Effect of carbon fabric type on properties of 2D carbon–carbon composites processed by liquid pitch impregnation

Abstract

The present study reports on the microstructural, physical, thermal and mechanical properties of two types (A and B) of carbon–carbon composites processed by an economical route. Skeleton composites were first made by pyrolysing laminated carbon fibre-reinforced phenolic composites and subsequently densified by liquid pitch impregnation–pyrolysis process. Both types of composite employed polyacrylonitrile-based 8-harness satin woven carbon fabrics, Type A being woven with tows of continuous fibres, whereas Type B used yarns of discontinuous fibres. Experimental results indicated that the Type A composite had better flexural and tensile modulus and strength values. However, the Type B composite had better interlaminar shear strength and through-thickness thermal conductivity, despite being less dense. These results are discussed and explained from microstructural and fractographic analysis using optical and scanning electron microscopes. Finally, the results were compared with those intended for similar applications.

Keywords

Carbon–carbon composites carbon fabric liquid pitch impregnation-pyrolysis microstructures thermal properties mechanical properties fractography

Introduction

Carbon–carbon (C_f/C_m) composites have a wide spectrum of properties and applications. These composites have been synthesized since the 1970s with the aim to combine the typical advantages of fibre-reinforced composites such as high specific strength, stiffness and in-plane fracture toughness with the refractory properties of structural graphite.¹ C_f/C_m composites are classified as intermediary between conventional synthetic graphite materials (i.e. the polygranular electrode materials or pyrolytic graphite) and carbon fibre-reinforced polymers. C_f/C_m composites have many unique properties, including high strength and high stiffness at ultrahigh temperatures (as high as 3000℃) in nonoxidising atmospheres; low density (1.6–2.00 g/cm³); high thermal conductivity, low thermal expansion coefficient and good thermal shock resistance; excellent frictional and wear characteristics; desirable neutronic properties; excellent chemical inertness, and suitable biocompatibility. The ultra-high temperature capabilities have made C_f/C_m composites the materials of choice for many important sectors, including the aerospace industry (rocket-propulsion systems, heat shields of re-entry space vehicles, aircraft brakes), high temperature nuclear and fusion reactors, road and rail transport braking systems.^1–7

Manufacturing of C_f/C_m composites generally involves three basic steps:

Step 1: Carbon Fibre Preforms Carbon fibres are obtained by carbonisation and graphitisation of an organic precursor made from Rayon, polyacrylonitrile (PAN) or pitch. Depending upon the end applications, the carbon fibres are arranged in preforms in 1, 2, 3,…, n directions, and the preforms are accordingly called UD (unidirectional), 2D, 3D,…., nD).¹

Step 2: Carbon Matrix Development The carbon matrix is obtained by either (a) liquid phase (resin or pitch) impregnation of the preforms followed by a pyrolysis/carbonisation treatment or (b) gas phase impregnation, hereby hydrocarbon gas is cracked at high temperature before chemical vapour infiltration (CVI) or (c) a hybrid process combining (a) and (b).^1,8–10 The impregnation-pyrolysis cycle is repeated until the product has the desired density. Although the CVI process takes a relatively long time and is costlier, its economic disadvantage is offset by much better properties in the final product.¹

Step 3: Graphitisation Treatment The impregnated preform is heated above the carbonisation temperature (∼1000–1300℃) in order to enhance high temperature capabilities of the final composites. However, the graphitisation treatment, if done as an intermediate process between impregnation-pyrolysis cycles, opens up porosity and aids further impregnation.¹

The properties of a particular C_f/C_m composite are largely determined by the fibre reinforcement(s), the matrix, and the fibre/matrix interfaces. More specifically, the type and volume of carbon fibres, preform arrangements, matrix carbon precursors, densification process and final heat-treatment temperature greatly influence the density, macrostructure (type, size and quantities of defects, i.e. pores, cracks, etc.) and matrix microstructure (orientation of graphitic planes and fibre/matrix interfaces): in turn, these latter features control the material properties.^1–4,7–18

In the present work, objective was to make two C_f/C_m composites, Type A and Type B through an economical route without employing costlier CVI process. These two composites had 2D fibre arrangements but different types and volumes of carbon fibre reinforcements. Such composites have potential application for aircraft braking systems. Their physical, thermal and mechanical properties were determined, analysed and correlated with microstructural features, and finally the results were compared with those of carbon brake materials reported in the open literature.

