Enhanced electrothermal effect of carbon fibrous composites decorated with high conductive carbon nanotube webs

Bingcheng WANG,Xu FENG,Ling LIU

School of Aerospace Engineering &Applied Mechanics,Tongji University,Shanghai 200092,China

KEYWORDS Composites;Carbon Fiber Reinforced Plastics (CFRPs);Carbon nanotubes;Electrical conductivity;Electrothermal effect;Deicing

Abstract Literature has demonstrated that Carbon Nanotubes (CNTs) can greatly enhance the electrical conductivity and matrix-dominated mechanical properties of fibrous composites.However,electrothermal coupling effect of CNTs on Carbon Fiber Reinforced Plastics (CFRPs) has scarcely been considered.This work prepared and introduced thin and porous CNT webs to the surface or/and interface of a CFRP to enhance its electrothermal properties.The results show that CNT webs can enhance the transverse electrical conductivities of the CFRP by 231%-519% in a current range of 50–150 mA,when compared to the base-CFRP.Also,the surface temperature of CNT webs decorated CFRP can be improved by 20.5–32.3 °C within 3 min showing a self-heating rate of 6.8–10.8 °C/min just with an applied voltage of 20–30 V,increased by 152%-177% when compared to the base-CFRP(2.7–3.9°C/min).Also,deicing can be finished within 4–10 min with a voltage of 18 V and an input power of 246 W/m2.Moreover,the electrothermal processes nearly have no negative effect on the mechanical properties of the CFRP.The relatively low input power and short response time for deicing make the CNT webs decorated CFRP may be a potential new generation for aeronautical deicing structure.

1.Introduction

Due to the high specific strength and stiffness as well as good corrosion resistance and fatigue resistance,Carbon Fiber Reinforced Plastics(CFRPs)have been widely used in airplane components such as wing skins,fuselages and tails.1–3However,freezing is one of the extreme service-environments that are occasionally faced by airplanes.The induced icing will worsen the aerodynamic characteristics,decrease the lift-drag ratio or the flight envelope to shrink,eventually reduce the flight performance and endanger the flight safety.4–6Therefore,deicing is always one of the focus points in the design of aeronautical structures.Among the proposed various deicing methods,4,7–9thermal deicing especially designing an electrothermal layer has been demonstrated as an effective method.10,11For example,early used metal electrothermal layers could effectively deice with advantages of easy executing,short response time and structural–functional integration.4However,with the development of CFRP wings,the use of traditional metal deicing layers would cause some serious problems,such as mismatch of thermal expansion coefficients between metals and CFRPs,or electrochemical corrosion of the metals.12To overcome these inevitable problems,conductive graphite or carbon was used to replace the metals to form new electrothermal deicing layers,13–16which can not only avoid the mismatch between metals and CFRPs or the electrochemical corrosion problems of metals,but also successfully reduce the weight of the structures.17Nevertheless,these kinds of electrothermal elements need special insulation treatment and then can be adhered with adjacent structures.15,16Most importantly,the structures and the deicing layers usually execute their functions individually.Thus,these limitations restrict their wide applications.

Since the discovery of Carbon Nanotubes(CNTs),18,19their usages have been extensively explored in some potential application fields such as materials,mechanics,bioengineering and aerospace20–22due to their super strength (about 50 GPa),excellent electrical conductivity (105-106S/m),low density(about 1.3 g/cm3) and very low percolation threshold in polymer composites (about 0.1vol%).23–26The excellent electrical conductivity of CNTs provide a wide application in developing a new type of sensors,27–29new electromagnetic materials,30,31and self-heaters32–37as well as a new deicing technology based on electrothermal coupling effect (Joule heating effect).Reports have shown that the electrical conductivity of CFRPs through-the-thickness can be improved by about several folds or several orders of magnitude38–42by introducing a little amount of CNTs into epoxy matrix,where the electrical conductivity highly depends on the number of conductive paths and junctions formed by the CNT networks.43–45This means that the denser the conductive paths,the higher the electrical conductivity.Therefore,a continuous and dense conductive network,CNT web,was prepared by Han et al.17and used to coat on the surface of a CFRP,showing that the surface electrical conductivity of a CFRP could be improved from original 2.3 × 10-10S/cm to 4.0 × 10-3S/cm and simultaneously accompanied by a weight reduction of 30%.Also,the in-plane electrical conductivity of the neat CNT webs could reach 1.53 × 102S/cm.46

