Microcellular SiC foams containing in situ grown nanowires for electromagnetic interference shielding

Microcellular SiC foams containing in situ grown nanowires for electromagnetic interference shielding

Journal of Industrial and Engineering Chemistry 80 (2019) 401–410 Contents lists available at ScienceDirect Journal of Industrial and Engineering Ch...

3MB Sizes 0 Downloads 3 Views

Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec

Microcellular SiC foams containing in situ grown nanowires for electromagnetic interference shielding Praveen Wilson, Sujith Vijayan, K. Prabhakaran* Department of Chemistry, Indian Institute of Space Science and Technology, Thiruvananthapuram 695 547, India

A R T I C L E I N F O

A B S T R A C T

Article history: Received 8 June 2019 Received in revised form 27 July 2019 Accepted 6 August 2019 Available online 14 August 2019

Microcellular SiC foams (MSiCFs) are produced by thermal setting of dispersions of silicon and NaCl powders in molten sucrose-glycerol solutions in a mould followed by carbonization, NaCl removal and reaction bonding at 1500  C. The acidic silica layer on silicon particle surface catalyses the setting of the pastes by OH condensation. The SiC nanowires grown in situ by a catalyst-free vapour–solid (VS) mechanism creates web-like architecture within the microcells (cell size 2 to 22 mm) of the foams. The MSiCFs with porosity in the range of 86.8–91.1 vol.% exhibit thermal conductivity and compressive strength in the ranges of 0.334–0.758 W m1 K1 and 0.97–2.38 MPa, respectively. The MSiCFs show excellent electromagnetic interference (EMI) shielding property in the X-band frequency region enhanced by the in situ grown SiC nanowires within microcells. The EMI shielding effectiveness (45.6 dB) and specific shielding effectiveness (137 dB g-1 cm3) are the highest reported for SiC foams. © 2019 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

Keywords: Microcellular SiC foams Nanowires Templating Reaction bonding EMI shielding

Introduction With expanding telecommunications network and assorted electronic devices in recent times, electromagnetic (EM) waves especially in the microwave region has become a major cause of environmental pollution [1–3]. Moreover, the EM waves also disrupt the functioning of the electronic devices unless a feasible protective shielding is provided. Metal-based protective shields are conventionally used for shielding the devices from spurious electromagnetic radiations, however, the weight of these materials poses a challenge in aerospace and aviation applications [4,5]. Though lightweight carbon and polymer-based materials exhibit good shielding characteristics, they are prone to oxidation/decomposition at high temperatures and therefore cannot be employed as EMI shielding materials in aerospace and aircraft applications where they are subjected to high temperatures [2,6–10]. For such applications, good shielding performance of the material should be complemented with low gravimetric mass as well as high thermal stability for meeting the large temperature fluctuations when operating at high altitudes and in space. Silicon carbide (SiC) is extremely stable at high temperatures and possesses low thermal expansion coefficient and excellent dielectric properties which makes it ideal for high temperature EMI shielding

* Corresponding author. E-mail address: [email protected] (K. Prabhakaran).

applications [11]. The shielding effectiveness of SiC can also be enhanced by engineering it in to porous structures (foams and aerogels) and by creating heterogeneous interfaces which attenuate the EM waves [12–18]. In addition to this, SiC foams are used as porous burners, catalytic support, thermal protection material, molten metal filter, and foam packing materials in distillation applications [19–25]. Several methods for the processing of SiC foams are reported in literature. They are broadly classified into the following categories: polymer foam replication, fugitive templating, emulsion templating, freeze-casting and direct foaming [26–38]. Carbothermal reduction and reaction bonding techniques are also employed to convert carbon foams into SiC foams [39–41]. Carbothermal reduction involves the reduction of SiO2, SiO and SiOC with carbon while reaction bonding technique involves the reaction between carbon and silicon at elevated temperatures. Alumina and yttria are used as additives to promote carbon-silicon reaction and densification of SiC [31,42]. EMI shielding performance of porous SiC/SiOC ceramics can be further improved by incorporating SiC nanowires/nanofibres within the pores. Increasing the population of SiC nanowires within the porous SiC substrate creates large number of interfaces and defects. SiC nanowire can induce electronic dipole polarization and the interfaces between the nanowire and the underlying substrate can enhance interfacial scattering. These two mechanisms facilitate the conversion of the electromagnetic energy into heat thus improving the absorption performance [43,44].

https://doi.org/10.1016/j.jiec.2019.08.020 1226-086X/© 2019 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.

402

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Incorporation of nanowires/nanofibres on the cell wall surfaces are achieved by two means: (i) by in situ growth during the ceramization of pre-ceramic polymer and (ii) by chemical vapour deposition/chemical vapour infiltration growth on the cell walls of ceramic foams [31,43–50]. Majority of the above mentioned preparation methods produces SiC foams of large cell size. However, smaller cell size is beneficial for better mechanical strength, higher EMI shielding effectiveness and lower thermal conductivity [45–49]. Microcellular ceramics are porous ceramics with cell size less than 30 m m and cell densities greater than 109 cells cm3 [50,51]. It is well known that the reflection of EM waves results from the large impedance mismatch between the shielding material and the free space. Incorporating porosities in the shielding material decreases the resulting effective permittivity of the material and thus reduces reflection loss. Foams with microcellular pores will have better impedance match with the free space by virtue of its large interfacial area as well the increased instances of multiple reflection within the cells resulting in improved absorption [2,10]. There are two major processing methods to fabricate microcellular ceramic foams viz gas bubble formation method and fugitive template methods [42,50,52–55]. We recently developed a NaCl templating method for the preparation of microcellular carbon foams with broad range of porosities (83.83–95 vol%) and cell size in the range of 2–13 mm [56,57]. The process involves thermal setting of pastes of NaCl powder in molten sucrose in a mould by  OH condensation followed by carbonization and NaCl removal. The idea of the present work is to incorporate silicon powder and sintering additives in the paste to produce microcellular SiC foams by reaction bonding after the NaCl removal. The heterogeneous SiC nanowire/microcellular SiC foams show high absorption dominated EMI shielding properties and possess reasonably high mechanical strength and low thermal conductivity. The use of biomass and NaCl makes it an eco-friendly and sustainable process with a scope for industrial-scale production.