Experimental details

Materials and processing

Two types of C_f/C_m composite materials (a) Type A: CCM2ND and (b) Type B: CCM3ND were produced by the liquid pitch impregnation-pyrolysis (LPIP) route.

Fiber reinforcements for type A composite:

Major: PAN-based 8-harness (H) satin woven fabric consisting of tows of continuous carbon fibres of T-300 grade (∼ 42% by volume). Its other salient features are fabric thickness ∼0.45 mm, tensile breaking load per 25 mm width of the fabric both in warp and weft direction ∼ 180 kg and carbon content ∼ 92%.

Minor: 2.5% (by weight) milled high modulus (HM) carbon fibres (length ∼ 0.1–2 mm) mixed with phenolic resin. It was used to reinforce the bulk matrix present in the cross over points of warp and weft threads of the fabric (schematic of the weaving style of 8 H-satin woven fabric shown in Figure 1(a)) and to improve overall thermal conductivity and friction and wear properties of the composite.

Fiber reinforcement for Type B composite:

Woven PAN-based 8 H-satin fabric, but it, unlike the fabric of Type A composite, consists of yarns of discontinuous carbon fibres. Its other salient features are: Fabric thickness ∼0.77 mm, tensile breaking load per 25 mm width of the fabric both in warp and weft direction ∼20 kg and carbon content ∼99%.

No additional reinforcement was used for phenolic resin.

The actual manufacturing process details for these two types of fibres were proprietary. However from their carbon contents (which were determined based on oxidative combustion using CHNS/O Elemental analyser (Model: vario EL III), it could be assumed that the Type A fibre reinforcement was carbonised (∼1200–1300℃) while the Type B fibre had undergone high temperature graphitisation (in excess of 1800℃) during manufacturing.^1,19 A graphitised fabric is expected to have better thermal conductivity and stability, owing to development of a high degree of fibre crystallinity during graphitisation.¹

In order to make the composites the fabrics were prepregged using resole-type phenolic resin polymer. They were cut and laid-up into annular discs of thickness ∼20 mm, inner diameter ∼140 mm and outer diameter ∼290 mm. During laying-up the orientation of each prepreg fabric layer was altered by a predefined angle with respect to the warp and weft directions of the fabrics to achieve better in-plane isotropy in the discs. The discs were then hot-moulded in a press, followed by carbonisation to obtain porous ‘skeleton’ preforms. Because of the difference in fabric thickness (indicated in preceding paragraph), total fibre contents, by volume, of both the composites could not be maintained at the same level and it was 45% for Type A and 30% for Type B composites respectively. The skeleton preforms were subsequently densified by multiple cycles of LPIP to get the final product. Each densification cycle consisted of pitch impregnation, carbonisation under pressure (80–90 MPa), and a graphitization heat treatment at 2600℃.

The final C_f/C_m composite discs were checked for internal defects by through-transmission ultrasonic inspection using dry coupling. Test specimens for property determinations were taken from defect-free discs.

Preparation of test specimens and testing

Test specimens were extracted randomly from different location of the discs, with two to five specimens being tested for each property. Unless specifically mentioned, the test specimen surfaces were finally prepared by surface grinding (R_a ∼ 1.6 µm) and profiles were cut using a computer numerical control (CNC) machine with carbide tools.

Average bulk density (ρ), i.e. weight by volume, was calculated by measuring the weights of test specimens of uniform dimensions. The open porosity values were determined using a QUANTACHROME Mercury Porosimeter and by intruding mercury at ≤30,000 psi into test specimens in the dilatometer. The test specimens were square cross-section bars 5 mm × 5 mm × 25 mm. The porosity values were calculated assuming theoretical densities of the composites to be 2.26 g/cm³.

Thermal diffusivity (α) parallel and perpendicular to the fabric layers was measured using the laser flash method⁹ as per ASTM Standard E1461 using NETZSCH LFA-427 laser thermal conductivity equipment. The thermal conductivity (k) was calculated using the equation

k = α \times C_{p} \times ρ

(1)

The specific heat (C_P) was measured as per ASTM Standard E1269 using a DSC 2920 differential scanning calorimeter.

