Recent Developments in Carbon Nanotubes-Reinforced Ceramic Matrix Composites: A Review on Dispersion and Densiﬁcation Techniques

: Ceramic matrix composites (CMCs) are well-established composites applied on commercial, laboratory, and even industrial scales, including pottery for decoration, glass–ceramics-based light-emitting diodes (LEDs), commercial cooking utensils, high-temperature laboratory instruments, industrial catalytic reactors, and engine turbine blades. Despite the extensive applications of CMCs, researchers had to deal with their brittleness, low electrical conductivity, and low thermal properties. The use of carbon nanotubes (CNTs) as reinforcement is an effective and efﬁcient method to tailor the ceramic structure at the nanoscale, which provides considerable practicability in the fabrication of highly functional CMC materials. This article provides a comprehensive review of CNTs-reinforced CMC materials (CNTs-CMCs). We critically examined the notable challenges during the synthesis of CNTs-CMCs. Five CNT dispersion processes were elucidated with a comparative study of the established research for the homogeneity distribution in the CMCs and the enhanced properties. We also discussed the effect of densiﬁcation techniques on the properties of CNTs-CMCs. Additionally, we synopsized the outstanding microstructural and functional properties of CNTs in the CNTs-CMCs, namely stimulated ceramic crystallization, high thermal conductivity, bandgap reduction, and improved mechanical toughness. We also addressed the fundamental insights for the future technological maturation and advancement of CNTs-CMCs.


Introduction
Ceramic materials are well known for their corrosion resistance, chemical inertness, high strength, and high thermal stability, which makes these materials suited for applications involving harsh environmental conditions or high-temperature exposure. Related industries such as metallurgical and chemical sectors demand low-density ceramic materials that possess high refractoriness and corrosion resistance for implementation in the revolutionary application. To meet the demand, numerous ceramic synthesis routes have been established: sol-gel [1,2], supercritical fluid [3,4], hydrothermal [5][6][7], solidstate [8,9], and polymer-derived ceramics [10,11]. Even with the superior properties and well-established synthesis techniques, monolithic ceramic materials have yet to become the first option applied in bulk structural components, because of the low thermal transport properties, low electrical properties, and brittleness (low fracture toughness K IC ). Throughout the years, the scientific and engineering communities have countered the elevated thermal, electrical, and mechanical deficiencies of monolithic ceramics by developing Figure 1. Comparison of the rate capacities and capacity retention of carbon nanotubes (CNTs), VO2, and VO2/CNTs (Adapted with permission from ref. [32]. Copyright 2018 Elsevier).
After the discovery of CNTs, CNTs-reinforced CMCs (CNTs-CMCs) have gained the attention of researchers for their excellent microwave absorbance [34], water decontamination [35], ceramic strengthening [36], and oxidation and ablation resistances [37]. To ensure the future of CNTs-CMCs, many scientific studies have been conducted, and many have failed. Researchers determined that the remarkable properties of the CNTs-CMCs were the result of how the key challenges along with the CNTs-CMCs processing were eliminated or solved. Generally, the challenges include achieving a homogenous CNTs distribution within the ceramic (see Figure 2), forming a suitable interfacial adhesion, lowering the thermal degradation, and optimizing the sinterability and densification of the CNTs-CMCs without causing structural damage to the CNTs and ceramic matrix. A previous study revealed an inconsistent improvement in CNTs-CMCs from many aspects, Figure 1. Comparison of the rate capacities and capacity retention of carbon nanotubes (CNTs), VO 2 , and VO 2 /CNTs (Adapted with permission from ref. [32]. Copyright 2018 Elsevier).
After the discovery of CNTs, CNTs-reinforced CMCs (CNTs-CMCs) have gained the attention of researchers for their excellent microwave absorbance [34], water decontamination [35], ceramic strengthening [36], and oxidation and ablation resistances [37]. To ensure the future of CNTs-CMCs, many scientific studies have been conducted, and many have failed. Researchers determined that the remarkable properties of the CNTs-CMCs were the result of how the key challenges along with the CNTs-CMCs processing were eliminated or solved. Generally, the challenges include achieving a homogenous CNTs distribution within the ceramic (see Figure 2), forming a suitable interfacial adhesion, lowering the thermal degradation, and optimizing the sinterability and densification of the CNTs-CMCs without causing structural damage to the CNTs and ceramic matrix. A previous study revealed an inconsistent improvement in CNTs-CMCs from many aspects, e.g., a CNTs reinforcement of 1 wt.% improved the K IC of CNTs-zirconia by about 9% [38], but the same CNTs content (1 wt.%) strengthened the K IC of CNTs-Al 2 O 3 to 36.74% [39]. For microwave application, an identical amount of CNTs addition also showed uncertain performance-the microwave absorption of iron acetylacetonate-modified polymethylsilsesquioxane polymer-derived ceramic composites, PMS(Fe), containing 3% of MWCNTs improved by 29% [34], while 3% of CNTs-Sc 2 Si 2 O 7 reached a 175% absorption [40], as compared to their counterpart. The role of the CNTs reinforcement mechanism has yet to be elucidated and established for high amounts of CNTs contained in the CNTs-CMCs. To explain the enhancement of properties of CNTs-CMCs, the mechanism and possible deformation at low-dimensional scales and subsequent postulated formulations need to be investigated.
Crystals 2021, 11, x FOR PEER REVIEW 4 of 38 e.g., a CNTs reinforcement of 1 wt.% improved the KIC of CNTs-zirconia by about 9% [38], but the same CNTs content (1 wt.%) strengthened the KIC of CNTs-Al2O3 to 36.74% [39]. For microwave application, an identical amount of CNTs addition also showed uncertain performance-the microwave absorption of iron acetylacetonate-modified polymethylsilsesquioxane polymer-derived ceramic composites, PMS(Fe), containing 3% of MWCNTs improved by 29% [34], while 3% of CNTs-Sc2Si2O7 reached a 175% absorption [40], as compared to their counterpart. The role of the CNTs reinforcement mechanism has yet to be elucidated and established for high amounts of CNTs contained in the CNTs-CMCs. To explain the enhancement of properties of CNTs-CMCs, the mechanism and possible deformation at low-dimensional scales and subsequent postulated formulations need to be investigated. Therefore, the present review contains four sections, comprehending the overall knowledge of the CNTs-CMCs processing-densification-property correlations. Section 2 introduces the critical challenges during the CNTs-CMCs fabrication and the end product characteristics. Section 3 discusses various processing methods for the dispersion of CNTs into the ceramic matrix to distinguish the processing's complexity and efficacy. Section 4 contributes to a thorough review of established CNTs-CMCs densification and sintering techniques and the effects on the characteristics of the end product. Finally, the functional properties of CNTs-CMCs are highlighted in Section 5. The simple advantage-disadvantage comparison on the CNTs-CMCs processing and densification and sintering routes are displayed, concluded, and summarized.

Homogenous Dispersion
This is the most difficult challenge when preparing CNTs-CMCs. The nanotubes are most likely bundled or agglomerated, and this causes a non-uniform distribution within the composites. A non-uniform CNTs distribution will lead to the degradation of the fracture toughness upon indentation analysis. A large amount of CNT bundles cause porosity and reduce the contribution of CNTs to the reinforcement mechanism [41]. It also impacts the anisotropic properties of ceramics [42], and it results in a significant difference in radial cracks perpendicular and parallel to the random CNTs orientation where the degree Therefore, the present review contains four sections, comprehending the overall knowledge of the CNTs-CMCs processing-densification-property correlations. Section 2 introduces the critical challenges during the CNTs-CMCs fabrication and the end product characteristics. Section 3 discusses various processing methods for the dispersion of CNTs into the ceramic matrix to distinguish the processing's complexity and efficacy. Section 4 contributes to a thorough review of established CNTs-CMCs densification and sintering techniques and the effects on the characteristics of the end product. Finally, the functional properties of CNTs-CMCs are highlighted in Section 5. The simple advantagedisadvantage comparison on the CNTs-CMCs processing and densification and sintering routes are displayed, concluded, and summarized.

Homogenous Dispersion
This is the most difficult challenge when preparing CNTs-CMCs. The nanotubes are most likely bundled or agglomerated, and this causes a non-uniform distribution within the composites. A non-uniform CNTs distribution will lead to the degradation of the fracture toughness upon indentation analysis. A large amount of CNT bundles cause porosity and reduce the contribution of CNTs to the reinforcement mechanism [41]. It also impacts the anisotropic properties of ceramics [42], and it results in a significant difference in radial cracks perpendicular and parallel to the random CNTs orientation where the degree of CNTs bridging is low [43]. Various dispersion methods have been developed to disperse CNTs homogenously and eventually enhance the CNTs-CMCs properties. Sol-gel [44], hydrothermal/solvothermal [45], and colloidal processing [46] are examples of applications of surface functionalization on CNTs to achieve a homogenous dispersion, while the in situ growth of CNTs by chemical vapor deposition (CVD) can achieve a homogenous CNTs distribution via growing CNTs at the designed growth site (the metal catalyst) within the ceramic precursor [47].

