, gsm Fig. 15. Comparison of EMI shielding results for single and double layer virgin veil configurations (fibre length = 6.4 mm) at 400 MHz, 800 MHz and 1200 MHz.

one on each side of a composite panel, to enhance the shielding performance through the creation of an extra reflection loss mechanism. Only virgin fibre veil with 6.4 mm-long fibre at 20 gsm, 60 gsm and 100 gsm were used due to the apparent insensitivity of shielding performance to fibre length, as long as good fibre dispersion is achieved. The shielding value was then compared with a single veil configuration with a similar net veil areal density and the result is presented in Fig. 14. The double veil configuration exhibits a higher shielding performance across the entire frequency range compared to the single veil layout at all areal densities. Unlike the single veil, the double veil configuration shows dependence on frequency with peak and trough points noted at various frequencies. This dependence is further illustrated in Fig. 15, which shows a clear enhancement in shielding performance with increasing fibre loading. Table 4 lists the maximum and minimum shielding values of each of the double veil configurations. The gains in shielding performance tabulated in Table 4 are attained by comparing the maximum and minimum shielding values of the double veil against the average value of the single veil layout respectively. The gain ranged between 12% and 80%. This increase may be partially attributed to the improvement in fibre dispersion, due to a lower fibre loading, but the main effect should be due to the double reflection loss. It is noticed that with two opposite layers of 50 gsm veils, the maximum gain attained is lower than with two pairs of 30 gsm veils. The reason for this is

Table 3 Correlation between formation indices and SE. Formation

NFI COV

20 gsm

40 gsm

60 gsm

Pearson correlation

p-Value

Pearson correlation

p-Value

Pearson correlation

p-Value

0.630 0.519

<0.01 <0.05

0.752 0.757

<0.001 <0.001

0.631 0.662

<0.01 <0.01

701

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702 Table 4 Comparison between single and double virgin veil configurations. Areal density, gsm

Single layer Average SE, dB

Min. SE, dB

Max. SE, dB

Min., %

Max., %

Virgin Virgin Virgin Recycled Recycled Recycled

20 60 100 20 60 100

31.13 40.68 47.31 25.95 38.62 39.59

35.10 46.81 52.97 31.55 40.85 45.88

50.98 73.14 78.40 44.58 61.06 70.91

12.75 15.07 11.96 21.58 5.77 15.89

63.76 79.79 65.72 71.79 58.10 79.11

Shielding effectiveness, dB

Type of fibre

90 80 70 60 50 40 30 20 D-20gsm S-20gsm

10

D-60gsm S-60gsm

D-100gsm S-100gsm

0 0

200

400

600

800

1000

1200

1400

1600

Frequency, MHz

Shielding effectiveness, dB

Fig. 16. Comparison of EMI shielding results for single and double layer recycled veil configurations (S = single, D = double).

70 60

Double layers configuration

50 Single layer configuration

40 30 20 10 400MHz

800MHz

1200MHz

0 0

20

40

60

80

100

120

Net veil areal density, gsm Fig. 17. Comparison of EMI shielding results for single and double layer recycled veil configurations (Lw = 16.4 mm) at 400 MHz, 800 MHz and 1200 MHz.

attributed to the sensitivity of the existing test rig which is limited to around 80 dB. Similarly, better shielding performance and frequency dependency are revealed from the recycled fibre with double veil configuration, as presented in Fig. 16. A continuous improvement in the shielding effectiveness with fibre loading for the double veil configuration at three different frequencies can be seen in Fig. 17. In contrast to this, the improvement levelled off for single veil with areal density heavier than 60 gsm. The corresponding gain in shielding performance can be seen in Table 4. With the double veil configuration, the minimum areal density required to meet the 40 dB threshold can be achieved by using two layers of 30 gsm recycled fibre veil, instead of 80 gsm for a single veil setup. 4. Conclusions The effects of fibre length, loading and veil configuration on the EMI shielding ability of GRP using virgin carbon fibre have been investigated. The shielding effectiveness seems to be unaffected by the length of virgin fibre (ranging from 6.4 mm to 14.4 mm),