500 ml zirconia bowl for 2 h at 200 RPM. 58.46 g of silicon powder was used for 100 g of sucrose in all the compositions anticipating a carbon yield of 24 wt% from sucrose. The NaCl to sucrose weight ratio was varied in the range of 1–3. The amount of alumina and yttria sintering additives used was 6 and 4 wt%, respectively, of the expected SiC formed from 58.46 g of silicon powder. The slurries obtained by ball milling were dried in an air oven at 70  C. The dried powder mixtures were heated in a glass tray at 185  C in an air over to melt the sucrose and thoroughly mixed using a wooden ladle. Adequate amount of glycerol is added during mixing to form pastes of mouldable consistency. The composition of the pastes used is given in Table 1. The paste was transferred to a stainless steel mould of dimension 10  10  2.5 cm3 (Fig. S1, Supplementary information) and then gently hot-pressed at 160  C in a hydraulic press for setting. After setting, the solid composite body (Fig. S2, Supplementary information) was extracted from the mould and annealed at 200  C for 2 h. A slow heating rate was used for the annealing process. The composite body was then carbonized at 750  C for 2 h in a box furnace under argon flow. A temperature ramp of 1  C min1 was used during heating and cooled naturally to room temperature. The argon flow was maintained during cooling as well. The carbonized composite body was then washed in distilled water at 90  C until the entire NaCl was leached out. This was confirmed by titrating the spent water against the aqueous AgNO3 solution. The presence of NaCl in spent water is indicated by the formation of white precipitate (AgCl) on titration with AgNO3 solution. Complete NaCl extraction is indicated by the absence of the white precipitate on titration. The washed body was dried and further sintered at 1500  C for 2 h at heating rate of 2  C min1 in an inert atmosphere furnace to obtain the final microcellular SiC foam. The microcellular SiC foams are designated as MSiCF-x, where x represents the NaCl to sucrose weight ratio used in the composition.

Experimental

The setting characteristics of the pastes was studied by measuring the torque as a function of time in a torque rheometer (Brabender Plasti-Corder, GmbH, Germany) with a rotor speed of 10 RPM at 160  C. 50 ml of the each composition was used for the study. Carbonization shrinkage and sintering shrinkages were calculated by measuring the initial and final dimensions of the respective bodies. Bulk density of the MSiCF was determined from the dimensions and weight of rectangular samples cut with a hacksaw blade and polished using emery paper. The skeletal density of the CCFs was measured using a helium pycnometer (AccuPyc II 1340 Pycnometer, Micromeritics, USA). The porosity of the foam samples were calculated from the following equation:   r ð1Þ Porosityð%Þ ¼ 1    b  100

Materials Analytical grade sucrose was procured from Sisco Research Laboratories, Maharashtra, India. NaCl, glycerol and methanol were procured from Merck India Ltd. Silicon powder with average particle size of 3.3 mm and surface area of 2.38 m2 g1 was procured from Metal Powder Company Ltd. (Madurai, India). Alumina powder of average particle size 0.34 mm and surface area 10.4 m2 g1 was procured from ACC Alcoa (Kolkata, India) and yttria powder of average particle size of 0.56 mm and specific surface area 2.9 m2 g1 was procured from Sigma Aldrich (USA). Distilled water was used for NaCl extraction. Synthesis Mixtures of sucrose, NaCl, silicon, alumina and yttria powders were ball-milled in a planetary ball mill (Fritsch, Germany) in methanol medium using zirconia balls of 10 mm diameter in a

Characterization

rs

where rb and rs are the bulk and skeletal density of the SiC foams, respectively. Thermal stability of the carbon–silicon composite foams was studied by thermogravimetry technique using a thermogravimetric analyzer (TGA, Q-50, TA Instruments, USA). Analysis of samples was performed both under air and N2 flow at a

Table 1 Compositions used for the preparation of MSiCF. Sample

Sucrose (g)

NaCl (g)

Silicon (g)

Alumina (g)

Yttria (g)

Glycerol (g)

MSiCF-1 MSiCF-1.5 MSiCF-2 MSiCF-2.5 MSiCF-3

100 100 100 100 100

100 150 200 250 300

58.46 58.46 58.46 58.46 58.46

5 5 5 5 5

3.34 3.34 3.34 3.34 3.34

2 18 32 45 64

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

heating rate of 10  C min1. The X-ray diffraction (XRD) analysis of the MSiCF samples was carried out in an X-ray diffractometer (X’pertPro, Philips, USA) using Cu Kα radiation. The microstructural analysis of the various samples was performed using scanning electron microscope (SEM; FEI Quanta FEG200, USA). EDAX analysis was performed using (QUANTAX 200 with XFlash16/30 SDD Detector, Bruker, USA). The cell sizes and diameter of SiC nanowires of the MSiCFs are measured from the SEM images using ImageJ software. The mechanical properties of MSiCF samples were analysed using a Universal Testing Machine (Instron 5050, Instron, USA) at a crosshead speed of 0.5 mm min1 following ASTM standard C365/C365M-05. Compressive strength and Young’s modulus were calculated from the stress–strain plots. Compressive strength was defined as the highest height of the plateau region of stress–strain plot and the Young’s modulus was measured from the initial linear slope of the stress–strain plot. N2 adsorption–desorption isotherms at 77 K were plotted using surface area analyser (Micromeritics Tristar II, USA). The samples were degassed under flowing N2 condition at 300  C for 12 h before the analyses. The surface area was calculated using BET (Brunauer– Emmett–Teller) equation in the relative pressure range of 0.05– 0.25. The thermal conductivity measurement of the MSiCF samples was carried out at room temperature using the transient plane source method (Hot Disk, TPS 2500S, Sweden). For measurement, a pair of samples measuring 40  40  15 mm3 from each set were cut and polished to obtain near-perfect flat surfaces so as to ensure no air gap exists when the sensor is sandwiched in between the samples (Fig. S3, Supplementary information). The electrical conductivity of MSiCFs and carbon–silicon composite foam was recorded using a four probe method. A constant current (I) was applied to the samples using a programmable source (Keithley 6221A). The voltage (V) between two pins separated by a distance (s) of 2 mm was measured using a nanovoltmeter. The bulk resistivity (r) was calculated from the current and voltage readings using the relation:   V r ¼ 2ps ð2Þ 1 The electromagnetic interference (EMI) shielding effectiveness (SE) and the complex relative permittivity of the MSiCFs and the composite foam was measured in the X-band (8.2–12.4 GHz) of the microwave region using a vector network analyser (VNA) (E5071C, Agilent Technologies, USA) and a corresponding waveguide. Rectangular samples of dimensions 22.7  10.2  5 mm3 were push-fitted in the cavity of the shim of the wave-guide ensuring perfect contact with the internal walls. The scattering parameters S11, S12, S21 and S22 were recorded and EMI shielding effectiveness (EMI SE) was deduced from the scattering parameters.