Suitable Interfacial Adhesion between CNTs and Ceramic
Functional properties such as thermal and mechanical properties are highly dependent on the interfacial adhesion and the surface morphology [29]. For instance, strong phonon scattering takes place when there is a weak CNTs-ceramic interaction, resulting in a large thermal resistance. Moreover, a weak CNTs-ceramic interfacial adhesion enhances the crack propagation and negatively influences other toughening mechanisms such as crack bridging [48][49][50]. CNTs with a large lattice oxygen content will deteriorate mechanical properties such as surface hardness and wear depth, where the surfaces between CNTs and the domain ceramic are incompatible and the adhesion is extremely weak [43,51]. A high chemical inertness of pristine CNTs is the major cause of the near-zero chemical affinity of CNTs toward the ceramics. CNTs are neither polar nor lipophilic and ordinary aromatic, where the π-electron population is not balanced by the aromatic hydrogens.
Surface functionalization helps to enhance the quality of CNTs adhesion and increase the chemical affinity toward the matrix. During the functionalization, the CNT walls are partially damaged and attached with functional groups such as carboxyl (-COOH) [41], hydroxyl (-OH) [52], amine (-NH 2 ) [53] groups, and C-F [54] bonds. The attached functional groups provide hydrophilic properties to CNTs for a stable water (or solvent) dispersion [55] and also to improve the interfacial adhesion with the ceramic matrix, which expresses the enhancement in terms of K IC [56], Young's modulus [57], corrosion resistance [58], and thermal conductivity [59]. Hence, the interface engineering is essential to understand and overcome the development of extremely hard, strong, and low-density CNTs-CMCs.

Thermal Degradation
During the densification and sintering of CNTs-CMCs, the temperature, pressure, environment, and duration are important for the crystallization of the ceramic and the maintaining of the primitive nature of CNTs. Studies in the literature have shown that high pressures, high sintering temperatures, and prolonged sintering durations can negatively affect the CNTs' nature and cause allotropic transformation and thermophysical damage [60]. Structural damage on the CNT walls was observed in CNTs-CMCs when they underwent spark plasma sintering (high temperature and pressure) and pressureless sintering (long sintering duration) [61].

Powder
Powder processing is an established CNTs-CMCs fabrication technique for the production of near-neat shape products. This technique is also useful for its batch production and the integrated weight fractions of the raw material. Powder processing requires two or three steps. Figure 3 compares the two-step and three-step powder processing. The two-step powder processing involves the mixing of CNTs and ceramic powder, followed by in situ densification and sintering such as spark plasma sintering (SPS) and hot-press sintering (HPS), while the three-step processing involves mixing, densification, and sintering, such as pressureless sintering (PLS) and microwave-assisted sintering (MAS). Several techniques could be applied at the mixing process: manual mixing [62], planetary milling [63], ball milling [38,64,65], and high-energy ball milling [66]. The mixing is commonly performed using dry CNTs and ceramic powder, but it often requires a solvent like alcohol or acetone as a medium, known as wet-mixing, where the solvent is later evaporated. Removing the aggregation and achieving a homogenous CNTs distribution are the ultimate goals of this step. This processing technique is suitable for preparing CMCs with large Crystals 2021, 11, 457 6 of 37 amounts of CNTs, and it is applicable to the thermal enhancement [62,67] and mechanical enhancement [68,69] of CNTs-CMCs.  [38,64,65], and high-energy ball milling [66]. The mixing is commonly performed using dry CNTs and ceramic powder, but it often requires a solvent like alcohol or acetone as a medium, known as wet-mixing, where the solvent is later evaporated. Removing the aggregation and achieving a homogenous CNTs distribution are the ultimate goals of this step. This processing technique is suitable for preparing CMCs with large amounts of CNTs , and it is applicable to the thermal enhancement [62,67] and mechanical enhancement [68,69] of CNTs-CMCs. Ye et al., improved the KIC and flexural strength of barium aluminosilicate (BAS) glass-ceramic composites by adding MWCNTs through wet powder processing [70]. They sonicated and ball-milled the MWCNTs-BAS in ethanol into the MWCNTs-BAS slurry. Next, the slurry was dried and hot-pressed (1600 °C, 1 h, 20 MPa, N2 atmosphere) into ceramic pellets for mechanical analysis. They noticed that the addition of 10 vol.% MWCNTs increased the flexural strength from 84 to 245 MPa (about 300%) and the KIC increased from 1.22 to 2.97 MPa·m 1/2 . Good MWCNTs-BAS interfacial bonding, the high elastic modulus of CNTs, and the CNTs bridging contributed to the enhanced mechanical properties. By contrast, the relative density of 15 vol.% CNTs-added composites decreased by 3% and the flexural strength only achieved 169 MPa, which was due to the effect of the agglomerated CNTs (see Figure 4). Moreover, Chan et al. reduced the lattice strain of zinc silicate CMCs upon the addition of CNTs [71]. They prepared CNTs through CVD on a cobalt oxide metal catalyst and through the manual mixing of the CNTs with zinc silicate glass powder. The CNTs-zinc silicate glass powder was further sintered in Ar atmosphere, and the zinc-silicate glassy phase changed to the polycrystalline willemite phase (Zn2SiO4). Through the phase analysis, CNTs showed an insignificant phase alteration to the domain Zn2SiO4 crystal structure but only introduced the graphite (C) and cobalt oxide (CoO) phases. They also noticed that the lattice strain of 3 wt.% CNTs-Zn2SiO4 reduced by 22.85% and the crystallite size of 1 wt.% CNTs-Zn2SiO4 CMC increased by 5.65%, as compared to the CNTs-free Zn2SiO4 composite. Ye et al., improved the K IC and flexural strength of barium aluminosilicate (BAS) glass-ceramic composites by adding MWCNTs through wet powder processing [70]. They sonicated and ball-milled the MWCNTs-BAS in ethanol into the MWCNTs-BAS slurry. Next, the slurry was dried and hot-pressed (1600 • C, 1 h, 20 MPa, N 2 atmosphere) into ceramic pellets for mechanical analysis. They noticed that the addition of 10 vol.% MWCNTs increased the flexural strength from 84 to 245 MPa (about 300%) and the K IC increased from 1.22 to 2.97 MPa·m 1/2 . Good MWCNTs-BAS interfacial bonding, the high elastic modulus of CNTs, and the CNTs bridging contributed to the enhanced mechanical properties. By contrast, the relative density of 15 vol.% CNTs-added composites decreased by 3% and the flexural strength only achieved 169 MPa, which was due to the effect of the agglomerated CNTs (see Figure 4). Moreover, Chan et al. reduced the lattice strain of zinc silicate CMCs upon the addition of CNTs [71]. They prepared CNTs through CVD on a cobalt oxide metal catalyst and through the manual mixing of the CNTs with zinc silicate glass powder. The CNTs-zinc silicate glass powder was further sintered in Ar atmosphere, and the zinc-silicate glassy phase changed to the polycrystalline willemite phase (Zn 2 SiO 4 ). Through the phase analysis, CNTs showed an insignificant phase alteration to the domain Zn 2 SiO 4 crystal structure but only introduced the graphite (C) and cobalt oxide (CoO) phases. They also noticed that the lattice strain of 3 wt.% CNTs-Zn 2 SiO 4 reduced by 22.85% and the crystallite size of 1 wt.% CNTs-Zn 2 SiO 4 CMC increased by 5.65%, as compared to the CNTs-free Zn 2 SiO 4 composite.

Colloidal
Colloidal processing involves several physical/chemical practices to disperse CNTs with water/solvent, eventually turning into CNT nanofluid. The options of the CNT dispersant are limited because it needs to meet the dispersion compatibility and the sinterability upon the densification and sintering with the ceramic domain. Critical dispersion steps, such as sonication, magnetic stirring, and chemical oxidation, are taken for the CNTs nanofluid production. The CNT nanofluid is added to the ceramic powder and densified for the final CNTs-CMCs product. Figure 5 displays the schematic of colloidal processing.