Flexural behaviour was studied by three-point bend testing at room temperature, with the specimens maintaining a span (L) to thickness (d) ratio of 18:1. The testing was done using an INSTRON 5500R with 5 kN load cell and a crosshead speed (CHS) of 0.5 mm/min. The failure mode was observed and the flexural strength (σ_f) and modulus (E_f) were calculated based on the equations given in ASTM Standard C 1341:

σ_{f} = \frac{3}{2} [\frac{P_{max} \times L}{{bd}^{2}}]

(2)

E_{f} = \frac{1}{4} [\frac{L^{3}}{{bd}^{3}} \times \frac{Δ P}{Δ m}]

(3)

where P_max is the maximum load attainable before failure; L is the support span length; b is the width and d is the thickness of the specimen; ΔP is the change in force in the linear portion of the load-deflection curve and Δm is the change in deflection corresponding to ΔP.

Tensile testing was done at room temperature using specimens with reduced and square cross-sections (10 mm × 10 mm) over a length of 30 mm using an INSTRON 5500R-4507-200kN machine with a constant CHS of 0.5 mm/min.

Interlaminar shear strength (ILSS) was measured by short-beam three-point bend testing at room temperature, with the specimens maintaining a span (L) to thickness (d) ratio of 5:1. As before, the testing was done with an INSTRON 5500 R with 5 kN load cell and a constant CHS of 0.5 mm/min. The failure mode was observed and the ILSS was calculated based on the equations given in ASTM Standard D-2344:

ILSS = \frac{3}{4} [\frac{P_{max}}{bd}]

(4)

where P_max is the maximum load attainable before failure, b is the width and d is the thickness of the specimen.

Features of both the types of woven PAN-based carbon fabrics were investigated using an FEI-Quanta 400 environmental scanning electron microscope (ESEM). In order to characterise microstructural features of both Type A and Type B composites, samples parallel and perpendicular to the fabric layers were ground and polished down to 1 µm diamond paste, followed by ultrasonic cleaning and drying. They were then examined under optical microscope using polarised light and also under ESEM. In addition, the fracture surfaces of the failed flexural- and tensile-tested specimens were examined by ESEM microscopy.

In order to find the extent of graphitisation in both the composites X-ray diffraction (XRD) was done on powdered samples to measure the d₀₀₂ interlayer spacing of the basal graphitic planes (using Bruker D8 Advanced X-ray equipment and CuK_α radiation ∼1.5406 Å). The d₀₀₂ interlayer spacing was calculated from the well-known Bragg equation:

λ = 2 d_{002} \sin θ

(5)

and the degree of graphitisation was calculated from the Franklin equation:¹

Graphitisation (%) = [\frac{3.44 - d_{002}}{0.086}] \times 100

(6)

Results and discussions for the two types of composite

Carbon fabric features

Figure 1(b) presents SEM micrograph of the carbon fabric used for the Type A C_f/C_m indicating that the fabric is made of tows woven in 8-H satin weave style (schematically represented in Figure 1(a)). Each tow (Figure 1(c)) consisted of several thousand continuous fibre filaments with diameters 7–7.5 µm (Figure 1(d)). Hence, Type A composite will have fibres aligned mostly in xy-plane and parallel with in a tow. Figure 2 presents SEM micrographs of the milled carbon fibres which were used in small fraction along with phenolic resins to serve the purpose described earlier. Its length and diameter were ∼0.05–0.2 mm and ∼6 µm, respectively (Figure 2(a)–(c)).

Figure 3 presents SEM micrographs of the carbon fabric used for the Type B C_f/C_m. Although also woven in 8-H satin style, the fabric is made of yarns, which are assemblies of monofilaments held together by a twist, Figure 3(b). The yarns consisted of discontinuous filaments/fibres, as seen in Figure 3(c); and the filament diameters are 8–8.5 µm, Figure 3(d). Ends of monofilaments are observed to be projecting out of the yarns (Figure 3(c)). The Type B fabric has fibres that are much less orderly within a yarn due to the twist. Thus many fibres in the Type B fabric under the influence of weaving process will have a significant z-direction component, in contrast to the Type A fabric where the fibres are mostly in xy-plane. As the Type B fabrics are also much coarser and has low fibre volume (see the subsection ‘Materials and processing’), when stacked interlocking of yarns (and layers) can take place by the fibre ends (Figure 3(c)) which are protruded out from one yarn (and layer) to the next one. The open ends of each monofilament in the yarns are also expected to provide additional contact area (equivalent to πd²/2, where d is diameter of the fibre) and are believed to act as anchors in the composite matrix. All these features of the Type B fabric such as z-direction component, interlocking of yarns (and fabric layers), anchoring effect of ends of monofialments are expected to influence through-thickness thermomechanical properties (specifically ILSS and through-thickness thermal conductivity) of the laminated composites.