403

consistency increases from 2 to 64 g per 100 g sucrose when the NaCl to sucrose weight ratio increases from 1 to 3. The mouldable consistency was ascertained by qualitatively inspecting its deformability at 160  C with the help of a wooden ladle. The setting characteristics of the pastes measured in terms of torque as a function of time at 160  C is shown in Fig. 1. The torque increases at a slow rate in the initial stage followed by a sudden rapid rise. The time at which an abrupt rise in torque is observed, increases with an increase in NaCl to sucrose weight ratio. The increase in torque is due to the evaporation of glycerol as well as by the condensation polymerisation of glucose and fructose anhydride produced from sucrose [56,57]. The delay in torque build up at higher NaCl to sucrose weight ratio is due to the longer time required for the evaporation of large concentration of glycerol. The pastes corresponding to NaCl to sucrose weight ratios of 1 and 1.5 reach a torque value of 50 N m at nearly 50 min. On the other hand, the pastes corresponding to NaCl to sucrose weight ratios of 2, 2.5 and 3 reach a torque value of 50 N m in 81, 136 and 208 min, respectively. It is interesting to note that paste produced from sucrose–NaCl mixture (without silicon powder and glycerol) at NaCl to sucrose weight ratio of 1 takes much longer time (150 min) to reach a torque value of 50 N m. This indicates that the acidic silica layer on the surface of silicon particles catalyses the OH condensation between glucose and fructose anhydride leading to faster setting. The actual setting time was determined by keeping the pastes prepared at various NaCl to sucrose weight ratios at 160  C in an air oven and physically inspecting the same after every 30 min. The time at which the soft paste turned into a hard solid is considered as the setting time. The setting time of the pastes prepared at various NaCl to sucrose weight ratios is given in Table 2. The paste filled in the rectangular mould could be removed as a solid sucrose polymer–silicon–NaCl composite body after the corresponding setting time. The sucrose polymer in the composite body undergoes carbonization during inert atmosphere heat treatment at 750  C to form carbon–silicon–NaCl composite. The composite undergoes shrinkage during the carbonization. There is a decrease in both linear and volumetric shrinkages of the bodies during carbonization from 4 to 1.36% and 10.8 to 5%, respectively, as we increase the NaCl to sucrose weight ratio from 1 to 3. The decrease in shrinkage is due to the fact that the NaCl and silicon particles do not shrink whereas only the sucrose polymer shrinks during the carbonization. The carbon–silicon–NaCl composite has uniform distribution of NaCl particles of size in the range of 3–23 m m as evidenced by the SEM image of fractured surface shown in

Results and discussion Processing of microcellular SiC foams The sucrose on melting, setting by polymerization, and subsequent carbonization at 750  C produces 24 wt% carbon. The amount of silicon powder used in the compositions is stoichiometrically equivalent to react with this carbon to produce SiC. Therefore, the sucrose to silicon powder weight ratio is kept the same in all the formulations. The NaCl to sucrose weight ratio is varied from 1 to 3 to modulate the porosity. The resin produced by the melting of sucrose in sucrose-NaCl–silicon powder mixture is not sufficient to form a paste of mouldable consistency. Therefore, glycerol is used as a plasticising additive as it does not retain any carbon during pyrolysis and homogeneously mixes with molten sucrose. The requirement of glycerol to achieve mouldable

Fig. 1. Torque-time curves of NaCl–silicon powder pastes in molten sucrose glycerol solution at various NaCl to sucrose ratios of a-1, b-1.5, c-2, d-2.5 & e-3 and paste of NaCl in molten sucrose at NaCl to sucrose weight ratio of 1 (f).

404

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Table 2 Setting time of compositions. Sample

MSiCF-1

MSiCF-1.5

MSiCF-2

MSiCF-2.5

MSiCF-3

Approximate setting time (h)

1

1.5

2

2.5

3.5

Fig. 2. High magnification SEM images of (a) carbon–NaCl–Si composite (b) carbon–Si composite foam after NaCl extraction, inset shows higher magnification SEM image of microcells.

Fig. 2a. These NaCl particles are completely leached out by immersing in water at 90  C for 48 h. The voids created by the removal of NaCl particles remains as microcells in the resulting carbon–silicon composite foam. The SEM image of fractured surface of carbon–silicon composite foam is shown in Fig. 2b. The uniform distribution of silicon particles in the carbon matrix is evidenced in the high-magnification SEM image. As envisaged from the 24 wt% carbon yield of sucrose, the carbon content in the carbon–silicon composite foam is very close to that required for reaction with the silicon present. The TGA graph of the carbon– silicon composite foams obtained from paste of various compositions is shown in Fig. 3. The weight loss of 23–24% observed up to 600  C is due to the burnout of carbon and a slight increase in weight after 600  C is due to the oxidation of silicon in the sample. The reaction bonding of silicon and carbon in the carbon-silicon composite foam during inert atmosphere heat treatment results in the formation of β-SiC foam. This is clearly evidenced by the XRD analysis as well as the change of colour from black to pale green.

The XRD pattern of carbon–silicon composite foam before and after high temperature heat-treatment is shown in Fig. 4. The peaks at 2u values of 28.7, 47.6, 56.3, 69.4, 76.6 and 88.2 corresponding to the reflections from the planes (111), (220), (311), (400), (331), and (422), respectively, of elemental silicon vanish after the high temperature heat treatment. On the other hand, peaks at 2u values of 35.6, 41.3, 59.9 and 71.8 corresponding to reflection from (111), (200), (220) and (311) planes of β-SiC appear [31,50]. The absence of any peak corresponding to carbon in the carbon–silicon composite foam indicates its amorphous nature. Further, the absence of peaks corresponding to silicon after the heat treatment at 1500  C indicates the complete conversion of silicon to silicon carbide. The carbon–silicon composite foam sample undergoes marginal shrinkage during the reaction bonding. The linear and volumetric shrinkages during reaction bonding are in the ranges of 0.2–2.3% and 1.17–7.7%, respectively. Large amount of porosity is expected in struts and cell walls of SiC foam obtained from the carbon–silicon composite foam. The small

Fig. 3. TGA graph in air of carbon-silicon composites foams prepared at various NaCl to sucrose weight ratios.

Fig. 4. XRD spectrum of carbon-silicon composite and MSiCF. SF indicates the peak corresponding to the stacking faults.