Colloidal
Colloidal processing involves several physical/chemical practices to disperse CNTs with water/solvent, eventually turning into CNT nanofluid. The options of the CNT dispersant are limited because it needs to meet the dispersion compatibility and the sinterability upon the densification and sintering with the ceramic domain. Critical dispersion steps, such as sonication, magnetic stirring, and chemical oxidation, are taken for the CNTs nanofluid production. The CNT nanofluid is added to the ceramic powder and densified for the final CNTs-CMCs product. Figure 5 displays the schematic of colloidal processing. N,N-Dimethylformamide (DMF) was first applied to disperse the low concentration of MWCNTs and SWCNTs with near-zero disruption to the structure [72]. Acetone was applied onto oxidized MWCNTs to achieve dispersion where acetone formed the solubilizing phase with the hydroxylated modified MWCNTs surface [73]. Carboxylated MWCNTs (MWCNTs-COOH) can be dispersed in diethyl ether with the addition of 4fluoroaniline in a N2 atmosphere [54]. Molecular dynamic (MD) simulations [74] and experimental studies [75] showed that pyrogallol and 1-naphthol had high binding coefficients to MWCNTs and increased threefold as the experimented MWCNTs diameter reduced from 100 to 10 nm (effect of agglomeration increased as diameter decreased). Solvents with a high density of aromatic structure are mostly dispersed SWCNTs at ease with the aid of planar polycyclic aromatic hydrocarbons as compared to linear structure solvents [76]. Previous studies in the literature on the dispersion medium, surfactant, and other additives are displayed in Table 1.  N,N-Dimethylformamide (DMF) was first applied to disperse the low concentration of MWCNTs and SWCNTs with near-zero disruption to the structure [72]. Acetone was applied onto oxidized MWCNTs to achieve dispersion where acetone formed the solubilizing phase with the hydroxylated modified MWCNTs surface [73]. Carboxylated MWCNTs (MWCNTs-COOH) can be dispersed in diethyl ether with the addition of 4-fluoroaniline in a N 2 atmosphere [54]. Molecular dynamic (MD) simulations [74] and experimental studies [75] showed that pyrogallol and 1-naphthol had high binding coefficients to MWC-NTs and increased threefold as the experimented MWCNTs diameter reduced from 100 to 10 nm (effect of agglomeration increased as diameter decreased). Solvents with a high density of aromatic structure are mostly dispersed SWCNTs at ease with the aid of planar polycyclic aromatic hydrocarbons as compared to linear structure solvents [76]. Previous studies in the literature on the dispersion medium, surfactant, and other additives are displayed in Table 1.

Sol-Gel
Sol-gel processing techniques are effective for fabricating CNTs-CMCs with a homogenous CNT dispersion and a uniform size distribution of ceramic crystals. CNTs are sonicated in an aqueous solution (or solvent) with minimum surfactant addition (known as the CNT sol). In order to prepare CNTs-CMCs gel, a series of solutions (ceramic precursor, gelling agent, and pH control agent) are added into the CNT sol, and ageing or the gelation process is carried out. The CNTs-ceramic gel is then dried, calcined at high temperature with (without) inert gas, and left with the CNTs-ceramic powder (see Figure 6).
Sol-gel processing techniques are effective for fabricating CNTs-CMCs with a homogenous CNT dispersion and a uniform size distribution of ceramic crystals. CNTs are sonicated in an aqueous solution (or solvent) with minimum surfactant addition (known as the CNT sol). In order to prepare CNTs-CMCs gel, a series of solutions (ceramic precursor, gelling agent, and pH control agent) are added into the CNT sol, and ageing or the gelation process is carried out. The CNTs-ceramic gel is then dried, calcined at high temperature with (without) inert gas, and left with the CNTs-ceramic powder (see Figure 6).  [89]. First, they prepared an Sr 2+ aqueous solution and TiO3 2− ethanol solution, with acetyl acetone as the chelating agent. Then, they added the acidified MWCNTs with sodium dodecyl benzene sulfonate (SDBS) as the CNT surfactant into the as-prepared SrTiO3 solution. The SrTiO3-MWCNTs composites were stirred for 3 h at 40 °C (gelation), dried overnight, and calcined at 800 °C before compounding with EP. The obtained dielectric constant of EP-compounded 11 vol.% SrTiO3-MWCNTs reached 283 and the dielectric loss was 0.07 at 1 kHz, where the dielectric constant of EP-compounded 11 vol.% neat MWCNTs reached 1600 and the dielectric loss was 25. They deduced that the SrTiO3 coating acted as barriers between CNTs, and it resulted in a low dielectric loss, which improved the dielectric  [89]. First, they prepared an Sr 2+ aqueous solution and TiO 3 2− ethanol solution, with acetyl acetone as the chelating agent. Then, they added the acidified MWCNTs with sodium dodecyl benzene sulfonate (SDBS) as the CNT surfactant into the as-prepared SrTiO 3 solution. The SrTiO 3 -MWCNTs composites were stirred for 3 h at 40 • C (gelation), dried overnight, and calcined at 800 • C before compounding with EP. The obtained dielectric constant of EP-compounded 11 vol.% SrTiO 3 -MWCNTs reached 283 and the dielectric loss was 0.07 at 1 kHz, where the dielectric constant of EP-compounded 11 vol.% neat MWCNTs reached 1600 and the dielectric loss was 25. They deduced that the SrTiO 3 coating acted as barriers between CNTs, and it resulted in a low dielectric loss, which improved the dielectric properties. The tribological properties of the metal/alloy surface can also be enhanced when it is coated with a thin layer of CNTs-CMCs. Xu et al. coated a 0.25 wt.% CNTs-Al 2 O 3 -ZrO 2 -ZrSiO 4 ceramic sol-gel mixture on A4 steel to enhance the wear rate at an escalated temperature [90]. Acidified CNTs were first dispersed in SDBS solution, transferred to the aluminium dihydrogen phosphate Al(PO 2 (OH) 2 ) 3 solution (containing the ceramic powder), and cured for 5 h before brushing on the steel surface. It showed that the CNT fillers (0.25 wt.%) effectively lowered the CNTs-CMCs-coated steel's wear rate with temperature: 57% (100 • C), 67% (300 • C), and 87.5% (500 • C), as compared to the bare ceramic coating.
Moreover, sol-gel processing is a favourable technique during the carbon-ceramic electrode (CCE) fabrication for its high CNT dispersion homogeneity. CNTs incorporating CCE were developed to detect nicotinamide adenine dinucleotide (NADH) [91], sulfonamide drugs [92], and purine derivative [93], as well as for wastewater treatment [94]. Ferrag et al. developed CNTs-Au-CCEs to detect the purine derivatives, uric acid (UA), xanthine (XA), and caffeine (CA) [93]. They mixed 3-mercaptopropyl-methyl-dimethoxy silane (MPDS, Si-based sol-gel network), hydrochloric acid, and the carbon mixture (50 mg CNTs: 150 mg graphite powder) in methanol, and they dried out the liquid mixture. The dried ceramic product was impregnated with Au solution and then molded in a narrow glass tube. Comparing the CNTs-Au-CCEs and the bare CCE, the diffusion coefficients for UA, XA, and CA improved: 3.5 times (UA), 1.56 times (XA), and 2.26 times (CA). The limit of detection (LOD) values of the CNTs-Au-CCE for UA, XA, and CA were 50, 63, and 354 nM, where the limits of quantification (LOQ) were 167, 209, and 1179 nM, respectively. Table 2 shows the dispersion medium, dispersant (or surfactant), gelling agent, and the gelation process when preparing the CNTs-CMCs gel.