Microstructures

The microstructures of both C_f/C_m composites are illustrated in Figures 4 and 5. These figures show the carbon fibres; the weaving pattern of the carbon fabric; the overall distribution of carbon matrices (derived from both the phenolic resin which was used for making the CFRP composites and the pitch used for densification of the porous skeleton preforms. The micrographs also show pores of various sizes and matrix cracks.

Optical metallography under polarised light

In polarised light images, Figures 4(a)–(e) and 5(a)–(e), the light regions represent fibres and matrix and the dark regions are pores. The pores were within and between the fibre tows (ref. for Type A)/fibre yarns (ref. for Type B) and within the fabric layers, and varied greatly in size. However, there were differences in pore morphology for the two types of composites. In Type A the lengths and widths of elongated pores (which were predominantly within fibre tows) were greater, cf. Figure 4(a) and (b). Type B contained many large rounded pores, distributed more or less randomly, see Figure 5(a) and (b). These differences in pore morphology and distribution are likely due to using tows for the Type A carbon fabric and yarns for the Type B carbon fabric. Other notable features differing for the two composites are

The interface gap between fabric layers is more discernible for Type A, cf. Figure 4(d) and (e).

Microstructures of Type A (Figure 4(b) and (c)) were nonuniform in terms of short fibres embedded in matrix at the cross-over regions between the warp and weft fibre threads. As stated earlier in section 2.1, milled high modulus fibres was used to reinforce phenolic resin which is likely to get accumulated at the fibre cross-over points of the carbon fabric. As the mixture of resin and milled fibre is applied manually, the milled fibre is likely to get settled down in the resin mixture over the time, resulting in having matrix rich areas at cross-over regions. In contrast, the overall microstructures of Type B composite were observed to be more or less uniform with occasional presence of matrix rich areas. The matrix rich regions, both in Type A and Type B composites, were found to be associated with micro-cracks (Figures 4(c) and 5(d)) which had generated during heat treatment possibly due to absence of reinforcing fibres.

The fibre/matrix bonding in Type A composite was observed to be incomplete, with unfilled regions along/around the fibres, see Figure 4(d) and (f). On the other hand, for Type B composite, the fibre/matrix bonding was found to be comparatively very good, see Figure 5(d) and (f). These observations were further confirmed through the SEM microscopy as illustrated in the following section.

Very good fibre/matrix bonding observed in Type B composite was attributed to the fact that the composite had pre-graphitised fibres (as mentioned in the subsection Materials and processing) which remained stable during composite heat treatment, i.e. the fibres did not shrink, rather matrix could shrink and maintained tight bond to fibres. This was not the case for the Type A composite, where the fibres shrink along with the matrix during heat treatment of composite resulting in lower fibre/matrix interface bond compared to that in Type B composite.

The probable and actual influence of fabric type is remarkable, not just with respect to porosity, inter-fabric layer interface, matrix micro-cracking and fibre/matrix interface, but also for the physical and mechanical properties, as will be shown in sections 3.4 and 3.5 and further discussed in the section Comparison of the Results with C_f/C_m Literature Data.

SEM metallography

The SEM micrographs in Figures 6 and 7, obtained from secondary electron (SE) and back scattered electron (BSE) imaging, show individual fibres, the matrices continuity around the fibres indicating the health of fibre/matrix interface bond and unfilled regions. The image contrast effects for the fibres, the matrix surrounding the fibres and unfilled regions (i.e. pores) are diagnostic, as follows. The fibres appear grey, unfilled regions are dark whereas the matrix around fibres is lighter, especially from SE imaging.

Degree of graphitisation (DOG)

Figure 8 shows the Bragg-angle diffraction peaks for both composites, and Table 1 gives the interlayer spacing {d₀₀₂} calculated using equation (5) and the DOG calculated using equation (6). Table 1 shows that the composites are graphitised to nearly the same level. This is interesting because the original carbon contents were different: 92% for the Type A fibres and 99% for the Type B fibres. Clearly, the overall production process, which was the same for both composites, has compensated for the differing carbon contents of the fibres, resulting in the similar DOG.