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

shrinkage observed is due to the densification of the struts and cell walls. Fig. 5 shows a digital photograph of carbon-silicon composite foam and SiC foam obtained by reaction bonding. Microstructure The low and high-magnification SEM images of MSiCF are shown in Fig. 6. From the low-magnification images it is observed that MSiCF-1 and MSiCF-1.5 contain large macropores which develop when water vapour generated during polymerization of glucose and fructose anhydride through OH condensation gets entrapped within the matrix. MSiCF-2, MSiCF-2.5, and MSiCF-3 do not show these features since the polymerization is sluggish due to higher glycerol quantities which allows water vapour to diffuse out gradually from the mould. Also, higher concentrations of solid NaCl particles in these compositions restricts the space for the water vapour to accumulate. The MSiCF exhibit uniform distribution of microcells with open cellular microstructure as observed in the high-magnification SEM images. The cell size determined from the images is in the range of 2–22 mm which is consistent in all the studied compositions. A typical cell size distribution of MSiCF is shown in Fig. S4, Supplementary information. The samples prepared with higher concentrations of NaCl show thinning of the cell walls as the concentration of the carbon and silicon producing SiC correspondingly decreases. At high magnification, the fractured struts of the foams show nano-sized grains (35–110 nm) of near-spherical geometry (Fig. S5, Supplementary information). Interestingly, the high-magnification images also show extensive web-like fibrous nanostructures of nearly 30–150 nm diameter occupying the space inside the microcells as well the surfaces of the cell wall. The EDAX analysis (Fig. 7) of the nanowires confirms the composition as 60.1 wt% of Si and 31.6 wt% of carbon. In addition, EDAX spectrum also shows the presence of aluminium and oxygen derived from the sintering additives. The carbon to silicon atomic ratio in the nanowires calculated from the EDAX data is 1.23. This indicates that the nanowires grown on the cell wall surfaces are carbon-rich SiC. A weak diffraction peak at 2u = 33.3 in the XRD spectrum (Fig. 4) of SiC is attributed to the stacking fault in (111) plane in the SiC nanowires [58,59]. The decoration of SiC and SiOC surfaces with one dimensional SiC nanostructures by chemical vapour infiltration or through carbothermal reduction in the presence of catalysts such as iron and nickel is well established in literature [44,60]. In addition, there are numerous reports on the growth of SiC nanowires on SiC

405

and SiOC surfaces without the aid of catalysts through vapour– solid (VS) and vapour–liquid–solid (VLS) mechanisms [43,61,62]. The VS mechanism is briefly explained thus: trace concentrations of SiO2 on the surface of Si particles react with carbon producing SiO and CO. This is succeeded by the reaction of one mole of SiO with 3 mol of CO producing SiC and 2 mol of CO2. The CO2 reacts with carbon to produce CO and drives the positive feedback to reaction (4) thus sustaining the growth of SiC crystals. The reactions involved in VS mechanism of SiC nanowire growth are shown below. In the present case, it appears that the growth of SiC nanowire is through the VS mechanism as catalyst such as Fe and Ni was not included in the preparation. SiO2(s) + C(s) → SiO(g) + CO(g)

(3)

SiO(g) + 3 CO(g) → SiC(s) + 2CO2(g)

(4)

CO2(g) + C(s) → 2CO(g)

(5)

Foam density and surface area The variation in the bulk density of the MSiCF is shown in Fig. 8. The bulk density of the foam is within the range of 0.284–0.419 g cm3. The density of MSiCF-1.5 (0.419 g cm3) is slightly higher than that of MSiCF-1 (0.411 g cm3) even though the former was prepared with higher concentration of NaCl in its composition. This is due to the large population of macropores in MSiCF-1 produced by the entrapment of water vapour as evidenced from the SEM image (Fig. 6). The density of the foams decrease from 0.419–0.284 g cm3 when the NaCl to sucrose weight ratio increases from 1.5 to 3. The average skeletal density of the foams determined from Helium pycnometry is 3.19 g cm3. The porosity calculated from the bulk and skeletal densities using Eq. (1) ranges from 86.9 to 91.1%. The effect of porosity and the in situ grown nanowires on the surface area of the foams is studied using N2 adsorption– desorption isotherm. The surface area of MSiCFs calculated using BET equation from the N2 adsorption–desorption isotherms is in the range of 7.6–9.98 m2 g1 (Table 3). MSiCF-1 has a surface area of 8.03 m2 g1 while MSiCF-1.5 exhibits marginally lower value of 7.6 m2 g1. With increasing porosity we observe a significant increase in the surface area. The increase in the surface area is essentially due to the increase in the area of cell walls as well as higher concentrations of the SiC nanowires at higher porosities.

Fig. 5. Digital photograph of (a) carbon–silicon composite foam (b) SiC foam.

406

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Fig. 6. Low-magnification (a, c, e) and high-magnification (b, d, f) SEM images of MSiCF-1, MSiCF-2 and MSiCF-3, respectively.

Compressive strength and thermal conductivity The SiC foams for EMI shielding and thermal insulation applications requires adequate mechanical strength. The stress– strain plot MSiCF is shown in Fig. 9a. The graphs are analogous to the stress–strain curves exhibited by brittle cellular foams with an elastic behaviour at low strains represented by the linear section followed by a plateau region signifying the collapse of the cells with increasing strain. At high strains, we observe rapid rise in the compressive stress which results from the densification of the foams after complete cellular collapse [47]. The compressive strength and Young’s modulus of the foams are shown in Fig. 9b. The compressive strength and Young’s modulus display dependence on the density of the foams. The compressive strength value

ranges from 0.97 to 2.38 MPa while the Young’s modulus is between 52 to 176 MPa. The peak values are exhibited by MSiCF-1.5 while the lowest values correspond to MSiCF-3 which is expected because these samples show the highest and lowest density, respectively. The thermal conductivity of the MSiCFs lie within the range of 0.334–0.758 W m1 K1 (Fig. 10a). It follows a similar trend to that of the bulk density with MSiCF-1.5 showing a high value of 0.758 W m1 K1 and MSiCF-3 showing a low value of 0.334 W m1 K1. It is well known that the thermal conductivity of porous materials is largely dependent on its bulk density [57]. High bulk density is indicative of thick cell walls and struts which are amenable to greater heat transport. The high porosity of foams is associated with thin cell walls and narrow struts which results in high tortuosity thus impeding heat transport. Fig. 10b depicts the

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

407

Fig. 7. EDAX spectrum of nanowires grown in MSiCF.

correlation between the thermal conductivity and density of the MSiCFs. Thermal conductivity shows a linear dependency on the density of the foams. Electrical properties

Fig. 8. Density of MSiCFs.