In Situ CNTs Growth
In situ synthesis is the attempt for CNTs-CMCs to grow CNTs on the ceramic matrix directly, and it is widely performed through chemical vapour deposition (CVD). Generally, the ceramic powder is prepared with metal salt (such as metal nitrate or acetate salt) in an aqueous solution or solvent to ensure complete metal ion impregnation within the ceramic domain. Next, the ceramic-metal solution is reduced at high temperature with the presence of H 2 as the metal salt reduces to metal oxides and further reduces to metal nanoparticles. The nanoparticles act as the active growth site for CNTs, so the carbon solubility is critical. The commonly used transition metal catalysts are Fe, Ni, and Co for their moderate-to-high carbon solubility at elevated temperatures. Carbon atoms for the CNTs growth come from the carbon precursor (CH 4 , C 2 H 2 , ethanol, etc.) and are carried by the flowing gas (such as Ar, N 2 , He) within the CVD system (see Figure 7). The deposition of carbon atoms (carbonaceous molecule from the carbon precursor) involves several mechanisms: carbon adsorption on the catalyst surface, the dissolution of carbon molecules (segregation of hydrogen bonding), carbon diffusion into the catalyst, the formation of the carbon network and the growth of CNTs. Table 3 displays the optimum parameters for in situ CNTs growth on ceramic. flowing gas (such as Ar, N2, He) within the CVD system (see Figure 7). The deposition of carbon atoms (carbonaceous molecule from the carbon precursor) involves several mechanisms: carbon adsorption on the catalyst surface, the dissolution of carbon molecules (segregation of hydrogen bonding), carbon diffusion into the catalyst, the formation of the carbon network and the growth of CNTs. Table 3 displays the optimum parameters for in situ CNTs growth on ceramic. Chen and co-workers prepared Si3N4-CNTs composites through in-situ CNTs growth via CVD [102]. The ceramic possessed an extraordinary electromagnetic shielding effect over the X-band frequencies (f = 8.2-12.4 GHz), where the total shielding effectiveness increased from 6.0 (neat Si3N4) to 30.4 dB (2.7 wt.% CNT). They impregnated Si3N4 with 10 wt.% cobalt acetate tetrahydrate (CoAc·4H2O) and fed the CVD system with Ar-carried Chen and co-workers prepared Si 3 N 4 -CNTs composites through in-situ CNTs growth via CVD [102]. The ceramic possessed an extraordinary electromagnetic shielding effect over the X-band frequencies (f = 8.2-12.4 GHz), where the total shielding effectiveness increased from 6.0 (neat Si 3 N 4 ) to 30.4 dB (2.7 wt.% CNT). They impregnated Si 3 N 4 with 10 wt.% cobalt acetate tetrahydrate (CoAc·4H 2 O) and fed the CVD system with Ar-carried acetone vapor (CH 3 ) 2 CO. Vapor deposition took place at 550 • C and lasted for 1 h. As per 10 wt.% CoAc catalyst, they managed to produce 2.7 wt.% feather-like CNTs (outer diameter, d = 10 nm) with a 0.75 I D /I G ratio (see Figure 8). The growth of CNTs interconnected the Si 3 N 4 grains and improved the shielding properties. The resulting Si 3 N 4 -2.7 wt.% CNTs composites improved the absorbance shielding effectiveness (SE A ) from >5 (neat) to 22 dB, and the reflectance shielding effectiveness (SE R ) increased from 2 to 8 dB over the X-band. To improve the Cu 2+ removal of the porous Al2O3 ceramic membrane, Tofighy and Mohammadi grew CNTs on the membrane by using an organometallic catalyst (Feroccene, Fe(C5H5)2) and cyclohexanol HOCH(CH2)5 as carbon feedstock [103]. Under transmission electron microscope (TEM) observation, the outer diameter of CNTs ranged from 30 to 40 nm, and it showed a rigid and flawless structure. The Al2O3-CNTs membrane was tested for the permeation flux and rejection of Cu 2+ ions in a weak acidic environment (pH = 6) for 10 min. The permeation flux reduced from 523.65 to 45.98 kg/m 2 h, while the Cu 2+ rejection increased from 2.76% to 20.27%. They discussed that the reduced permeation flux and the increased Cu 2+ rejection were the outcomes of the CNTs filling the membrane pore void, as it increased the membrane resistance. Ding et al. grew CNTs on FeCl3-implanted polysiloxane at 800-1500 °C for high electromagnetic waves (EMW)-absorbing applications [104]. Instead of using gas-carried precursor vapor, they dissolved the FeCl3polysiloxane in ethanol and carried out the reaction for 2 h in Ar atmosphere. During the CVD at 900 °C, the ID/IG ratio was 1.01, and the measured CNTs crystallite size (La, Equation (1), where λ is the laser wavelength of the Raman spectroscope) [105,106] was 16.65 To improve the Cu 2+ removal of the porous Al 2 O 3 ceramic membrane, Tofighy and Mohammadi grew CNTs on the membrane by using an organometallic catalyst (Feroccene, Fe(C 5 H 5 ) 2 ) and cyclohexanol HOCH(CH 2 ) 5 as carbon feedstock [103]. Under transmission electron microscope (TEM) observation, the outer diameter of CNTs ranged from 30 to 40 nm, and it showed a rigid and flawless structure. The Al 2 O 3 -CNTs membrane was tested for the permeation flux and rejection of Cu 2+ ions in a weak acidic environment (pH = 6) for 10 min. The permeation flux reduced from 523.65 to 45.98 kg/m 2 h, while the Cu 2+ rejection increased from 2.76% to 20.27%. They discussed that the reduced permeation flux and the increased Cu 2+ rejection were the outcomes of the CNTs filling the membrane pore void, as it increased the membrane resistance. Ding et al. grew CNTs on FeCl 3 -implanted polysiloxane at 800-1500 • C for high electromagnetic waves (EMW)absorbing applications [104]. Instead of using gas-carried precursor vapor, they dissolved the FeCl 3 -polysiloxane in ethanol and carried out the reaction for 2 h in Ar atmosphere. During the CVD at 900 • C, the I D /I G ratio was 1.01, and the measured CNTs crystallite size (L a , Equation (1), where λ is the laser wavelength of the Raman spectroscope) [105,106] was 16.65 nm. However, more defects formed on the CNTs, causing the I D /I G ratio to increase to 2.46 and L a to decrease to 6.84 nm when the reaction temperature was 1500 • C. Transmission electron microscope (TEM) analysis revealed that the CNTs experienced a structure breakdown, and the SiC grew in various morphologies such as microspheres, needle-like, and nanowires at high temperature. In this case, CNTs provide a casting mold for the growth of SiC micro-and nanostructures, and the formed SiC showed a low reflection loss of −58.37 dB at 10.11 GHz. In situ synthesis is a simple method to synthesis large volumes of CNTs-embedded CNTs-CMCs with a minimum amount of bundled CNTs and agglomeration. Besides, this method provides CNTs-CMCs with a highly ordered unidirectional orientation, which is enables the comparison of the toughening mechanism between microfibers-reinforced ceramic and CNTs-CMCs. However, the quality of CNTs within the ceramic is not consistent, because the CNTs growth mechanism on the active metal nanoparticles is greatly dependent on various parameters-temperature, deposition duration, amount and purity of metal catalyst, gas flow, and carbon precursor.