Table 1.

Interlayer spacing of graphitic planes {d₀₀₂} and degree of graphitisation (DOG) for the Type A and Type B composites.

Composites	Interlayer spacing {d₀₀₂}	DOG
Type A: CCM2ND	3.368	83.87
Type B: CCM3ND	3.369	82.42
Ideal graphite1	3.354	100

Physical and thermal properties of the composites

The density, porosity and thermal conductivity values of the composites are given in Table 2. The thermal conductivity was calculated using equation (1) and reported for 200℃ and both in-plane and through-thickness directions. Even though the matrix precursor materials and processing conditions were same, the Type A composite is denser by ∼7% than Type B, corresponding to about 36% less porosity. These differences can be attributed to the higher fibre volume in the Type A composite.

Table 2.

Physical and thermal properties of the Type A and Type B composites.

Material	Fibre vol. (%)	Physical properties			Thermal conductivity at 200℃ (W/m·K)
Material	Fibre vol. (%)	Bulk density (g/cm³)	Open porosity (%)	Total porosity (%)	In-plane direction	Thickness direction
Type A: CCM2ND	45	1.85	8.00	22.50	110.4	13.6
Type B: CCM3ND	30	1.73	12.96	30.67	103.2	27.2

The thermal conductivity vales in-plane direction for both the composite were found to be comparable and they are one order magnitude higher than that of through-thickness direction. This is due to the fact that majority of heat gets conducted along the length of the fibres present in the in-plane direction. However, the Type B composite, despite its higher porosity, lesser density and also not having fibres in thickness direction, exhibited comparatively much higher thermal conductivity in the through-thickness direction. The increase in thermal conductivity in through-thickness direction in the Type B composite is attributed to the use of yarns rather than tows. As mentioned in subsection 3.1, due to the twist many fibres in the yarns have significant z-direction component providing the way for heat conduction perpendicular to the fabric layers, i.e. in thickness direction of the composite. Additionally, connectivity for the heat conduction across the fabric layers increases due to better inter-fabric layer and fibre/matrix bonding (as inferred from comparisons of Figures 4(e) and (f), 5(e) and (f), and 6 and 7) resulting from the interlocking of layers by the protruded fibre ends. Better bonding reduces the resistance to heat flow at each interface Further, Type B composite has less fibre volume fraction, i.e. less number of fabric layers (as indicated in the subsection Materials and processing) which has also contributed to enhanced through-thickness conductivity due to fewer planes of resistance to heat flow.

Mechanical properties of the composites

The flexural and tensile strengths and moduli, and the ILSSs of the composites are listed in Table 3. Typical flexural stress–displacement and tensile stress–strain plots are shown in Figures 9 and 10, respectively.

Table 3.

Mechanical properties of the Type A and Type B composites.

Composites	Mechanical properties
	Strength (MPa)		Modulus (GPa)		ILSS (MPa)
	Flexural	Tensile	Flexural	Tensile	ILSS (MPa)
Type A: CCM2ND	165–194 (avg. 178)	122–128 (avg. 125)	33–42 (avg. 40)	60–65 (avg. 62.5)	8–9 (avg. 9)
Type B: CCM3ND	105–110 (avg. 108)	78–82 (avg. 80)	20–26 (avg. 23.5)	30.5–36 (avg. 33)	11–13 (avg. 12)

From Table 3 and Figures 9 and 10 it is evident that the Type A composite had much better flexural and tensile properties, with advantages of 65% and 70% (flexural strength and modulus) and 56% and 89% (tensile strength and modulus) compared to the Type B composite. However, the Type B composite had a 25% better ILSS. This is attributed to the fact discussed earlier in the subsection Carbon fabric features, i.e. use of yarns in the fabric, where many fibres have a significant z-direction component besides interlocking of yarns (and layers) by protruded fibre ends from one yarn (and layer) to the next one meaning that that there is less likelihood of inter laminar/ planar separation of fabric layers.

The differences in strengths and moduli between the two types of composites may be explained from three main factors:

Density and fibre volume fraction: These are higher in the Type A composite, see Table 2. Hence this composite had higher strengths and moduli.