Table 3 BET surface area of MSiCFs. Sample

MSiCF-1

MSiCF-1.5

MSiCF-2

MSiCF-2.5

MSiCF-3

BET surface area (m2 g1)

8.03

7.6

9.6

9.71

9.98

The direct current electrical conductivities of the MSiCFs measured using a four probe instrument is shown in Fig. 11. The electrical conductivity of the foams is in the range of 1.02  104– 11.1 104 S cm1. Contrary to the expected, the electrical conductivity increases with a decrease in the foam density. This effect is explained by the increase in the SiC nanowire population within the microcells with increase in the porosity of the foams. The one dimensional nanowires with high carrier concentration form interconnecting networks between cell walls thus contributing to higher electrical conductivity [18]. The dielectric properties of the MSiCFs were measured in the frequency range of 8.2–12.4 GHz (X-band). Dielectric materials are characterized by their relative complex permittivity (e = e0  je00 ) where, e0 is the real component and e00 is the imaginary component of the complex permittivity representing storage and loss capability, respectively. According to the Debye theory, the real part (e0 ) of the relative complex permittivity is related to the polarization relaxation and the imaginary part (e00 ) is related to the dielectric loss of the material [63–65]. The dielectric loss tangent (tan d = e00 /e0 ) characterizes the microwave attenuation property of the material. Higher tan d indicates superior microwave attenuation capacity. Fig. 12 shows the complex permittivity and the

Fig. 9. Stress–strain plots (a) and compressive strength and Young’s modulus (b) of MSiCFs.

408

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Fig. 10. Thermal conductivity (a) and the correlation between thermal conductivity and density (b) of MSiCFs.

apparent correlation with porosity either. The tan d of MSiCF-2 and MSiCF-2.5 increase with the increase of frequency of EM waves. The highest average tan d value of 0.55 is exhibited by MSiCF-2.5. EMI shielding

Fig. 11. Electrical conductivity of MSiCFs.

dielectric loss of the MSiCFs as a function of frequency. It can be seen that the real component (e') of the complex permittivity of all MSiCFs decreases with increase in frequency (Fig. 12a). At high EM wave frequencies, the rearrangement of dipoles is not fast enough to match the alternating electric field causing a decline in the dielectric polarization [15]. The average value of e' of MSiCFs lies within the range of 2.5–4.0 with MSiCF-2.5 showing the highest value. Meanwhile, the e00 values of all but MSiCF-2.5 also exhibit declining trend with increase in the EM wave frequency (Fig. 12b). The e00 of MSiCF-2.5 shows a slight increase with increase in the frequency. The average value of e00 ranges from 0.69 to 2.18 and does not follow any obvious trend in relation to porosity with MSiCF-2.5 exhibiting the highest value of 2.18. The dielectric loss tangent (tan d) of MSiCFs varies between 0.25 and 0.55 showing no

An ideal EMI shielding material should exhibit high value of shielding effectiveness over entire range frequency in the selected band. EMI shielding takes effect through factors such as absorption through attenuation of incident wave (SEA), reflection from the incident plane (SER) and multiple internal reflection (SEM) within the shielding material [15,18,66]. Reflection of the EM waves occur when there is an impedance mismatch between the free space and the shielding material. Absorption is explained by the dielectric loss involving conduction and relaxation losses. The relaxation loss is effected by the polarization caused by dipoles, defects, grain boundaries and functional groups in the material while conduction loss occurs due to its electrical conductivity. Multiple internal reflections are caused by porous materials or composites with two or more phases with different dielectric properties. So, the total shielding effectiveness (SET) can be summarized using the equation: SET = SEA + SER + SEM

(6)

In cases where SET > 15, the multiple internal reflection component is usually neglected. The Eq. (3) can be reorganized as: SET  SEA + SER

(7)

Fig. 13a is a typical graph showing total shielding effectiveness of MSiCF and contribution from absorption and reflection components towards the same as a function of frequency in the X-band region of microwave. The total shielding effectiveness in MSiCFs is more or less independent of frequency. Majority of the

Fig. 12. The real permittivity (a), imaginary permittivity (b) and the dielectric loss tangent (c) of MSiCFs as a function of frequency (8.2–12.4 GHz).

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

409

Fig. 13. The SER, SEA and SET of MSiCF-2.5 over a frequency range of 8.2–12.4 GHz (a), the shielding effectiveness (b) and specific shielding effectiveness (c) of MSiCFs at 11 GHz.

reported SiC and SiOC based materials shows large variation in the shielding effectiveness with frequency as the dipoles fail to realign with the EM wave to induce dielectric polarization loss [15]. The SEA and SER values observed are in the ranges 24.9–38.8 dB and 3.4–6.8 dB, respectively. The SET varies in the range of 28.3–45.6 dB with a maximum for MSiCF-2.5 and minimum for MSiCF-3. However, the SET values of MSiCF is found to be lower than that of the carbon-silicon composite foams. This is due to the high electrical conductivity of the carbon–silicon composite foams. For example, the carbon–silicon composite foam prepared at NaCl to sucrose ratio of 2.5 shows a SET of 53 dB due to a high electrical conductivity of 790 S cm1. EMI shielding effectiveness of a typical carbon-silicon composite foam as a function of frequency in the Xband region is show in Fig. S6, Supplementary information. However, the carbon–silicon composite foams are prone to oxidation above 300  C and are not suitable for EMI shielding applications at high temperature. We observe that the absorption component of the total shielding effectiveness of MSiCFs is significantly higher than the shielding through reflection with SER values showing only a slight variation with respect to the samples. The low reflection component is due to a narrow mismatch in the impedance of the material and free space attributed to microcellular nature of the highly porous (86.8–91.1 vol.%) material. Generally, impedance matching is accomplished by incorporating pores in the material where the porous material is considered as mixture of air and the solid matrix. The effective permittivity (eeff) of the porous absorber material is thus determined by the Maxwell-Garnett (MG) theory:

eef f   ¼  e1

ðe2 þ 2e1 Þ þ 2f  ðe2  e1 Þ ðe2 þ 2e1 Þ  f ðe2  e1 Þ

ð8Þ

where e1 and e2 are the permittivity of the solid matrix and air, respectively and f is the volume fraction of the air in the effective medium [15,65,67]. This is a valuable characteristic of a radar absorbing material used in stealth-based defence application wherein the EM waves have to be absorbed rather than reflected back [68–71]. The total shielding effectiveness and its absorption and reflection components of SiC foams prepared at various NaCl to sucrose weight ratios is shown in Fig. 13b. The SET values obtained for MSiCFs are higher than that reported recently for SiC foams prepared from dough (24.2 dB) and SiOC foams prepared from polysiloxane (18 dB) [14,18]. The shielding effectiveness of a material depends on the density and available surface for EMI attenuation. The higher SET values obtained for MSiCFs is due to their microcellular structure, nanograin size and the presence of vast networks of SiC nanowires within the microcells. Microcells in the foams enhance the microwave absorption through multiple internal reflection mechanism and subsequent dissipation of EM energy as heat [14,15]. In addition, small amount of free carbon that may be present contribute to the EMI attenuation by ohmic losses. The higher SET