Hydrothermal/Solvothermal Processes
Hydrothermal and solvothermal processes are well-established CNTs-CMCs techniques for homogenous CNT dispersions with a low number of impurities. Generally, acid-functionalized CNTs are dispersed in water (hydrothermal) or solvent (solvothermal), together with the ceramic salt with (without) the surfactant. The surfactant reduces the interfacial resistance among the ceramic salt, CNTs, and the dispersion medium. The aqueous mixture of CNTs and salt is magnetically stirred or sonicated for CNTs disentanglement and to improve the CNTs-CMCs homogeneity. Next, the mixture is stored in an autoclave and heated from 100 to 250 • C for several hours to several days, highly dependent on the crystallization of the domain ceramic ( Figure 9). Zhao et al. prepared low-bandgap MWCNTs-Bi2S3 composites for methylene blue (MB) wastewater treatment [109]. Before the hydrothermal processing, they applied acid treatment (stirred with HNO3, 80 °C, 12 h) to functionalize CNTs. The CNT wall with the carboxyl group -COOH resulted in a higher miscibility with water. The composite was prepared by stirring the bismuth nitrate (Bi(NO3)3·5H2O), urea, and MWCNTs in deionized water and heating in an autoclave at 120 °C for 12 h. The hydrothermally treated composites were centrifuged, rinsed with alcohol-water, and dried overnight into MWCNTs-Bi2S3 ceramic powder. Using 5 wt.% MWCNTs, the bandgap value of Bi2S3 powder reduced from 1.245 to 0.875 eV, and the Brunauer-Emmett-Teller (BET) surface area increased from 35.2 to 36.1 m 2 /g. With the low-bandgap value and larger surface area, the photocatalytic activity (the MB degradation) under visible-light irradiation reached 90.75% in the first cycle and was maintained at about 75% after the fourth cycle. Besides, MWCNTs as the filler effectively improved the sensing device, e.g., ammonia sensor. Singh et al. shortened the response and recovery time of a MoS2-based ammonia sensor by adding a relatively low amount of CNTs during the hydrothermal process [45]. They prepared the ceramic ammonia sensor by sonicating 1 g of ammonium tetra-thiomolybdate (NH4)2MoS4, 10 mL of hydrazine monohydrate N2H4·H2O, and 3 mg of MWCNTs in deionized water. Next, the mixture was transferred to an autoclave, heated at 200 °C for 12 h, and dried in an oven for 2 h. The obtained black powder (after drying) showed a larger surface area of 20.23 m 2 /g, where the surface area of neat MoS2 was 4.22 m 2 /g. For the ammonia response determination, neat MoS2 required 400 s to detect the NH3 content when the concentration was low as 150 ppm, while the MoS2-MWCNTs hybrid was more capable of rapid detection (response time = 65 s), which was six times faster than the neat one and was due to the larger surface area. Li et al. discussed the dielectric performance of the poly(1-butene) polymer filled with modified polydopamine and Ba(Zr0.2Ti0.8)O3-MWCNTs as a diaphragm in capacitors [110]. The Ba(Zr0.2Ti0.8)O3-MWCNTs were prepared through sol-gel processing on barium acetate Ba(COOCH3)2 (dissolved in acetic acid and DI) and tetrabutyl titanate Ti(CH3(CH2)3OH)4 (dissolved with isopropanol zirconium in glycol methyl ether), and then by carrying out the hydrothermal processing with MWCNTs in NaOH aqueous solution. The temperature and duration of the hydrothermal process was set at 180 °C for 24 h, and the obtained Ba(Zr0.2Ti0.8)O3-MWCNTs was then transferred on the poly(1-butene) polymer film. The analysis showed that the dielectric constant of the film improved from 4.4 to 28.40 with the addition of 0.5 vol.% CNTs at 1 kHz, and the dielectric loss was 0.0077, which makes this composite a potential candidate as an intermediate phase and insulating layer.
In addition to hydrothermal treatment, solvothermal processing also offers promising results. Vadivel et al., improved the BiPO4 nanorod's capacitance and its photocatalytic activity on methyl orange degradation by introducing MWCNTs via the Zhao et al. prepared low-bandgap MWCNTs-Bi 2 S 3 composites for methylene blue (MB) wastewater treatment [109]. Before the hydrothermal processing, they applied acid treatment (stirred with HNO 3 , 80 • C, 12 h) to functionalize CNTs. The CNT wall with the carboxyl group -COOH resulted in a higher miscibility with water. The composite was prepared by stirring the bismuth nitrate (Bi(NO 3 ) 3 ·5H 2 O), urea, and MWCNTs in deionized water and heating in an autoclave at 120 • C for 12 h. The hydrothermally treated composites were centrifuged, rinsed with alcohol-water, and dried overnight into MWCNTs-Bi 2 S 3 ceramic powder. Using 5 wt.% MWCNTs, the bandgap value of Bi 2 S 3 powder reduced from 1.245 to 0.875 eV, and the Brunauer-Emmett-Teller (BET) surface area increased from 35.2 to 36.1 m 2 /g. With the low-bandgap value and larger surface area, the photocatalytic activity (the MB degradation) under visible-light irradiation reached 90.75% in the first cycle and was maintained at about 75% after the fourth cycle. Besides, MWCNTs as the filler effectively improved the sensing device, e.g., ammonia sensor. Singh et al. shortened the response and recovery time of a MoS 2 -based ammonia sensor by adding a relatively low amount of CNTs during the hydrothermal process [45]. They prepared the ceramic ammonia sensor by sonicating 1 g of ammonium tetra-thiomolybdate (NH 4 ) 2 MoS 4 , 10 mL of hydrazine monohydrate N 2 H 4 ·H 2 O, and 3 mg of MWCNTs in deionized water. Next, the mixture was transferred to an autoclave, heated at 200 • C for 12 h, and dried in an oven for 2 h. The obtained black powder (after drying) showed a larger surface area of 20.23 m 2 /g, where the surface area of neat MoS 2 was 4.22 m 2 /g. For the ammonia response determination, neat MoS 2 required 400 s to detect the NH 3 content when the concentration was low as 150 ppm, while the MoS 2 -MWCNTs hybrid was more capable of rapid detection (response time = 65 s), which was six times faster than the neat one and was due to the larger surface area. Li et al. discussed the dielectric performance of the poly(1-butene) polymer filled with modified polydopamine and Ba(Zr 0.2 Ti 0.8 )O 3 -MWCNTs as a diaphragm in capacitors [110]. The Ba(Zr 0.2 Ti 0.8 )O 3 -MWCNTs were prepared through sol-gel processing on barium acetate Ba(COOCH 3 ) 2 (dissolved in acetic acid and DI) and tetrabutyl titanate Ti(CH 3 (CH 2 ) 3 OH) 4 (dissolved with isopropanol zirconium in glycol methyl ether), and then by carrying out the hydrothermal processing with MWCNTs in NaOH aqueous solution. The temperature and duration of the hydrothermal process was set at 180 • C for 24 h, and the obtained Ba(Zr 0.2 Ti 0.8 )O 3 -MWCNTs was then transferred on the poly(1-butene) polymer film. The analysis showed that the dielectric constant of the film improved from 4.4 to 28.40 with the addition of 0.5 vol.% CNTs at 1 kHz, and the dielectric loss was 0.0077, which makes this composite a potential candidate as an intermediate phase and insulating layer.
In addition to hydrothermal treatment, solvothermal processing also offers promising results. Vadivel et al., improved the BiPO 4 nanorod's capacitance and its photocatalytic activity on methyl orange degradation by introducing MWCNTs via the solvothermal process [111]. They dispersed bismuth nitrate (Bi(NO 3 ) 3 , Bi 3+ precursor), monoammonium phosphate (NH 4 H 2 PO 4 , PO 4 3− precursor), and untreated MWCNTs (with PVP as a surfactant) in ethylene glycol (EG). The weight ratio of the mixture Bi 3+ :PO 4 3− :MWCNTs was 1.5:0.85:0.05. The mixture was then sealed in an autoclave and heated at 160 • C for 12 h. The measured specific capacitances (current density = 1 A/g) of neat MWCNTs and pristine BiPO 4 were 45 and 292 F/g, respectively, while the BiPO 4 -MWCNTs reached 368 F/g at the identical current density. The stable photocatalytic activity of the composite under indigo irradiation (464 nm) also reached 95% of methyl orange degradation after a 150 min test time. The enhanced capacitance and photocatalytic activities were due to the large surface area of BiPO 4 -MWCNTs (38.25 m 2 /g), which was 1.53 times larger than that of pristine BiPO 4 . Solvothermal processing also enables the oxide coating on the wall of CNTs. Zhang and Hao coated layers of rare earth oxide on the CNTs wall toward the effective thermal decomposition of ammonium perchlorate (AP, a type of oxidizer in solid propellant) [112]. They enlarged the surface area of CNTs with HNO 3 , and they modified the CNTs polarity with cetyl trimethylammonium bromide (CTAB) and poly(sodium 4styrenesulfonate) (PSS). The rare earth nitrate salt (Y 3+ , Nd 3+ , and Sm 3+ ), modified CNT, sodium acetate (NaAc), and polyethylene glycol (PEG) were dissolved and stirred in EG, and they were heated in the reaction system in a 200 • C oil bath for 2.5 h. The AP's thermal decomposition peak temperature reduced from 452 to 326.5 • C with the addition of 2 wt.% Y 2 O 3 -CNTs, and the DTA heat released increased from 0.281 to 2.308 kJ/g. It showed that the thermal decomposition of AP decreased because of the large surface area and the partially filled 4d orbital in the rare earth material, which is favorable for electron transfer. Tables 4 and 5 show the optimized parameters of the surfactant-CNTs ratio and the hydrothermal (solvothermal) processing.  EG -180 3 [32] Hydrothermal and solvothermal processes bring benefits in preparing high-quality CNTs-CMCs with low synthesis temperatures. However, the processing duration (involving heating, cooling, and drying) is longer than those of other processing techniques. In addition, the starting materials such as the ceramic precursor and the surfactants (or dispersants) for CNT dispersion are expensive, and the final yield is low because the molecular ratio of ceramics in the precursor is mostly low. For example, the stable hydrogen evolution reaction (HER) candidate, molybdenum sulfide/reduced graphene oxide-carbon nanotube composites (MoS 2 /RGO/CNTs), was where the MoS 2 compound was produced from ammonium thiomolybdate salt (NH 4 ) 2 MoS 4 [113]. The weight ratio of one mole of MoS 2 is about 61% per one mole of the (NH 4 ) 2 MoS 4 . A small amount of end product explains that the hydrothermal and solvothermal processes are not favorable in CNTs-CMCs mass production, but they are fit for research purposes and the high-purity requirement.

Densification and Sintering Techniques
Densification and sintering are the critical treatments for CNTs-CMCs interaction by shaping the powder and supplying energy (thermal/electrical) to the matrix itself. This process inputs high energy in promoting the crystallization of the domain ceramic crystal and bond-breaking of C=C bonds of CNTs. Generally, a CNTs-CMCs pellet with a desired dimension is produced through an applied pressure before or during the thermal treatment. Hereafter, four common techniques are reviewed: spark plasma sintering (SPS), hot-press sintering (HPS), pressureless sintering (PLS), and microwave-assisted sintering (MAS). The rapid heating in SPS and MAS offers a time-efficient route by heating the CNTs-CMCs at a high temperature for a short duration (<15 min); however, the setup and fabrication are expensive. On the other hand, HPS and PLS produced a cheaper composite with flexible size and shape, yet at the risk of a lower density and long sintering duration (>1 h).

Spark Plasma Sintering (SPS)
The success of spark plasma sintering (SPS) has been highlighted by noteworthy achievements of, for example, an outstanding increase in tensile strength of ceramic material composed of CNTs. During the SPS process, the CNTs-ceramic mixed powder is sintered under the simultaneous effect of a pulsed direct current (PDC) and an applied pressure in a vacuum chamber. The powder is placed in a die, pellet-pressed by the plunger, and current-heated by passing a DC [116] through the die and the sample (see Figure 10). The heating rate, the application of pressure, and the current are the three main parameters in determining the microstructural and mechanical properties of the sintered pellet. produced from ammonium thiomolybdate salt (NH4)2MoS4 [113]. The weight ratio of one mole of MoS2 is about 61% per one mole of the (NH4)2MoS4. A small amount of end product explains that the hydrothermal and solvothermal processes are not favorable in CNTs-CMCs mass production, but they are fit for research purposes and the high-purity requirement.

Densification and Sintering Techniques
Densification and sintering are the critical treatments for CNTs-CMCs interaction by shaping the powder and supplying energy (thermal/electrical) to the matrix itself. This process inputs high energy in promoting the crystallization of the domain ceramic crystal and bond-breaking of C=C bonds of CNTs. Generally, a CNTs-CMCs pellet with a desired dimension is produced through an applied pressure before or during the thermal treatment. Hereafter, four common techniques are reviewed: spark plasma sintering (SPS), hot-press sintering (HPS), pressureless sintering (PLS), and microwave-assisted sintering (MAS). The rapid heating in SPS and MAS offers a time-efficient route by heating the CNTs-CMCs at a high temperature for a short duration (<15 min); however, the setup and fabrication are expensive. On the other hand, HPS and PLS produced a cheaper composite with flexible size and shape, yet at the risk of a lower density and long sintering duration (>1 h).