The high electrical conductivity of CNTs or CNT webs modified CFRPs shows potential high electrothermal or selfheating performance.47,48Research indicated that CNT/polymer heaters can self-heat to 140 °C,which is improved by about 133% when compared with traditional platinum heaters.49Chu et al.50found that the electrical conductivity of a Glass Fiber Reinforced Plastic (GFRP) interleaved with CNT webs could reach 64.9 S/cm and deicing would be accomplished within 3.7 min with an input power of 3733 W/m2.Brampton et al.51designed and fabricated an asymmetric CFRP laminate interleaved with aligned CNT webs,and the surface temperature of the CFRP can be heated to 34 °C within 28 s when a voltage of 5 V and an input power of 492 W/m2are applied.Buschhorn et al.52introduced CNT webs to the surface of a modal wing,revealing that CNTs has distinguished potential in the field of electrothermal deicing.Yao et al.53incorporated 20 layers of CNT webs into the middle interface of a CFRP and found that the CNT webs could decrease the electrical resistance of the CFRP by 25%.When the applied current was perpendicular to the fiber direction,the surface temperature could arrive 60 °C within 5 min at an input power of 1143 W/m2.They also found that if 40 layers of CNT webs are incorporated into a GFRP,the deicing could be accomplished within 15 s with a constant 4900 W/m2power supply,while if the CNT webs are replaced by two layers of carbon fibers,the obtained deicing time is 135 s.54A recent research carried out by Tarfaoui et al.55also reported that by introducing a special shaped CNT web to the surface of a GFRP,whose surface temperature could be heated to 28°C at a voltage of 6 V and an input power of 63 W/m2,deicing requirement could be realized within 1 h.The aforementioned studies show that the introduced CNT webs could significantly improve the electrothermal performance of CFRPs or GFRPs due to the enhanced electrical conductivity and the formed denser conductive paths.Moreover,functionalized CNTs will co-bond well with the polymer matrix and highly integrate with the CFRPs or GFRPs,which reveals that CNTs could execute their functional-structural performances simultaneously.But previous studies mainly introduce oneor multi-layer of CNT webs either to the middle interface or to the surface of a CFRP or a GFRP,and usually need a relatively large input power such as greater than 1000 W/m2to completely deice within 6 min,or a small input power such as 63 W/m2to finish deicing within 1 h.42–46Also,how to connect wires,which material direction is the optimal direction of input power,as well as whether the input power correlates with the material directions,have scarcely been considered until now.

Thus,this paper aims to explore how the introduced positions of CNT webs affect the electrical conductivity,electrothermal self-heating rate and corresponding mechanical properties of a CFRP,in which effect of material directions is also considered.Firstly,the CNT webs were prepared and respectively introduced to the surface or/and interface of a CFRP to prepare various CFRP laminates.Then,the electrical conductivities depending on the axes of the CFRP laminates were measured under different currents.Based on the measured electrical conductivities results,the effects of CNT webs on the electrothermal self-heating rate of the CFRP were analyzed comparatively.Simultaneously,deicing was carried out to demonstrate the self-heating performance of the CNT webs modified CFRP.Finally,electrothermal effect on the transverse tensile mechanical properties of the CFRP was also evaluated.

2.Experiment

2.1.Materials

The used multi-walled CNTs (8–15 nm in diameter and 5–15 μm in length) were purchased from Chengdu Organic Chemistry Co.Ltd.of China.CNTs were originally obtained as suspensions in N,N-dimethylformamide with a weight loading about 2.0%.A carbon fiber/epoxy prepreg lamina(Shenying Carbon Fiber Co.Ltd.,China) was utilized to prepare the CFRP laminates,in which the carbon fiber is named as SYT49-12 K (Shenying Carbon Fiber Co.Ltd.,China) and the epoxy resin is a typical bisphenol-A with a brand of YPH-69 (Kunshan Yubo Composite Material Co.Ltd.,China).The density,monolayer thickness and fiber volume fraction of the prepreg are supplied as 476 g/cm3,0.30 mm and(63 ± 2)%,respectively.