of MSiCF-1.5 (41 dB) compared to that of MSiCF-1 (31 dB) is due to its higher density [72]. On the other hand, increase in shielding effectiveness from MSiCF-1.5 to MSiCF-2.5 is due to the increase in the available surface for the attenuation of EM radiation due to increase in the number of microcells as well as the population of SiC nanowires with the stacking faults. These SiC nanowires form extensive interfaces with the SiC matrix which induces EM absorption through dipole polarization [18,43,44,73,74]. The sample MSiCF-3 shows the lowest SET value of 28.3 dB which can be ascribed to the fact that it has the lowest density of all the studied samples. It appears that the concentration of SiC nanowires in the microcells is insufficient to overcome the effect due to decrease in density. The specific shielding effectiveness (Fig. 13c) of MSiCF calculated on the basis of density of the foams is in the range of 98–137 dB g1 cm3. The specific shielding effectiveness obtained is much higher than that reported for SiC foam prepared from dough (24.8 dB g1 cm3) and SiOC foam prepared from polysiloxane (13.85 dB g1 cm3) [14,18]. With a focus on developing inexpensive, lightweight, thermally and chemically stable material with high EMI shielding properties, the microcellular SiC foams promises great advantage for aerospace and aviation applications. Conclusion Microcellular SiC foams (MSiCFs) were produced by moulding and thermal setting of pastes of silicon and NaCl powder mixtures in molten sucrose–glycerol solutions followed carbonization, NaCl removal by leaching with hot water and reaction bonding at 1500  C. The microcell size was in the range of 2–22 mm. The carbon rich SiC nanowires, grown in situ during the reaction bonding via a catalyst-free vapour–solid (VS) mechanism, form web-like architecture within the microcells. The density, thermal conductivity and compressive strength of the foams were in the ranges of 0.284–0.419 g cm3, 0.334–0.758 W m1 K1 and 0.97–2.38 MPa, respectively. The excellent electromagnetic interference shielding property in the X-band frequency of MSiCFs is due to superior impedance matching as a result of high porosity, enhanced polarization at foam-nanowire interfaces and defect sites in the SiC nanowire, internal reflection within the microcell and ohmic losses associated with the presence of small amount of free carbon. The MSiCFs showed shielding effectiveness (SE) and specific shielding effectiveness (SSE) as high as 45.6 dB and 137 dB g-1 cm3, respectively. The lightweight and high-absorption dominated shielding effectiveness make the MSiCFs suitable candidates for EMI shielding materials in aircraft and aerospace structures. Acknowledgements The authors are thankful to the Indian Institute of Space Science and Technology (IIST), Thiruvananthapuram, India, for the financial support. The authors also thank the Materials Science and