Double layer

Gain in SE

as long as a network of interconnecting fibres with good fibre dispersion was established. Fibres longer than 19.6 mm were likely to become entangled during the fibre dispersion process and this reduced the shielding value due to non-uniformities in the veil areal density. A linear relationship between veil areal density and shielding effectiveness was observed. A double veil arranged in a sandwich configuration provided a higher shielding performance via two factors: the achievement of a better formation from the use of veils with a lower areal density and the presence of a double-wave reflection mechanism. However, unlike the single veil, the shielding ability depended on frequency. Despite this, at least 12% improvement was evident for the sandwich laminate. This gain even increased to 80% at certain frequencies without the penalty of extra material costs. The feasibility of EMI shielding using recycled carbon fibre has been studied and positive results have been achieved. Compared to a veil made from 14.4 mm-long virgin fibres, the recycled fibre veil has a 12% lower shielding performance. A linear correlation between fibre loading and shielding was evident up to 60 gsm only, beyond which a plateau was reached and small decrease was then observed. This might be attributed to the difficulty of achieving good dispersion with the recycled fibre due to the presence of long fibres. Nevertheless, an 80 gsm veil made from recycled fibre provided an EMI attenuation of 40 dB, which should be adequate to meet the FCC Class B requirement for commercial applications as suggested by Colaneri and Shacklette [35]. The shielding performance could be improved further by removing the long fibre portion from the recycled fibre length distribution in order to attain a better veil formation. Nevertheless, with a double veil configuration, higher shielding can be achieved from a lower fibre loading. Similar to the virgin fibre veil, a frequency-dependent shielding performance is observed from the recycled fibre under the double veil configuration. Acknowledgements The authors are grateful to the following organisations for the funding of this research: Engineering and Physical Sciences Research Council (EPSRC), Aston Martin Lagonda, BAE Systems plc, Cleanaway Ltd. and The Advanced Composites Group. Many thanks also to Dr. T. Wu from University of Nottingham for writing the KS 400 image analysis programme. References [1] Petterson J, Nilsson P. Recycling of SMC and BMC in standard process equipment. J Thermoplast Compos Mater 1994;7(1):56–63. [2] Schinner G, Brandt J, Richter H. Recycling carbon–fiber-reinforced thermoplastic composites. J Thermoplast Compos Mater 1996;9(3):239–45. [3] Fromonteil C, Bardelle P, Cansell F. Hydrolysis and oxidation of an epoxy resin in sub- and supercritical water. Ind Eng Chem Res 2000;39(4):922–5. [4] Hyde JR, Lester E, Kingman S, Pickering S, Wong KH. Supercritical propanol, a possible route to composite carbon fibre recovery: a viability study. Composites A 2006;37(11):2171–5. [5] Diem W. Still too exotic. Ward’s Auto World 2004;40(11):23. [6] Anonymous. Aircraft demand prompts Toray to boost carbon fibre capacity. Reinf Plast 2007;51(9):7.

702

K.H. Wong et al. / Composites: Part A 41 (2010) 693–702

[7] Nowaczek W. Polyester moulding compounds made from recycled thermosetting resins. Polym Recycl 2000;5(2):91–7. [8] Perrin D, Guillermain C, Bergeret A, Lopez-Cuesta J-M, Tersac G. SMC composites waste management as reinforcing fillers in polypropylene by combination of mechanical and chemical recycling processes. J Mater Sci 2006;41(12):3593–602. [9] Bream CE, Hornsby PR. Comminuted thermoset recyclate as a reinforcing filler for thermoplastics: Part I characterisation of recyclate feedstocks. J Mater Sci 2001;36(12):2965–75. [10] Kouparitsas CE, Kartalis CN, Varelidis PC, Tsenoglou CJ, Papaspyrides CD. Recycling of the fibrous fraction of reinforced thermoset composites. Polym Compos 2002;23(4):682–9. [11] Conroy A, Halliwell S, Reynolds T. Composite recycling in the construction industry. Composites A 2006;37(8):1216–22. [12] Ogi K, Shinoda T, Mizui M. Strength in concrete reinforced with recycled CFRP pieces. Composites A 2005;36(7):893–902. [13] Unser JF, Staley T, Larsen D. Advanced composites recycling. SAMPE J 1996;32(5):52–7. [14] Vilcakova J, Saha P, Kresalek V, Quadrat O. Pre-exponential factor and activation energy of electrical conductivity in polyester resin/carbon fibre composites. Synth Met 2000;113(1-2):83–7. [15] Luo X, Chung DDL. Electromagnetic interference shielding using continuous carbon–fiber carbon-matrix and polymer–matrix composites. Composites Part B: Eng 1999;30(3):227–31. [16] Das NC, Chaki TK, Khastgir D, Chakraborty A. Electromagnetic interference shielding effectiveness of conductive carbon black and carbon fiber-filled composites based on rubber and rubber blends. Adv Polym Technol 2001;20(3):226–36. [17] Pickering SJ, Kelly RM, Kennerley JR, Rudd CD, Fenwick NJ. A fluidised-bed process for the recovery of glass fibres from scrap thermoset composites. Compos Sci Technol 2000;60(4):509–23. [18] Huang CY, Chiou TW. The effect of reprocessing on the EMI shielding effectiveness of conductive fibre reinforced ABS composites. Eur Polym J 1998;34(1):37–43. [19] Walker NJ. Papers – a new dimension in carbon fiber materials. In: Carbon fibres: technology, uses and prospects; 1986. p. 55–63. [20] Richardson PF, Dunn SN, Wilharm MT, Nielson SF. Effect of acrylamide copolymers on wet laid sheets produced from high l/d ratio polyester fibers. In: TAPPI proceedings – nonwovens conference. Nashville, TN (USA); 1988. p. 57–64. [21] Kennerley JR. Recycling fibres recovered from composite materials using a fluidised bed process. PhD Thesis, Department of Mechanical Engineering, University of Nottingham; 1998. [22] Hull D, Clyne TW. An introduction to composite materials. Cambridge: Cambridge University Press; 1996. [23] Helmer RJN, Covey GH, Raverty WD, Vanderhoek N. Optimising simulated commercial paper for pulp quality analysis. Appita J 1987;54(3):289–95. [24] Cresson T, Luner P. The characterisation of paper formation. Part 2: The texture analysis of paper formation. Tappi J 1990;73(12):175–85.