Types of carbon fabric: The Type A composite used woven carbon fabric with tows of straight and continuous fibres. However, the Type B composite used yarns made of discontinuous fibres held together by a twist, see Figure 2. The discontinuous fibres and the twist have reduced the mechanical properties in comparison to the more efficient fibre type and arrangement in the Type A composite. Further, as discussed in the subsection Optical metallography under polarised light, use of pre-graphitised carbon fabric resulted in strong fibre/matrix bonding in Type B composite

Fibre/matrix bonding: It is generally accepted that a strong bond between fibres and matrix is essential to achieve optimum strength and stiffness in composite systems where the fibre is brittle/ductile and the matrix is ductile. C_f/C_m composites differ significantly from this generalised assumption since the fibres and matrix are both brittle. In this case strong fibre/matrix bonding, and hence strong interfaces, result in poor tensile strength, but weak bonding and weak interfaces give poor shear strength. An optimum is obtained from intermediate bond and interface strengths that result in a higher strength mixed-mode (tensile + shear) fracture, whereby matrix cracks are deflected by de-bonding at the fibre/matrix interfaces.^1,11

In addition to points (1) and (2) this may help to explain the higher strengths of the Type A composite, and also – as will be discussed in the section Fractography and analysis of fracture surface features – the significant extent of mixed mode fracture compared to the Type B composite.

Fractography and analysis of fracture surface features

Examples of the flexural and tensile fracture features are shown in Figures. 11 and 12, respectively. These are described and discussed in the following subsections.

Type A composite

Figure 11(a) and (b) shows that flexural specimens tended to fail firstly by shear owing to separation of fabric layers, and later by tensile fracture at and near the tensile-bending surface. Similarly, Figure 12(a) and (b) shows that tensile specimens failed by predominantly mixed-mode fracture, i.e. shear owing to separation of fabric layers before tensile failure. This similar fracture behaviour is attributable to weak bonding between the fabric layers as well as the fibre/matrix interfaces, as discussed in section 3.5. Figures 11(c) and 12(c) provide additional evidence of weak bonding, since they show large amounts of fibre pull-out at many levels, indicating the existence of weak fibre/matrix bonding and a high degree of crack deflection.

These fractographic observations explain (a) the lesser ILSS (Table 3) of the Type A composite, owing to weak bonding between the fabric layers and (b) the high strength, owing to predominantly mixed-mode fracture.

Type B composite

Figures 11(d) and (e) and 12(d)–(f) show that the flexural and tensile specimens failed mainly by tensile fracture with little or no shearing owing to separation of fabric layers. This indicates strong interlayer bonding, as would be expected from the relatively high ILSS in Table 3. Also, Figures 11(f) and 12(g) show relatively small amounts of fibre pull-out, indicating strong fibre/matrix bonding. All these results help to explain the restricted amounts of crack deflection, resulting in less irregular fracture surfaces and lower strengths and moduli compared to the Type A composite.

Comparison of the results with C_f/C_m literature data

The most widespread application of C_f/C_m composites is in braking systems, owing to the high temperature capability combined with excellent friction and wear characteristics, which are interrelated: the interface temperature at the friction surface plays an important role in controlling the wear phenomena of C_f/C_m brake materials.^21,22 For brake discs a higher friction interface temperature, which is mainly controlled by the thermal conductivity (both in-plane and through-thickness), affects the stability of the wear-controlling smooth friction film formed during braking. Higher thermal conductivities, especially in the through-thickness direction, aid heat dissipation and lower the friction interface temperature, thereby reducing nonuniform and uncontrolled wear that causes system vibrations.^22,23

The present study shows that the through-thickness thermal conductivity of a 2D laminated C_f/C_m composite is much improved (see Table 2) by using graphitised carbon fabrics woven with yarns of discontinuous monofilaments (Type B), rather than carbonised fabrics woven with tows of continuous carbon fibres (Type A). In view of this important result, the properties of the Type A and Type B composites are compared in Table 4 with literature data for other C_f/C_m brake materials.

Table 4.

Properties of carbon-carbon (C_f/C_m) composite materials developed for friction and wear purpose.