410

P. Wilson et al. / Journal of Industrial and Engineering Chemistry 80 (2019) 401–410

Technology Division, National Institute of Interdisciplinary Science and Technology (NIIST), Thiruvananthapuram, India, for EMI shielding measurements. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jiec.2019.08.020. References [1] A.K. Singh, A. Shishkin, T. Koppel, N. Gupta, Compos. B Eng. 149 (2018) 188, doi: http://dx.doi.org/10.1016/j.compositesb.2018.05.027. [2] J. Ling, W. Zhai, W. Feng, B. Shen, J. Zhang, W.G. Zheng, ACS Appl. Mater. Interfaces 5 (2013) 2677, doi:http://dx.doi.org/10.1021/am303289m. [3] Y.J. Wan, P.L. Zhu, S.H. Yu, R. Sun, C.P. Wong, W.H. Liao, Small 14 (2018), doi: http://dx.doi.org/10.1002/smll.201800534. [4] Y. Yuan, X. Sun, M. Yang, F. Xu, Z. Lin, X. Zhao, Y. Ding, J. Li, W. Yin, Q. Peng, X. He, Y. Li, ACS Appl. Mater. Interfaces 9 (2017) 21371, doi:http://dx.doi.org/10.1021/ acsami.7b04523. [5] Z. Zeng, Y. Zhang, X.Y.D. Ma, S.I.S. Shahabadi, B. Che, P. Wang, X. Lu, Carbon N. Y. 140 (2018) 227, doi:http://dx.doi.org/10.1016/j.carbon.2018.08.061. [6] L. Zhang, M. Liu, S. Roy, E.K. Chu, K.Y. See, X. Hu, ACS Appl. Mater. Interfaces 8 (2016) 7422, doi:http://dx.doi.org/10.1021/acsami.5b12072. [7] Q. Liu, D. Zhang, T. Fan, Appl. Phys. Lett. 93 (2008), doi:http://dx.doi.org/ 10.1063/1.2957035. [8] Y. Yang, M.C. Gupta, K.L. Dudley, R.W. Lawrence, Nano Lett. 5 (2005) 2131, doi: http://dx.doi.org/10.1021/nl051375r. [9] P.R. Agarwal, R. Kumar, S. Kumari, S.R. Dhakate, RSC Adv. 6 (2016) 100713, doi: http://dx.doi.org/10.1039/C6RA23127F. [10] B. Shen, Y. Li, D. Yi, W. Zhai, X. Wei, W. Zheng, Carbon N. Y. 102 (2016) 154, doi: http://dx.doi.org/10.1016/j.carbon.2016.02.040. [11] C. Liu, D. Yu, D.W. Kirk, Y. Xu, RSC Adv. 7 (2017) 595, doi:http://dx.doi.org/ 10.1039/c6ra25142k. [12] S. Dong, X. Zhang, D. Zhang, B. Sun, L. Yan, X. Luo, J. Eur. Ceram. Soc. 38 (2018) 29, doi:http://dx.doi.org/10.1016/j.jeurceramsoc.2017.07.034. [13] Y. Kong, X. Shen, S. Cui, M. Fan, Ceram. Int. 40 (2014) 8265, doi:http://dx.doi. org/10.1016/j.ceramint.2014.01.025. [14] C. Liang, Z. Wang, L. Wu, X. Zhang, H. Wang, Z. Wang, ACS Appl. Mater. Interfaces 9 (2017) 29950, doi:http://dx.doi.org/10.1021/acsami.7b07735. [15] C. Liang, Z. Wang, Chem. Eng. J. 373 (2019) 598, doi:http://dx.doi.org/10.1016/j. cej.2019.05.076. [16] N. Song, H. Zhang, H. Liu, J. Fang, Ceram. Int. 43 (2017) 6786, doi:http://dx.doi. org/10.1016/j.ceramint.2017.02.095. [17] H. Liu, H. Cheng, J. Wang, G. Tang, J. Alloys Compd. 491 (2010) 248, doi:http:// dx.doi.org/10.1016/j.jallcom.2009.09.163. [18] W. Duan, X. Yin, Q. Li, X. Liu, L. Cheng, L. Zhang, J. Eur. Ceram. Soc. 34 (2014) 257, doi:http://dx.doi.org/10.1016/j.jeurceramsoc.2013.08.029. [19] Z. Taslicukur, C. Balaban, N. Kuskonmaz, J. Eur. Ceram. Soc. 27 (2007) 637, doi: http://dx.doi.org/10.1016/j.jeurceramsoc.2006.04.129. [20] S. Gianella, D. Gaia, A. Ortona, Adv. Eng. Mater. 14 (2012) 1074, doi:http://dx. doi.org/10.1002/adem.201200012. [21] C. Duong-Viet, H. Ba, Z. El-Berrichi, J.M. Nhut, M.J. Ledoux, Y. Liu, C. Pham-Huu, New J. Chem. 40 (2016) 4285, doi:http://dx.doi.org/10.1039/c5nj02847g. [22] C.M.V. Bolivar, A. Antonini, S. Biamino, M. Pavese, P. Fino, C. Badini, Adv. Eng. Mater. 12 (2010) 617, doi:http://dx.doi.org/10.1002/adem.200900328. [23] X. Gao, X. Li, X. Liu, H. Li, Z. Yang, J. Zhang, Chem. Eng. Sci. 135 (2015) 489, doi: http://dx.doi.org/10.1016/j.ces.2014.11.044. [24] X. Li, Q. Shi, H. Li, Y. Yao, A.N. Pavlenko, X. Gao, J. Eng. Thermophys. 26 (2017) 452, doi:http://dx.doi.org/10.1134/S1810232817040026. [25] H. Li, L. Fu, X. Li, X. Gao, AIChE J. 61 (2015) 4509, doi:http://dx.doi.org/10.1002/ aic.14944. [26] S.S. Hossain, S. Sarkar, N.K. Oraon, A. Ranjan, J. Mater. Sci. 51 (2016) 9865, doi: http://dx.doi.org/10.1007/s10853-016-0220-1. [27] Z. Zhang, F. Wang, X. Yu, Y. Wang, Y. Yan, K. Li, Z. Luan, J. Am. Ceram. Soc. 92 (2009) 260, doi:http://dx.doi.org/10.1111/j.1551-2916.2008.02842.x. [28] Y.-H. Koh, E.-J. Lee, B.-H. Yoon, J.-H. Song, H.-E. Kim, H.-W. Kim, J. Am. Ceram. Soc. 89 (2006) 3646, doi:http://dx.doi.org/10.1111/j.1551-2916.2006.01311.x. [29] N.D. Shcherban, S.M. Filonenko, P.S. Yaremov, S.A. Sergiienko, V.G. Ilyin, D.Y. Murzin, J. Mater. Sci. 52 (2017) 3917, doi:http://dx.doi.org/10.1007/s10853-016-0652-7. [30] J. Yan, A. Wang, D.P. Kim, J. Phys. Chem. B 110 (2006) 5429, doi:http://dx.doi. org/10.1021/jp057378+. [31] S. Vijayan, P. Wilson, R. Sreeja, K. Prabhakaran, J. Am. Ceram. Soc. 99 (2016) 3866, doi:http://dx.doi.org/10.1111/jace.14450. [32] F. Xue, K. Zhou, N. Wu, H. Luo, X. Wang, X. Zhou, Z. Yan, I. Abrahams, D. Zhang, Ceram. Int. 44 (2018) 6293, doi:http://dx.doi.org/10.1016/j.ceramint.2018.01.019. [33] R. Mouazer, S. Mullens, I. Thijs, J. Luvten, A. Buekenhoudt, Adv. Eng. Mater. 7 (2005) 1124, doi:http://dx.doi.org/10.1002/adem.200500163. [34] R. Mouazer, I. Thijs, S. Mullens, J. Luyten, Adv. Eng. Mater. 6 (2004) 340, doi: http://dx.doi.org/10.1002/adem.200400009.