2.2.Preparation

The CNT suspension in Section 2.1 was filtered in a pressure tank with an aid of 0.1–0.2 MPa and a filter membrane with a pore size of 0.8 μm,as shown in Fig.1(a).The as-prepared CNT web was immersed in nitric acid about 1 min to remove surfactant,and then washed with distilled water and dried at 60 °C for 1 h.The thickness of the final obtained CNT web is 25–35 μm with an average diameter about 280 mm (see Fig.1(b)),which looks rather smooth and whose corresponding SEM image is given in Fig.1(c).It suggests that the CNT web is microscopically porous and CNTs distribute uniformly within the web.This means that the epoxy resin could flow into the porosities during the curing process of the CFRP and finally form an integrated structure with the CFRP.

According to previous studies,47–56electrothermal CFRPs can be designed as co-cured structure,in which the electrothermal layer is coated on one surface of the structure or interleaved at its interface (Fig.1(d)),that is,the electrothermal layer is just embedded nearby the surface area of a structure.Based on this,here three kinds of CFRP laminates,named as Base-CFRP (without CNT web),S-CFRP (with one CNT web co-cured on the surface of the CFRP,①in Fig.1(d)),and S/I-CFRP (with two CNT webs respectively introduced to the surface and its adjacent interface of the CFRP,①and②in Fig.1(d)),were designed and prepared.In Fig.1(d),the CFRP laminates consist of 4-layer unidirectional prepreg laminas.To easily compare the measured results,a coordinate system is defined,in which,x-,y-,andz-axis respectively correspond to the axis of carbon fibers,the transverse direction of carbon fibers,and the thickness direction (see Fig.1(d)).It is worth noting that the Base-CFRP was prepared mainly for comparison.These CFRP laminates were then cured in a hot press,firstly heated up to 80 °C in a heating rate of 2–3 °C/min and kept for 30 min,then continuously heated to 130 °C and held for another 90 min,and finally cooled down to room temperature.The applied pressure was set as 0.6 MPa during the whole curing process.The cured thickness of these three CFRP laminates is measured about 1.21–1.23 mm,suggesting that one or two CNT webs have little effect on the thickness of the cured laminates.

2.3.Characterization

The morphologies of the cured laminates were characterized by using a SEM machine (Philips XL 30 ESEM).Sidesections and surfaces of the cured CFRP laminates with CNT webs were observed at an applied voltage of 20 kV in order to estimate the interfacial bonding between CNTs and epoxy matrix or between CNT web interlayer and adjacent CFRP layer.

It is commonly known that the electrical conductivity of CFRPs highly depends on the axis of carbon fibers,and various conductivities would certainly cause different self-heating effect.Previous investigations50–56scarcely considered effect of material directions on the electrical conductivity of CNTs decorated CFRP laminates.Therefore,relatively large(200 mm × 200 mm) CFRP specimens were firstly cut from the cured laminates and used to perform the measurements.For each CFRP laminate,three specimens were cut and tested for evaluating the repeatability of the measurement.Electrical conductivities along thex-axis(shown in Fig.1(e)),y-axis andz-axis (similar to Fig.1(e)) were simultaneously measured by using two-probe method and calculated according to Ohm’s law.In each case,input end (+) and output end (-) of each specimen were polished to make partial carbon fibers exposed to eliminate contact resistance,as shown in Fig.1(e).Then,conductive silver glue was coated on the exposed carbon fibers and filled the polished zones completely,where copper wires were drawn out and connected to a Keithley 2450 source meter.During the measuring processes,various electrical currents(A)were applied to find the changing trend of the electrical conductivities.

Fig.1 Preparing CNT web and CFRP laminates as well as measuring electrical conductivities and deicing.

Based on the testing results of the electrical conductivity,a reasonable measuring direction with a proper electrical conductivity will be chosen to perform the following electrothermal effect.It should be noted that improvement of electrothermal self-heating rate is based on reducing the electrical resistance of the CFRP laminates,and it is more efficient to generate heat when a voltage is applied during the measurement process.15,50,54Therefore,the surface temperatures of the CFRP laminates versus applied voltages were measured.It is found that the surface temperature of the laminates tends to be stable within 3 min,and therefore,the stable temperature field was recorded by using a thermal infrared camera FLIR ONE Pro.Also,from 0 to 3 min,temperatures at the surface center of the laminates were continuously measured by TASI-8620 thermocouple,which were used to calculate the self-heating rate.