Material attributes	Materials used in present study		Materials reported in open literature
Material attributes	C_f/C_m: Type A	C_f/C_m: Type B	C_f/C_m¹³	C_f/C_m − 1¹⁴	C_f/C_m − 2¹⁴	C_f/C_m ¹⁵	C_f/C_m¹⁶
Fibre type	PAN-based fabric woven with tows + chopped fibres	PAN-based fabric woven with yarns	PAN-based fabric + chopped fibres	PAN-based fabric	PAN-based fabric	NA	PAN-based fabric + chopped fibres
Weave pattern/ filament size	8H-satin/3K	8H-satin/1.5K equivalent	NA	8H-satin/2K	8H-satin/2K	NA	8H-satin
Matrix type	Resin/pitch	Resin/pitch	Pitch/CVI	Pitch	Pitch	CVI	Pitch
Process	LPIP	LPIP	LPIP + CVI	LPIP	LPIP	CVI	LPIP
Dimensionality	2D	2D	2D	2D	2D	NA	2D
Fibre vol. (%)	45	30	NA	NA	NA	NA	NA
Bulk density (g/cm³)	1.85	1.73	>1.75	1.81	1.72	1.7–1.75	1.87
Open porosity (%)	8	13	NA	NA	NA	NA	NA
Thermal cond. (W/m K)				NA	NA
In-plane	110.4	103.2	80–90			150	91
Thickness	13.6	27.2	25–30			40	27
Temperature (℃)	200	200	NA			NA	NA
Flexural strength (MPa)	165–194	105–110	100–140	148	120	120	NA
Flexural modulus (GPa)	33–42	20–26	NA	NA	NA	29	NA
Tensile strength (MPa)	122–128	78–82	80–110	NA	NA	70	NA
Tensile modulus (GPa)	60–65	31–36	NA	NA	NA	35	NA
ILSS (MPa)	10–11	10–13	10–12	NA	NA	12	NA

NA-Not Available; CVI-Chemical Vapour-Infiltration; LPIP-Liquid Pitch Impregnation and Pyrolysis; ILSS-Interlaminar Shear Strength.

From Table 4, it is seen that almost all the composites were reinforced with PAN-based carbon fabric in 2D layers, i.e. without fibres in the through-thickness direction. The matrix precursors vary, but most were processed by the liquid pitch impregnation and pyrolysis route similar to that used for the Type A and Type B composites.

Table 4 also shows that the Type A composite has superior flexural and tensile properties, but a comparatively high density and low through-thickness thermal conductivity. On the other hand, the Type B composite is seen to have strength, density and thermal conductivities similar to those of the composites investigated by Lee et al.,¹³ Chen and Ju¹⁴ Trefilov¹⁵ and Shin et al.¹⁶ which include materials processed with CVI: these are known for their better fibre/matrix bonding and excellent friction and wear characteristics.¹ Thus it may be inferred that the Type B composite would make a good brake system material, with the advantage that LPIP processing is cheaper than CVI.¹

Summary and conclusions

Two types (Type A and Type B) of carbon fibre-reinforced carbon (C_f/C_m) composites were investigated for their microstructural, physical, thermal and mechanical properties. The composites were obtained via the liquid pitch impregnation–pyrolysis (LPIP) process, using laminated carbon fibre-reinforced polymer composites as precursors. Both types of composite employed PAN-based 8 H-satin carbon fabric, Type A being woven with tows of continuous fibres, whereas Type B used yarns of discontinuous fibres. The following conclusions are drawn:

The Type B composite possessed better fibre/matrix interface bonding and fabric interlayer bonding, owing to the use of yarns of discontinuous fibres in the fabric. Due to the twist, many fibres in the yarns have a significant z-direction component. In addition, there are possibilities of interlocking of the yarns (and layers) by the protruded fibre ends resulting in less likelihood of planar separation of fabric layers.

Although the composites underwent the same number of densification cycles, Type A was ∼7% denser than Type B. This difference was due to about 36% less porosity in Type A, as a consequence of the higher fibre volume (45% compared to 30%).

Even though less dense, Type B had much better through-thickness thermal conductivity than Type A, while the in-plane conductivities were similar. The better through-thickness conductivity of Type B is attributable to many fibres having significant z-direction component, less number of fabric layers in the composite, better fibre/matrix interface bonding and fabric interlayer bonding.

The Type A composite possessed much better flexural and tensile properties than Type B, but less ILSS. These contrasting results have been explained from fractographic analysis: they are due to differences in the starting materials for the carbon fabrics (tows of continuous straight fibres for Type A; yarns of discontinuous fibres held by a twist for Type B) and consequent differences in bond strengths for the fibre/matrix interfaces and the fabric interlayers.