[35] H. Ge, G. Wang, B. Yuan, B. Dong, H. Li, Ceram. Int. 40 (2014) 11705, doi:http:// dx.doi.org/10.1016/j.ceramint.2014.03.182. [36] S. Ungureanu, G. Sigaud, G.L. Vignoles, C. Lorrette, M. Birot, A. Derré, O. Babot, H. Deleuze, A. Soum, G. Pécastaings, R. Backov, J. Mater. Chem. 21 (2011) 14732, doi:http://dx.doi.org/10.1039/C1JM12084K. [37] M. Fukushima, P. Colombo, J. Eur. Ceram. Soc. 32 (2012) 503, doi:http://dx.doi. org/10.1016/j.jeurceramsoc.2011.09.009. [38] P. Colombo, Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 364 (2006) 109, doi: http://dx.doi.org/10.1098/rsta.2005.1683. [39] J. Li, W. Yuan, C. Deng, H. Zhu, J. Eur. Ceram. Soc. 37 (2017) 1131, doi:http://dx. doi.org/10.1016/j.jeurceramsoc.2016.10.025. [40] J.H. Eom, Y.W. Kim, I.H. Song, H.D. Kim, Mater. Sci. Eng. A 464 (2007) 129, doi: http://dx.doi.org/10.1016/j.msea.2007.03.076. [41] K.Y. Lim, Y.W. Kim, I.H. Song, Ceram. Int. 39 (2013) 6827, doi:http://dx.doi.org/ 10.1016/j.ceramint.2013.02.014. [42] J.H. Eom, Y.W. Kim, I.H. Song, H.D. Kim, J. Eur. Ceram. Soc. 28 (2008) 1029, doi: http://dx.doi.org/10.1016/j.jeurceramsoc.2007.09.009. [43] B. Du, C. He, A. Shui, X. Zhang, C. Hong, Ceram. Int. 45 (2019) 1208, doi:http:// dx.doi.org/10.1016/j.ceramint.2018.09.306. [44] W. Duan, X. Yin, C. Luo, J. Kong, F. Ye, H. Pan, J. Eur. Ceram. Soc. 37 (2017) 2021, doi:http://dx.doi.org/10.1016/j.jeurceramsoc.2016.12.038. [45] S.E. Zakiyan, H. Azizi, I. Ghasemi, Compos. Sci. Technol. 157 (2018) 217, doi: http://dx.doi.org/10.1016/j.compscitech.2018.02.002. [46] A. Ameli, P.U. Jung, C.B. Park, Carbon N. Y. 60 (2013) 379, doi:http://dx.doi.org/ 10.1016/j.carbon.2013.04.050. [47] M. Letellier, C. Delgado-Sanchez, M. Khelifa, V. Fierro, A. Celzard, Carbon N. Y. 116 (2017) 562, doi:http://dx.doi.org/10.1016/j.carbon.2017.02.020. [48] S.S. Sundarram, W. Li, Polym. Eng. Sci. 53 (2013) 1901, doi:http://dx.doi.org/ 10.1002/pen.23452. [49] F.X. Alvarez, D. Jou, A. Sellitto, Appl. Phys. Lett. 97 (2010) 2010, doi:http://dx. doi.org/10.1063/1.3462936. [50] Y.W. Kim, H.D. Kim, C.B. Park, J. Am. Ceram. Soc. 88 (2005) 3311, doi:http://dx. doi.org/10.1111/j.1551-2916.2005.00597.x. [51] Y.W. Kim, S.H. Kim, C. Wang, C.B. Park, J. Am. Ceram. Soc. 86 (2003) 2231, doi: http://dx.doi.org/10.1111/j.1151-2916.2003.tb03641.x. [52] Y.W. Kim, S.H. Kim, H.D. Kim, C.B. Park, J. Mater. Sci. 39 (2004) 5647, doi:http:// dx.doi.org/10.1023/B:JMSC.0000040071.55240.85. [53] Y.W. Kim, S.H. Kim, I.H. Song, H.D. Kim, C.B. Park, J. Am. Ceram. Soc. 88 (2005) 2949, doi:http://dx.doi.org/10.1111/j.1551-2916.2005.00509.x. [54] P. Colombo, E. Bernardo, Compos. Sci. Technol. 63 (2003) 2353, doi:http://dx. doi.org/10.1016/S0266-3538(03)00268-9. [55] M. Shibuya, T. Takahashi, K. Koyama, Compos. Sci. Technol. 67 (2007) 119, doi: http://dx.doi.org/10.1016/j.compscitech.2006.03.022. [56] P. Wilson, S. Vijayan, K. Prabhakaran, Carbon N. Y. 118 (2017) 545, doi:http://dx. doi.org/10.1016/j.carbon.2017.03.084. [57] P. Wilson, S. Vijayan, K. Prabhakaran, Mater. Des. 139 (2018) 25, doi:http://dx. doi.org/10.1016/j.matdes.2017.10.063. [58] S. Farhan, R. Wang, K. Li, Ceram. Int. 42 (2016) 11330, doi:http://dx.doi.org/ 10.1016/j.ceramint.2016.04.054. [59] W. Dai, J. Yu, Y. Wang, Y. Song, F.E. Alam, K. Nishimura, C.-T. Lin, N. Jiang, J. Mater. Chem. A 3 (2015) 4884, doi:http://dx.doi.org/10.1039/C4TA06417H. [60] X. Qi, J. Liang, C. Yu, S. Ma, X. Liu, B. Xu, Mater. Lett. 116 (2014) 68, doi:http://dx. doi.org/10.1016/j.matlet.2013.10.098. [61] J. Pan, X. Yan, X. Cheng, W. Shen, S. Li, X. Cai, Ceram. Int. 42 (2016) 12345, doi: http://dx.doi.org/10.1016/j.ceramint.2016.05.007. [62] C. Vakifahmetoglu, E. Pippel, J. Woltersdorf, P. Colombo, J. Am. Ceram. Soc. 93 (2010) 959, doi:http://dx.doi.org/10.1111/j.1551-2916.2009.03448.x. [63] B. Zhao, X. Guo, W. Zhao, J. Deng, G. Shao, B. Fan, Z. Bai, R. Zhang, ACS Appl. Mater. Interfaces 8 (2016) 28917, doi:http://dx.doi.org/10.1021/acsami.6b10886. [64] B. Zhao, X. Guo, Y. Zhou, T. Su, C. Ma, R. Zhang, CrystEngComm 19 (2017) 2178, doi:http://dx.doi.org/10.1039/c7ce00235a. [65] B. Zhao, X. Guo, W. Zhao, J. Deng, B. Fan, G. Shao, Z. Bai, R. Zhang, Nano Res. 10 (2017) 331, doi:http://dx.doi.org/10.1007/s12274-016-1295-3. [66] K. Sreedeviamma Dijith, S. Vijayan, K. Prabhakaran, K. Peethambharan Surendran, J. Ind. Eng. Chem. 78 (2019) 330, doi:http://dx.doi.org/10.1016/j. jiec.2019.05.037. [67] B. Zhao, L. Liang, Z. Bai, X. Guo, R. Zhang, C. Zuo, J. Deng, CrystEngComm 19 (2017) 6095, doi:http://dx.doi.org/10.1039/C7CE01464C. [68] H. Zhang, J. Zhang, H. Zhang, Compos. A Appl. Sci. Manuf. 38 (2007) 602, doi: http://dx.doi.org/10.1016/j.compositesa.2006.02.014. [69] C. Wang, M. Chen, H. Lei, K. Yao, H. Li, W. Wen, D. Fang, Compos. B Eng. 123 (2017) 19, doi:http://dx.doi.org/10.1016/j.compositesb.2017.05.005. [70] H. Zhang, J. Zhang, H. Zhang, Smart Mater. Struct. 15 (2006) 759, doi:http://dx. doi.org/10.1088/0964-1726/15/3/011. [71] H. Zhang, J. Zhang, H. Zhang, Comput. Mater. Sci. 38 (2007) 857, doi:http://dx. doi.org/10.1016/j.commatsci.2006.05.024. [72] D.X. Yan, P.G. Ren, H. Pang, Q. Fu, M.B. Yang, Z.M. Li, J. Mater. Chem. 22 (2012) 18772, doi:http://dx.doi.org/10.1039/c2jm32692b. [73] H. Liu, H. Tian, H. Cheng, J. Nucl. Mater. 432 (2013) 57, doi:http://dx.doi.org/ 10.1016/j.jnucmat.2012.08.026. [74] H. Zhang, Y. Xu, J. Zhou, J. Jiao, Y. Chen, H. Wang, C. Liu, Z. Jiang, Z. Wang, J. Mater. Chem. C. 3 (2015) 4416, doi:http://dx.doi.org/10.1039/c5tc00405e.