After that,deicing tests were performed based on the above electrothermal self-heating results.A sealing tape with a thickness of 3 mm was used to form a 180 mm×180 mm groove on the surface of the laminate,as shown in Fig.1(f).Then the groove was filled with 100 mL of water and frozen in a refrigerator for 10 h in order to form a layer of ice.Certainly,the thickness of the ice layer is 3 mm.Based on the electrothermal self-heating results,a moderate voltage of 18 V was applied along they-axis of the laminates.Times required for the ice layer beginning to separate from the CFRP surface as well as the ice completely melting into water were measured.

At last, just when the pain was so great that Medio Pollito thought he must die, the cook lifted up the lid of the pot to see if the broth was ready for the King s dinner

Usually,electrothermal process may decrease the interfacial bonding strength between the carbon fibers and epoxy matrix,which can be evaluated by using transverse tensile testing (90°tension ory-axis) of the CFRP.Another reason why the 90°tensile testing is chosen is that the obtained optimal electrothermal rate and deicing efficiency occur along they-axis of the CFRP laminates.Therefore,to evaluate the electrothermal effect on the mechanical properties of CFRP laminates,three thicker laminates with 8-layer CFRP prepregs were also prepared,which are similar to the Base-CFRP,S-CFRP and S/I-CFRP,respectively.The only difference is that the additional layers of CFRP prepregs were laid up on the bottom side (or opposite side of CNT web location)of the laminates to satisfy the thickness requirement of 90° tensile testing according to ASTM D3039 standard.These three laminates were cured using the aforementioned cure process.Dimensions of 90°tensile specimens are 175 mm×25 mm× (2.41–2.53 mm)(thickness varies with different laminates),strengthening with end tabs (25 mm × 25 mm × 1.5 mm).The 90° tensile tests were conducted on a universal testing machine (Z250,Zwick/Roll)with a load rate of 2 mm/min.During the testing,various voltages of 0,10,20,30 V were respectively applied to the specimens to determine the effect of electrothermal process on the 90°tensile mechanical properties of the CFRP.Also,for evaluating the repeatability of the tested results,3–5 specimens were measured for each laminate.

3.Results and discussion

3.1.Microscopic infiltration state of CNT webs

SEM images of the surficial and interfacial CNT webs cocured with the CFRP are given in Fig.2.Fig.2(a) shows that the CNT web is impregnated well with epoxy matrix and seems attached tightly on the surface of the CFRP.Some small nanoscale shallow pits are visible on the surface.A zoomed zone of Fig.2(a) is given in Fig.2(b),suggesting that CNTs (even located within/near the pits) are co-bonding well with epoxy and there are no visible gaps between the walls of CNTs and the epoxy,revealing good interfacial bonding.CNTs within the CNT web connect and overlap with each other,which mean that uniform conductive networks and dense conductive paths are formed by introducing CNT web to the surface of the CFRP.When the CNT web is interleaved into the interface of the CFRP,as shown in Fig.2(c),a significant interlayer is formed with a relatively constant thickness about 38 μm.Meanwhile,the formed CNT interlayer is highly integrated with adjacent CFRP layers and indicated as a perfect microstructure.A close view of the CNT interlayer is given in Fig.2(d),and it also demonstrates that CNTs are cobonding well with epoxy resin and distribute uniformly.This also suggests that a uniform conductive network is formed at the interface of the CFRP.Therefore,Fig.2 proves that both the surficial and interfacial CNT webs have formed an integrated structure with the CFRP.Simultaneously,the formed CNT conductive paths may significantly improve the electrical conductivity of the CFRP.

Fig.2 SEM image of CNT web infiltrated with epoxy on surface and interface of CFRP.