The Type B composite, which was processed solely by the relatively economical LPIP route, is a promising material for aircraft brakes and other braking systems.

Footnotes

Acknowledgements

The present manuscript has been finalised after a critical review and suggestions by Dr. R.J.H. Wanhill, National Aerospace Laboratory (NLR), the Netherlands. The authors are most grateful for his assistance. The authors gratefully acknowledge DRDO funding and the assistance from DMRL and M/s. Midhani, Hyderabad, for the mechanical and microscopy studies. They are also grateful for additional help from colleagues in their organisations; and B. Jana and N. Eswara Prasad particularly thank Dr. K. Tamilmani, Distinguished Scientist, DG (Aero) DRDO and Sri P Jayapal, CE (A) – CEMILAC, for support and encouragement.

Declaration of conflicts of interest

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Savage

. Carbon-carbon composites, London: Chapman & Hall, 1993.

Xiong

Huang

B-y

J-h

. Preparation and mechanical properties of carbon/carbon composites with high textured pyrolytic carbon matrix. Carbon 2006; 44: 463–467.

Venugopalan R, Sathiyamoorthy D and Tyagi AK. Development of carbon/carbon composites for nuclear reactor applications. BARC Newsletter, issue 325, March–April 2012, pp. 16–20.

Venkat Rao

Mahajan

Mittal

. Effect of architecture of mechanical properties of carbon/carbon composites. Compos Struct 2008; 83: 131–142.

Zhao

Guo

Feng

. An artificial neural network model of the chemical vapor deposition processes for carbon-carbon composites. Carbon Sci Tech 2008; 1: 60–65.

Wang

W-X

Takao

Matsubara

. Tensile strength and fracture toughness of C/C and metal infiltrated composites Si-C/C and Cu-C/C. Composites Part A 2008; 39: 231–242.

Kogo

Kikkawa

Saito

. Comparative study on tensile fracture behavior of monofilamentand bundle C/C composites. Composites Part A 2006; 37: 2241–2247.

Abdo

Shamseldin

. Comparative study of friction and wear of two generation of CVI C-C composite. Emirates J Eng Res 2007; 12: 57–67.

Shu

Jie

Qizhong

. Effect of braking speeds on the tribological properties of carbon/carbon composites. Mater Trans 2010; 51: 1038–1043.

10.

Fitzer

. The future of carbon-carbon composites. Carbon 1987; 25: 163–190.

11.

Lalit

. High performance carbon-carbon composites. Sadhana 2003; 28: 349–358.

12.

Yin

Xiong

Zhang

. Microstructure and ablation performances of dual-matrix carbon/carbon composites. Carbon 2006; 44: 1690–1694.

13.

Lee J, Park J-k, Ha H-s, et al. Development of high performance carbon/carbon brake disk for an advanced aircraft. In: Proc. of Eleventh International Conference on Composite Materials-ICCM-11, Gold Coast, Australia, 14-18 July 1997, vol I, pp. I-297-I-301.

14.

Chen

. Friction and wear of PAN/pitch-, PAN/CVI-and pitch/resin/CVI-based carbon/carbon composites. Wear 1994; 174: 129–135.

15.

Trefilov

. Ceramic and ceramic matrix composites, UK: Chapman & Hall, 1995.

16.

Shin

H-k

Lee

H-B

Kim

K-S

. Tribological properties of pitch-based 2-D carbon-carbon composites. Carbon 2001; 39: 959–970.

17.

Wielage

Richter

Mucha

. Influence of fabric parameters on microstructure, mechanical properties and failure mechanisms in carbon-fibre reinforced composites. J Mater Sci Technol 2008; 24: 953–959.

18.

Rohini Devi

Rama Rao

. Carbon-carbon composites-an overview. Def Sci J 1993; 43: 369–383.

19.

Kaw Autar

. Mechanics of Composite Materials, New York: CRC Press, 1997.

20.

Iqbal

Dinwiddie

Porter

. Effect of heat treatment on thermal properties of pitch-based carbon fiber and PAN-based carbon fiber carbon-carbon composites. Ceram Eng Sci Proc 2011; 32: 245–254.

21.

Venkataraman

Sundararajan

. The influence of sample geometry on the friction behaviour of carbon-carbon composites. Acta Mater 2002; 50: 1153–1163.

22.

Byrne