[25] Nazhad MM, Harris EJ, Dodson CTJ, Kerekes RJ. The influence of formation on tensile strength of paper made from mechanical pulps. In: Tappi papermakers conference and trade fair. Vancouver, BC (Canada); 2000. p. 141–6. [26] Ichiura H, Nozaki M, Kitaoka T, Tanaka H. Influence of uniformity of zeolite sheets prepared using a papermaking technique on VOC adsorptivity. Adv Environ Res 2003;7(4):975–9. [27] Dudgeon CD. Polyester resins. In: Dostal CA, editor. Engineered materials handbook. Composites. Ohio: ASM International; 1987. p. 90–6. [28] Pike GE, Seager CH. Percolation and conductivity: a computer study. I. Phys Rev B 1974;10(4):1421. [29] Farsi M. EMI/RFI shielding: theory and technique. Interference Technology Engineer’s Master 1988:228–38. [30] Chiang WY, Chiang YS. Effect of titanate coupling agent on electromagnetic interference shielding effectiveness and mechanical properties of PC–ABS–NCF composite. J Appl Polym Sci 1992;46(4):673–81. [31] Ott HW. Noise reduction techniques in electronic systems. New York: John Wiley and Son; 1988. [32] Bushko WC, Stokes VK, Wilson J. EMI shielding effectiveness of fiber-filled plastics – material testing issues. In: ANTEC. New York (USA); 1999. p. 1499– 502. [33] Bigg DM, Mirick W, Stutz DE. Measurement of EMI shielding of plastic composites using a dual chamber facility. Polym Test 1985;5(3):169–81. [34] Bigg DM. Effect of chemical exposure on the EMI shielding of conductive plastics. Polym Compos 1987;8(1):1–7. [35] Colaneri NF, Shacklette LW. EMI shielding measurements of conductive polymer blends. IEEE Trans Instrum Meas 1992;41(2):291–7. [36] Jou WS, Wu TL, Chiu SK, Cheng WH. The influence of fiber orientation on electromagnetic shielding in liquid–crystal polymers. J Electron Mater 2002;31(3):178–84. [37] Miyake S, Umezu Y, Sagawa Y, Morita T, Yoshino R. Investigation related to construction method and performance of an electromagnetic shielded enclosure. In: IEEE international symposium on electromagnetic compatibility. Cherry Hill, NJ (USA); 1991. p. 120–5. [38] Ghose S, Pande DC. Identification of electromagnetic leakage from seams and joints and their prevention. In: 8th International conference on electromagnetic interference and compatibility. Chennai (India); 2003. p. 339–42. [39] Bigg DM, Stutz DE. Plastic composites for electromagnetic interference shielding applications. Polym Compos 1983;4(1):40–6. [40] Lee BO, Woo WJ, Park HS, Hahm HS, Wu JP, Kim MS. Influence of aspect ratio and skin effect on EMI shielding of coating materials fabricated with carbon nanofiber/PVDF. J Mater Sci 2002;37(9):1839–43. [41] Liang X, Ling L, Lu C, Liu L. Resistivity of carbon fibers/ABS resin composites. Mater Lett 2000;43(3):144–7. [42] Shinagawa S, Kumagai Y, Urabe K. Conductive papers containing metallized polyester fibers for electromagnetic interference shielding. J Porous Mater 1999;6(3):185–90. [43] Green SB, Salkind NJ, Using AkeyTM, SS SP. For windows: analyzing and understanding data. New Jersey: Prentice Hall; 2000.