3.2.Electrical conductivities of laminates

Three-Dimensional (3D) electrical conductivities of the CFRP laminates are presented in Fig.3.It can be seen that electrical conductivities are highly correlative with the material directions (x-,y-andz-axis) and the introduced positions of CNT webs (surface,interface or both),as shown in Fig.3(a)-(c).They gradually increase with the increased currents,in which the maximum conductivity occurs along thex-axis direction or in the case of CNT webs introduced to both the surface and interface of the CFRP.Moreover,electrical conductivities are all improved no matter where the CNT webs are introduced.As for the S/I-CFRP,the obtained conductivities along thex-andz-axis are respectively increased by 8%-41% and 88%-100% in the current range of 600–2000 mA when compared to the base laminate,while the gain alongy-axis is improved by 231%-519% in a current range of 50–150 mA when compared to the base laminate.The increment of S/ICFRP is much more than that of S-CFRP,that is,the more the CNT webs,the higher the electrical conductivities.

Fig.3 Electrical conductivities of CFRP laminates and an approximate conductive mechanism.

As can be seen from Fig.3(d),when measuring the electrical conductivity along thex-axis,all the carbon fibers within each lamina align along thex-axis and form in-plane parallel circuits,and simultaneously various laminas also form parallel circuits.In addition,the electrical resistances along the axis of carbon fibers are the minimum.Therefore,the measured electrical conductivities of the laminates along thex-axis are the highest.When the CNT webs are introduced to the surface or/and interface (Fig.3(d)),they increase the resistance number of parallel circuits,and thus the electrical conductivities are enhanced significantly(see Fig.3(a)).If measuring the conductivity along they-axis direction,the transversal resistances of the carbon fibers are relatively large,the large transversal resistances further form series circuits in each lamina,thus the inplane resistance is relatively larger along they-axis direction,and then the large in-plane resistances of multi-layer CFRP laminas and CNT webs further form out-of-plane parallel circuits (Fig.3(d)).Therefore,the total resistance is larger than that along thex-axis,i.e.the conductivity decreases along they-axis direction.However,along thez-axis direction,the carbon fibers are parallel within the in-plane,and form series circuits through-the-thickness direction.But the current flows from the entire upper surface to the entire bottom surface,i.e.the cross-sectional area of the total resistance is large enough,and an increased cross-sectional area will decrease the resistance according to Ohm’s law.Therefore,the obtained conductivity is larger than that along they-axis direction but less than that along thex-axis direction.As mentioned above,the obtained 3D electrical conductivities roughly increase with the increased currents.This is because the 3D electrical conductivities of the CFRP with and without CNT webs are all belong to the range of semi-conductors.However,as for the semi-conductors,the input current is not proportional to the input voltage,that is,the electrical conductivity increases with the increase of current.Therefore,the obtained results of CNT webs modified CFRP satisfy this common principle.

3.3.Electrothermal self-heating rate

In fact,it is more convenient to connect the conductive wires along the in-plane direction in practical applications.Therefore,the electrothermal self-heating rate of the CFRP is performed by inputting various voltages along thex-andy-axis.Fig.4(a) gives the temperature increment (ΔT) at the surficial center of the CFRP laminates versus time when a voltage of 1.2 V is applied along thex-axis.The original room temperature is 23.0°C,and thus ΔTis the difference between the tested temperature and the original temperature of 23.0°C.It can be seen from Fig.4(a) that ΔTof all the laminates tend to be stable within 3 min and the obtained electrothermal selfheating rate is 1.0 °C/min for the Base-CFRP and S-CFRP,and 2.3 °C/min for the S/I-CFRP,respectively.Certainly,the electrothermal self-heating rate will increase with the increased voltages,but the time corresponding to the stable temperature is found as 3 min no matter what the voltage is.The Base-CFRP and S-CFRP show similar electrothermal rate along thex-axis,because their conductivities are almost equal to each other.However,the electrothermal rate of the S/ICFRP is improved by 130% when compared to the Base-CFRP due to the introduced two CNT webs.That is,introducing two CNT webs would significantly improve the self-heating performance of the CFRP.This is because the input electric power (U2/R,whereUis the voltage,andRis the resistance)varies inversely withR(namely the Joule effect).Introducing CNT webs will decreaseRof the CFRP to some degree,and thus improve the self-heating performance of the CFRP.Fig.4(b) gives ΔTof the laminates depending on different voltages along thex-axis,where ΔTis the stable temperature increment (after electrifying 3 min).It can be seen that ΔTincreases continuously with the increased voltages due to the Joule effect,and thus self-heats the CFRP.Similarly,the SCFRP and S/I-CFRP show higher electrothermal rates than the Base-CFRP when the voltage is greater than 1.5 V.When the voltage reaches 3.2 V,ΔTis obtained as 13.5,17.6,27.4°C for the Base-CFRP,S-CFRP and S/I-CFRP,respectively,showing electrothermal rates of 4.5,5.9,9.1 °C/min,respectively.That is,the electrothermal rate of the S/I-CFRP is improved by 102% at the voltage of 3.2 V when compared to the Base-CFRP.It should be noted that the rated current of the source meter is limited as 3 A,and therefore,the input voltage is relatively small due to the lowRof the CFRP.In addition,Fig.4(c)gives the corresponding surface temperature distributions of the laminates at voltages of 1.6,2.4,3.2 V applied along thex-axis.It shows that the more the CNT webs,the more uniform the temperature distribution.This reveals that the CNT webs enhance the uniformity of the selfheating area.