Spark Plasma Sintering (SPS)
The success of spark plasma sintering (SPS) has been highlighted by noteworthy achievements of, for example, an outstanding increase in tensile strength of ceramic material composed of CNTs. During the SPS process, the CNTs-ceramic mixed powder is sintered under the simultaneous effect of a pulsed direct current (PDC) and an applied pressure in a vacuum chamber. The powder is placed in a die, pellet-pressed by the plunger, and current-heated by passing a DC [116] through the die and the sample (see Figure 10). The heating rate, the application of pressure, and the current are the three main parameters in determining the microstructural and mechanical properties of the sintered pellet.  Zhan et al., was the first to perform the SPS process on a CNTs-ceramic composite. They fabricated a SWCNTs-aluminum oxide composite with an applied pressure of 63 MPa and a high-temperature plasma of 1150 • C in a 3 min sintering period. It was noticeable that the electrical conductivity increased from 1 × 10 −12 to 1050 S/m and the fracture toughness, K IC , increased from 3.3 to 9.7 MPa·m 1/2 [117,118]. These pioneer works had inspired the present researchers in using the SPS process to fabricate CNTs-ceramic composites. Recently, homogenization was first applied to avoid the aggregation of CNTs during the SPS process. Lamnini et al. applied 5 h high-efficiency attrition milling at 4000 rpm on the MWCNTs-8YSZ (8 vol.% yttria-stabilized zirconia) powder, and they applied SPS to the powder to form pellets at 1400 • C with a 50 MPa plunging pressure for 5 min. The enhanced K IC of the MWCNTs-8YSZ composites (3.2 MPa·m 1/2 ) was attributed to the MWCNTs crack deflection and crack bridging between ceramic particles [38]. Momohjimoh et al. compared MWCNTs and SiC as dopants in the alumina matrix for electrical conductivity performance. After performing 50 MPa and >1400 • C sintering on the composites, 2 wt.% MWCNTsalumina achieved a low-density ceramic matrix with a high electrical conductivity of 101.118 S/m [119]. Table 6 shows the reported research on fabricating CNTs-CMCs through SPS technique.

Hot-Press Sintering (HPS)
Pressure-assisted sintering/hot-press sintering is the "in situ temperature and pressure sintering" technique, which turns CNTs-CMCs powder into pellets or films. The schematic setup is shown in Figure 11. The graphite dies, inside a vacuum chamber, hold the mixed powder. The mixed powder is pressed to several MPa and heated up for several hours. This technique is suitable for ceramics that need extreme-condition fabrication. Through pressure-assisted sintering, the thermo-physical and mechanical strength properties of the sintered sample are notably enhanced. Crystals 2021, 11, x FOR PEER REVIEW 18 of 38

Pressureless Sintering (PLS)
Pressureless sintering (PLS) differs from pressure-assisted sintering during the thermal treatment. For pressure-assisted sintering (such as SPS and HPS), the densification process of CNTs-CMCs powder takes place in the furnace during the sintering process. For the PLS route, the CNTs-CMCs mixed powder is uniaxially pressed to pellets before the sintering, with a manual hydraulic press or other pelleting apparatus. Next, the pellet is placed in an electrical/tubular furnace and fired to form the final product. Only during the specific temperature firing/sintering, the high energy breaks the strong covalent and ionic bonding within the ceramics and opens a path for the vacancy diffusion mechanism [128]. The gas type/flow is optional and greatly dependent on the thermal stability and purity of CNTs at different elevating temperatures. The schematic setup is displayed in Figure 13.

Pressureless Sintering (PLS)
Pressureless sintering (PLS) differs from pressure-assisted sintering during the thermal treatment. For pressure-assisted sintering (such as SPS and HPS), the densification process of CNTs-CMCs powder takes place in the furnace during the sintering process. For the PLS route, the CNTs-CMCs mixed powder is uniaxially pressed to pellets before the sintering, with a manual hydraulic press or other pelleting apparatus. Next, the pellet is placed in an electrical/tubular furnace and fired to form the final product. Only during the specific temperature firing/sintering, the high energy breaks the strong covalent and ionic bonding within the ceramics and opens a path for the vacancy diffusion mechanism [128]. The gas type/flow is optional and greatly dependent on the thermal stability and purity of CNTs at different elevating temperatures. The schematic setup is displayed in Figure 13.  Table 7 show a series of reported CNTs-CMCs fabrication via PLS technique.  Table 7 show a series of reported CNTs-CMCs fabrication via PLS technique.  [137] Pressureless sintering is the least complex sintering method where the furnace setup is considerably low cost and simpler as compared to SPS and HPS. However, the inhomogeneous CNT dispersion and the long sintering duration are needed to achieve sufficient time and energy savings.

Microwave-Assisted Sintering (MAS)
Microwave-assisted sintering emerged as a new sintering technique for CNTs-CMCs fabrication. The processed CNTs-CMCs interact with the electromagnetic field (microwave), where the existence of electric or magnetic fields is highly material-dependent. Microwaves irradiate the CNTs-CMCs and generate heat within the CMC. The generated heat is responsible for the thermophysical changes and the CNTs-CMCs interfacial interactions.
Wang et al. coated SWCNTs with SiC via microwave irradiation at 675 W, on the acidified SWCNTs and the chlorotrimethylsilane (a type of chromatographic agent) [138]. After 10 min of microwave irradiation, the CNTs-silane mixture in the vessel formed a tree-like solid, and the energy dispersive X-ray spectroscopy coped transmission electron microscope (EDX-TEM) analysis showed that the SiC coated the wall of SWCNTs (see Figure 14a,b). Bhandavat et al. compared the influence of conventional pyrolysis and microwave sintering on the Si(B)CN polymer-derived ceramic with 5 wt.% MWCNTs reinforcement [139]. MWCNTs were first dispersed in toluene and homogenized with boron-modified poly(urea)methyl vinyl silazane (PUVMS, acting as Si(B)CN polymeric precursor) before the thermal treatment. By comparing the microwave sintering (900 W, 5 min) and pyrolysis (800 • C, 4 h), the oxidation temperature of the microwave-sintered CNTs-Si(B)CN ceramic (736.8 • C) was higher than that of the pyrolysis ceramic (730 • C), and the weight loss of the microwave-sintered ceramic was 17.4% lower than that of the pyrolysis ceramic. However, the 5 min microwave sintering provided insufficient time for ceramic growth, resulting in a ceramic yield of only 50%. Abbaspour and Ghaffarinejad developed CNTs-CCEs for DNA determination by the microwave-assisted sintering of methyltrimethoxysilane (MTMOS) and CNTs at a ratio of 100 µL:25 mg [140]. They performed low-power microwave irradiation (300 W, 10 min) on the MTMOS-CNTs sol-gel mixture with methanol and HCl as a medium and gelling agent, respectively. This CNTsreinforced CCE successfully oxidized adenine and guanine in DNA, and the limit of DNA detection was as low as 0.05 µg/mL. A series of microwave assisted sintered CNTs-CMCs is displayed at Table 8. CNTs-CCEs for DNA determination by the microwave-assisted sintering of methyltrimethoxysilane (MTMOS) and CNTs at a ratio of 100 μL:25 mg [140]. They performed lowpower microwave irradiation (300 W, 10 min) on the MTMOS-CNTs sol-gel mixture with methanol and HCl as a medium and gelling agent, respectively. This CNTs-reinforced CCE successfully oxidized adenine and guanine in DNA, and the limit of DNA detection was as low as 0.05 μg/mL. A series of microwave assisted sintered CNTs-CMCs is displayed at Table 8.  The heating rate of microwave sintering is not limited as the thermal heating furnace is, which makes this sintering capable of reaching high temperatures rapidly and offers a faster sintering path. Yet, microwave sintering can only be applied to composites with certain criteria: ceramics with low microwave absorbances and low additions of CNTs. This is because high-microwave-absorbance ceramics generate heat rapidly, eventually causing defects in the structure itself; CNT is an excellent microwave absorption material, easily deformed into amorphous/porous carbon or decomposed into gas, leaving a CNTfree ceramic.   [144] The heating rate of microwave sintering is not limited as the thermal heating furnace is, which makes this sintering capable of reaching high temperatures rapidly and offers a faster sintering path. Yet, microwave sintering can only be applied to composites with certain criteria: ceramics with low microwave absorbances and low additions of CNTs. This is because high-microwave-absorbance ceramics generate heat rapidly, eventually causing defects in the structure itself; CNT is an excellent microwave absorption material, easily deformed into amorphous/porous carbon or decomposed into gas, leaving a CNTfree ceramic.