Fig.4 Electrothermal effect of laminates along x-axis.

Fig.5(a) shows the electrothermal rate of the laminates along they-axis at a voltage of 6.0 V,which suggests that the CNT webs modified CFRP also reaches the maximum temperature within 3 min.The Base-CFRP nearly keeps unchanged during the electrifying time.At this level of voltage,the obtained self-heating rates within 3 min are calculated as 1.0 °C/min and 2.3 °C/min for the S-CFRP and S/I-CFRP,respectively.The S/I-CFRP also shows the highest selfheating rate at they-axis direction,which is improved by 666% and 130% when respectively compared with those of the Base-CFRP and the S-CFRP.Likewise,Fig.5(b)indicates ΔTvarying with the voltages,which shows that the electrothermal self-heating rate increases significantly with the increased voltages due to the Joule effect,in which,the S-CFRP shows a moderate self-heating rate and the S/I-CFRP shows the best electrothermal rate.At a voltage of 20–30 V,ΔTobtained from the S/I-CFRP is 20.5–32.3 °C,showing an self-heating rate of 6.8–10.8 °C/min,improved by 152%177% when compared to the Base-CFRP (2.7–3.9 °C/min,ΔT=8.0–11.7 °C).However,the gain in the S-CFRP is about 44%-105% (ΔT=11.5–24.0°C).The results mean that the CNT webs can greatly improve the electrothermal self-heating rate of the CFRP by introducing one or two CNT webs.It can be seen from Fig.5(a) and (b) that the obtained ΔTalong they-axis direction is much higher than that obtained along thex-axis direction(see Fig.4(a)and(b)).This is because the input voltage is significantly increased and the Joule heat is proportional to the square of the voltage.Fig.5(c)also gives some typical temperature distributions of the laminates.It shows that the temperature fields along they-axis direction are more uniform than those along thex-axis (see Fig.4(c)).

Fig.5 Electrothermal effect of laminates along y-axis.

It can be concluded from Figs.4 and 5 that introducing CNT webs can significantly enhance the electrothermal selfheating rate of the CFRP,especially in the in-plane transverse direction.In other words,Joule heat effect caused by CNT webs is more effective than the carbon fibers.Moreover,it may be a more reasonable choice to simultaneously introduce the CNT webs to the surface and its near interface of the CFRP.Furthermore,ΔTcan reach 20.5–32.3 °C just within 3 min when a voltage of 20–30 V is applied,which has the potential to be used as an electrothermal deicing system.

3.4.Deicing performance

The above results have demonstrated that the in-plane transverse direction of the CFRP shows the optimal electrothermal self-heating rate.In order to avoid a high temperature decreasing the mechanical properties of the CFRP,57,58here a moderate voltage of 18 V is applied to the CNT webs decorated CFRP to evaluate its deicing performance.Fig.6 shows the time required for the ice layer beginning to separate from the CFRP surface and to completely melt into water during the deicing process.It shows that the required time for deicing is basically consistent with the electrothermal effect trend shown in Fig.5(b).The Base-CFRP needs 15 min to realize ice separating from the CFRP surface and 33 min to completely melt the ice into water,as shown in Fig.6(a)and(b).However,the gains for the S-CFRP are respectively read as 7 min and 15 min (Fig.6(a) and (b)).As for the S/I-CFRP,the corresponding time is measured as 4 min and 10 min,respectively,for ice layer separating from CFRP surface (Fig.6(c)) and ice completely melting into water (Fig.6(d)).

Fig.6 Deicing process of CFRP decorated with CNT web.

3.5.Electrothermal effect on transverse tensile properties

Electrothermal process may decrease the interfacial bonding strength between the carbon fibers and epoxy matrix.Usually,transverse tensile testing could be conducted to evaluate the interfacial bonding properties of CFRP.Therefore,90°tension of unidirectional CFRP with and without CNT webs is carried to evaluate the electrothermal effect on the transverse mechanical properties of the studied CFRP.As shown in Fig.7,transverse tensions of Base-CFRP,S-CFRP and S/I-CFRP at voltages of 0,10,20,30 V are presented in Fig.7(a)-(d),and the given stress–strain curve for each laminate and each case is the one which is close to the average result (It should be noted that three specimens are tested in each case).In Fig.7(b)-(d),corresponding temperature distribution within the gauge length of the testing specimens for each laminate is also given.It can be seen that the temperature along the length of the specimens distributes uniformly and the obtained temperatures are very close to those shown in Fig.5(c).Also,the stress–strain curves of these three laminates under each voltage almost approach to each other.It reveals that the CNT webs and the applied voltages (or electrothermal effect) have little effect on the interfacial or transverse mechanical properties of the CFRP.The calculated tensile strengths and moduli of these three laminates are shown in Fig.7(e),suggesting that they are nearly equal to each other,and the deviation is basically within the range of error.

When the CNT webs are simultaneously introduced to the surface and interface of the CFRP,it will slightly improve the transverse tensile mechanical properties because the interfacial CNTs will strengthen the interlaminar resin rich zone (shown in Fig.2(c) and (d)).Also,some CNTs within the interface would penetrate into adjacent CFRP layer with the flow of resin during the cure process and attach on the surface of carbon fibers (Fig.7(f)).Therefore,local interfacial bonding between carbon fibers and epoxy can be improved by the CNTs,which would further enhance the 90°tensile mechanical properties of the unidirectional CFRP.Thus,the strengths and moduli of the S/I-CFRP are slightly higher than those of the Base-CFRP and S-CFRP (below 65 °C).As for the S-CFRP,since the thin CNT web is mainly attached on one surface of the CFRP,it nearly has no effect on the mechanical properties of the CFRP.

Fig.7 Transverse stress–strain curves,tensile strengths and moduli and typical fractured surfaces of CFRP with and without CNT web under various voltages.

4.Conclusions

CNT webs are introduced to the surface or/and interface of a CFRP,and how the introduced CNT webs affect the electrical conductivities,electrothermal self-heating rate,deicing capability and transverse mechanical properties of the CFRP is investigated systematically.

(1) The electrical conductivities of the CFRP are highly correlative with the material directions and the introduced number of CNT webs.The maximum conductivity occurs along the carbon fiber axis or in the case of CNT webs introduced to both the surface and interface of the CFRP.With CNT webs,the electrical conductivities of the CFRP along thex-axis,z-axis,andy-axis can be improved by about 8%-41%(600–2000 mA),88%-100% (600–2000 mA) and 231%-519% (50–150 mA),respectively,when compared to the Base-CFRP.

(2) The CNT webs decorated CFRP shows higher electrothermal self-heating rate.Along thex-axis,temperature increment can reach 27.4 °C within 3 min,showing a self-heating rate of 9.1 °C/min just with an input voltage of 3.2 V for the S/I-CFRP,which is improved by 102% when compared to the Base-CFRP.If a voltage of 20–30 V is applied along theyaxis,temperature increment of the S/I-CFRP would arrive 20.5–32.3 °C within 3 min,showing a selfheating rate of 6.8–10.8 °C/min,which is improved by 152%-177% when compared to the Base-CFRP (2.7–3.9 °C/min).Moreover,when a voltage of 18 V is applied along they-axis,the S/I-CFRP will realize to separate the ice layer from the surface of CFRP within 4 min and completely melt the ice into water within 10 min,showing an input power of 246 W/m2.

(3) The CNT webs and electrothermal process have little effect on the transverse mechanical properties of the CFRP.Summarily,this study reveals that the low input voltage/power and short response time needed to deice make the CNT webs decorated CFRP may be a potential new generation of deicing structure in aerospace.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This study was supported by the National Natural Science Foundation of China (No.11772233).