Microstructural Properties
The microstructural properties of the pristine CMCs changed upon the addition of CNTs. The introduction of a foreign material (CNT) to the CMCs impacts the grain size, crystallite size, lattice microstrain, unit-cell lattice parameter, and other parameters. Grain size reduction is a common observed phenomenon upon addition of CNTs into CMCs. Through SPS (1800 • C) with colloidal CNT, the grain size growth of Al 2 O 3 retards from 75 µm (CNT-free) to approximately 0.2 µm (5 wt.% MWCNTs) [145]. The retardation of grain size is discussed to be due to the CNTs tending to form a strong entangled network surrounding the grain, limiting the grain growth even if a high temperature/energy is supplied for grain growth. Similar research was carried out by Satam et al., where they compared the CNT dispersion process into polycrystalline Al 2 O 3 : sol-gel versus powder mixing [146]. They recorded 5 vol.% sol-gelled MWCNTs with a higher-homogeneity CNT dispersion showing a greater grain size growth retardation, approximately 69.2% smaller than pristine Al 2 O 3 , while powder-mixed CNTs only restrained the grain size by about 46.2%. However, both sol-gelled and powder-mixed CNTs-Al 2 O 3 enhanced the elastic modulus (E) from 418 GPa (pristine Al 2 O 3 ) to 448 GPa and 434 GPa, respectively. Sharma and Kothiyal reported that the percentage of the crystallite portlandite (trigonal Ca(OH) 2 ) phase in Portland cement increased from 23.5% to 66.5% with the addition of 0.125 wt.% colloidal MWCNTs cured for 4 weeks [147].
To the best of the authors' knowledge, there is no direct observation of the crystallization mechanism in CNTs-CMCs. However, Wu et al. reported that CNTs induced crystallization in an olefin block copolymer (OBC) by introducing more nucleation sites to the polymer [148]. They observed that the size of crystallite spherulites in the polymer reduced, but the amount of spherulites increased rapidly with the incorporation of MWC-NTs. They calculated that the times taken for CNTs-induced spherulites to reach 10% and 50% crystallinities were 33.8 and 37.2 times faster than those for the CNT-free polymer, respectively. The dimensional growth and the nucleation sites of spherulites in polymers can also be estimated with the n and k values in classic isothermal polymer crystallization kinetics-the Avrami equation [149] in Equation (2): where k is the crystallization rate constant and n is the crystal growth geometry-dependent Avrami exponent. The k values of 0.1 wt.% CNTs-OBC and pristine OBC were recorded as 0.0874 and 2.2 × 10 −6 , respectively, which indicates that the MWCNTs introduced more nucleation sites and reduced the induction period for OBC crystals. Pristine OBC was categorized as a three-dimensional crystallization growth with an n value close to 3 (n = 2.84), while CNTs-OBCs were more likely to exhibit two-dimensional crystal growth as the n value was 2.41. In addition, the crystallization activation energy of 0.1 wt.% CNTs-OBC was recorded at 422.84 kJ/mol, which is lower than that of pristine OBC, which was recorded at 522.5 kJ/mol. The incorporation of CNTs helped in promoting more nucleation sites, switching the crystal growth to a lower dimension and reducing the activation energy, resulting in a rapid crystallization.

Thermal Properties
CNTs have a positive influence on the thermal properties of CMCs, namely, thermal conductivity, diffusivity, resistivity, capacity, and the coefficient of thermal expansion. CMCs consist of strong bonds and light atoms e.g., O-Si-O bonds. At T > 0 K, the atoms in CMCs vibrate in a collective mode: phonon vibration. Phonon vibrations can be transmitted across or stored in the solid, and these are greatly affected by vibrations of adjacent atoms through bonding and the surrounding temperature, triggering elastic deformation. At working temperatures lower than the Debye temperature (T < T D ), the Einstein model approximates that the heat transport in a material is the result of high mobility, and the temperature dependency is T [150]. However, this is nearly negligible for ceramic materials with localized charge. Debye introduced the approximation of heat transport in the form of phonons, and the temperature dependency is T 3 [151], which is applicable in metals, ceramic, glass, etc. [152].
Phonon vibrations are a quantum theory-derived concept that describes the energies and directions of motion in collective modes propagating through the material. The concept of the phonon is analogous to light particles-photons that involve energy and frequency. The photon energy depends on the frequency of the electromagnetic wave (ν), while the phonon energy (ε) depends on the velocity of sound (v s ) and wavevector-derived angular frequency (ω).
Phonon-phonon interactions, crystal imperfections, and grain boundaries are the main factors that influence the thermal conduction in CMCs. Thermal energy travels through the ceramic as phonons (elastic waves) at the speed of sound until they are fully scattered. The phonon scattered as a form of wave through phonon-phonon interaction. When the in-phase phonon waves overlap at a particular atom, the vibrational amplitude is superimposed and produces a different phonon with a different spring constant. In addition, phonons scatter at lattice imperfections, especially for materials lacking a periodically arranged lattice. The thermal conductivity (κ th ) and diffusivity (D) of CMCs decrease as the crystalline ceramic approaches its melting point. At the melting point, the ordered lattice arrangement in the ceramic starts to distort and form the semi-crystalline and amorphous phase, causing the number of scattering sites to increase. The increased scattering sites give rise to thermal resistivity ( ) and degrade the κ th and D of the ceramic. By contrast, the dominant mechanisms of the κ th and D are highly dependent on thermal radiation in the form of light at high temperature. At this temperature, ceramic conducts heat through absorbing and emitting photons in a rapid sequence. This thermal conduction mode increases κ th and D values with the increase in working temperature, and this is an important parameter of thermal management in transparent-based glass materials. In addition, the grain size and grain boundaries determine the thermal conduction in ceramics. During the synthesis of ceramics, controllable high temperatures are involved in promoting the grain growth and reducing the grain boundaries. The closer the grains, the faster the phonon vibration or photon transmission between atoms, and this improves the thermal conduction of CMCs.
The thermal capacity (C p ) of ceramics is based on the bond strength, mass of the atom, and bond length. Light atoms in CMCs have a short bond length that contribute as a function of the distance of the potential well, analogous to Hooke's Law. Two spheres are attached to a spring at an equilibrium position, which represents the atom at the bond length (see Figure 15). When the bond is compressed or stretched from its equilibrium position, the short bond length, strong bonding constant, and low mass cause the CMCs to have a high vibration frequency with a small disturbance in the lattice, resulting in a high C p .
CNTs have introduced their unique thermal transport properties into CMCs by reducing the crystal defects. During the CNTs-CMCs fabrication, CNTs possess a high κ th that channels thermal energy from the heat source into the CMCs crystal, introduce more nucleation sites, reduce the induction period, and enhance the crystallization. The crystals in CMCs appear as single or poly-crystalline instead of amorphous phases. However, CNTs form a network around the ceramic grain, retard the grain growth, and increase the grain boundaries. This may result in a low thermal transport of the CMCs. CNTs fill the grain boundary through tube-tube aggregation and conduct phonon vibration together with electron transmission as CNTs are electron-rich materials. Table 9 shows previous research on applying CNTs to CMCs in achieving better thermal management. CNTs have introduced their unique thermal transport properties into CMCs by reducing the crystal defects. During the CNTs-CMCs fabrication, CNTs possess a high κth that channels thermal energy from the heat source into the CMCs crystal, introduce more nucleation sites, reduce the induction period, and enhance the crystallization. The crystals in CMCs appear as single or poly-crystalline instead of amorphous phases. However, CNTs form a network around the ceramic grain, retard the grain growth, and increase the grain boundaries. This may result in a low thermal transport of the CMCs. CNTs fill the grain boundary through tube-tube aggregation and conduct phonon vibration together with electron transmission as CNTs are electron-rich materials. Table 9 shows previous research on applying CNTs to CMCs in achieving better thermal management. Chemical vapour infiltration PLS CNTs-SiC network coated alumina pipe Thermal insulation: 60−12.5 °C/min at 750 °C flame [157] * VACNTs: vertically aligned CNTs; YSZ: yttrium-stabilized zirconia.  Chemical vapour infiltration PLS CNTs-SiC network coated alumina pipe Thermal insulation: 60−12.5 • C/min at 750 • C flame [157] * VACNTs: vertically aligned CNTs; YSZ: yttrium-stabilized zirconia.

Optical Properties
The optical properties are the behaviour of materials when they are exposed to electromagnetic radiation (light). Glass-based materials are optically transparent, have a high light transmittance, a high reflectance (smooth surface), and a high optical density-dependent refraction. By contrast, glass-ceramics or ceramics-based materials mostly appear as opaque, and the light transmitted through them is nearly zero. When the material interacts with light, the predominant mechanism that occurs in the materials is electronic polarization and electron transition between different energy states.
Electronic polarization takes place during the electric field of light distorting the electron cloud in the material. The electronic polarization mechanism can be divided into two conversions: photon-phonon and photon-photon. In photon-phonon conversion, light energy is scattered and converted into elastic waves during the propagation through the materials. The produced elastic waves appear to be phonon vibration and, subsequently, generate heat. In photon-photon conversion, it involves a rapid light absorption and reemission at the same wavelength, which is known as reflection. As the energy levels of electrons are quantized and the ground state overlaps with the excited state, light provides the electrons with sufficient energy to shift from lower energy levels (ground state) to a higher energy levels (excited state). At the higher energy levels, electrons are unable to hold the energy for prolonged periods and instantly fall back to the ground state, emitting electromagnetic radiation at the same wavelength as the light energy supplied to the electrons. Figure 16 shows the possible phenomenon when ceramics interacts with light.
The optical properties are the behaviour of materials when they are exposed to electromagnetic radiation (light). Glass-based materials are optically transparent, have a high light transmittance, a high reflectance (smooth surface), and a high optical density-dependent refraction. By contrast, glass-ceramics or ceramics-based materials mostly appear as opaque, and the light transmitted through them is nearly zero. When the material interacts with light, the predominant mechanism that occurs in the materials is electronic polarization and electron transition between different energy states.
Electronic polarization takes place during the electric field of light distorting the electron cloud in the material. The electronic polarization mechanism can be divided into two conversions: photon-phonon and photon-photon. In photon-phonon conversion, light energy is scattered and converted into elastic waves during the propagation through the materials. The produced elastic waves appear to be phonon vibration and, subsequently, generate heat. In photon-photon conversion, it involves a rapid light absorption and reemission at the same wavelength, which is known as reflection. As the energy levels of electrons are quantized and the ground state overlaps with the excited state, light provides the electrons with sufficient energy to shift from lower energy levels (ground state) to a higher energy levels (excited state). At the higher energy levels, electrons are unable to hold the energy for prolonged periods and instantly fall back to the ground state, emitting electromagnetic radiation at the same wavelength as the light energy supplied to the electrons. Figure 16 shows the possible phenomenon when ceramics interacts with light. Figure 16. Light phenomenon when the material is exposed to electromagnetic radiation (light). Transparent samples (glass) refract the incident ray; almost all the material reflects light but at different angles of reflection, i.e., a smooth surface reflects the light at an identical angle; light transmits through translucent and transparent materials (glass-ceramics or ceramics-based materials); high-light-absorption materials absorb light and may scatter as a form of heat (CNTs-added ceramics).
However, electronic polarization is not applicable to most ceramics, as there is a separation (capped at 4 eV) between the valence band and conduction band-the bandgap energy (Eg). Due to the remarkable Eg, the electronic transition in the ceramic is meaningful when it comes to application purposes. Eg is classified into direct and indirect, and the electronic transition is divided into allowed and forbidden transitions. The direct bandgap is a vertical transition (electron-photon interaction) at the same wavevector where the energy changes but the momentum is conserved, e.g., X-X transition; the indirect bandgap is an oblique transition (electron-photon-phonon interaction) at different wavevectors where the energy and the momentum change, e.g., X-Γ transition. The difference between Figure 16. Light phenomenon when the material is exposed to electromagnetic radiation (light). Transparent samples (glass) refract the incident ray; almost all the material reflects light but at different angles of reflection, i.e., a smooth surface reflects the light at an identical angle; light transmits through translucent and transparent materials (glass-ceramics or ceramics-based materials); high-light-absorption materials absorb light and may scatter as a form of heat (CNTs-added ceramics).
However, electronic polarization is not applicable to most ceramics, as there is a separation (capped at 4 eV) between the valence band and conduction band-the bandgap energy (E g ). Due to the remarkable E g , the electronic transition in the ceramic is meaningful when it comes to application purposes. E g is classified into direct and indirect, and the electronic transition is divided into allowed and forbidden transitions. The direct bandgap is a vertical transition (electron-photon interaction) at the same wavevector where the energy changes but the momentum is conserved, e.g., X-X transition; the indirect bandgap is an oblique transition (electron-photon-phonon interaction) at different wavevectors where the energy and the momentum change, e.g., X-Γ transition. The difference between the allowed and forbidden transitions is the momentum matrix element characterization: different from zero (allowed transition) and equals zero (forbidden transition). The sketched band structure of a material with an indirect bandgap is displayed in Figure 17. It shows that the direct E g is larger than the indirect E g , meaning that the material is capable of performing a direct band transition when the energy supplied is sufficient. Researchers have extensively applied CNTs to optical absorbance properties such as bandgap reduction to domain crystals. Table 10 shows previous studies in the literature of the application of CNTs in achieving ceramics with low bandgap applications. different from zero (allowed transition) and equals zero (forbidden transition). The sketched band structure of a material with an indirect bandgap is displayed in Figure 17. It shows that the direct Eg is larger than the indirect Eg, meaning that the material is capable of performing a direct band transition when the energy supplied is sufficient. Researchers have extensively applied CNTs to optical absorbance properties such as bandgap reduction to domain crystals. Table 10 shows previous studies in the literature of the application of CNTs in achieving ceramics with low bandgap applications.

Mechanical Properties
Introducing CNTs into CMCs to make a stronger ceramic composite is not a modern advancement but it has been developed decades ago. Previous studies in the literature have shown the effect of CNTs on the enhancement of mechanical properties when CNTs are dispersed and densified with appropriate techniques with optimized CNTs

Mechanical Properties
Introducing CNTs into CMCs to make a stronger ceramic composite is not a modern advancement but it has been developed decades ago. Previous studies in the literature have shown the effect of CNTs on the enhancement of mechanical properties when CNTs are dispersed and densified with appropriate techniques with optimized CNTs compositions. Generally, the CNTs-CMCs mechanical properties enhancement refer to the increase in fracture toughness (K IC ), flexural strength, and Young's modulus, upon low CNTs addition. The K IC of the CNTs-CMCs was mainly estimated by measuring the crack lengths from the Vickers indentation (also known as the indentation fracture indention technique) [120,136,162,163] and single-edge notched beam (SENB) method [51,[164][165][166]. For CNTs-CMCs materials, the Evans and Charles equation [167] is widely used to estimate the K IC from the crack length, which is shown in Equation (3): where P is the applied load and c is the radial crack length. The lack of CNT pullout from the CMCs and CNTs bridging caused the Vicker's surface hardness (H V ) and the % of the CNTs-CMCs to not show significant improvements, and they increased the risk of mechanical defects such as a low K IC and low bending strength [166] (see Figure 18). Table 11 displays the recent works (dispersion, densification, sintering, and optimized CNTs composition) of the CNTs-CMCs compared with the CNTfree CMCs to illustrate the improvement from the mechanical perspectives.
compositions. Generally, the CNTs-CMCs mechanical properties enhancement refer to the increase in fracture toughness (KIC), flexural strength, and Young's modulus, upon low CNTs addition. The KIC of the CNTs-CMCs was mainly estimated by measuring the crack lengths from the Vickers indentation (also known as the indentation fracture indention technique) [120,136,162,163] and single-edge notched beam (SENB) method [51,[164][165][166]. For CNTs-CMCs materials, the Evans and Charles equation [167] is widely used to estimate the KIC from the crack length, which is shown in Equation (3): where P is the applied load and c is the radial crack length. The lack of CNT pullout from the CMCs and CNTs bridging caused the Vicker's surface hardness (HV) and the ρ% of the CNTs-CMCs to not show significant improvements, and they increased the risk of mechanical defects such as a low KIC and low bending strength [166] (see Figure 18). Table 11 displays the recent works (dispersion, densification, sintering, and optimized CNTs composition) of the CNTs-CMCs compared with the CNT-free CMCs to illustrate the improvement from the mechanical perspectives.

Summary and Outlook
The present review article aimed to provide an overview of the critical challenges, the recent breakthroughs on the dispersion and densification and sintering routes, and the enhanced characteristics of the CNTS-reinforced ceramic matrix composites. CNTs have strengthened the ceramic matrix composites in many aspects such as density, fracture toughness, thermal conductivity, and wear resistance, while only a low amount of CNTs reinforcement is needed. Those beneficial properties could be improved or perfected if the CNT fillers are distributed homogenously within the domain matrices, and the agglomeration issue has thrived. With such fascinating properties, CNTs should have improved the CMCs' functional properties linearly or exponentially with the increase in CNT fillers. However, most research on the CNTs-CMCs has resulted in an "optimum" addition of CNTs for the best CMCs' performance due to the CNTs agglomeration at a large CNTs amount in the matrix.
To counter these issues, researchers have investigated the processing techniques and densification and sintering routes of CNTs-CMCs, such as the sol-gel processing and SPS, to maximize the functional properties. Every mentioned processing technique (Section 3) and densification and sintering route (Section 4) has its pros and cons in the perspective of the techniques' complexity, duration, energy consumption, and the starting materials (see Tables 12 and 13). All these pros-cons considerations are the parameters that determine the CNTs distribution, interfacial adhesion between CNTs and the domain ceramic, and the thermal stability and the sinterability of CNTs-CMCs. The desired properties of the CNTs-CMCs end-products are also the main concerns during the fabrication process. Still, many studies maintain the momentum to optimize the existing CNTs-CMCs fabrication processes, discover novel fabrication methods, and evaluate the final composites' performance from the microstructural properties to the functional properties before commercializing them. We addressed the fundamental insights for the future technological maturation and advancement of CNTs-CMCs.