Vous êtes sur la page 1sur 10

Tribology International 102 (2016) 497506

Contents lists available at ScienceDirect

Tribology International
journal homepage: www.elsevier.com/locate/triboint

Structural transformation of carbon/carbon composites for aircraft


brake pairs in the braking process
Shuai Wu, Yunqi Liu, Yicheng Ge, Liping Ran, Ke Peng, Maozhong Yi n
State Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083, China

art ic l e i nf o

a b s t r a c t

Article history:
Received 15 April 2016
Received in revised form
5 June 2016
Accepted 12 June 2016
Available online 14 June 2016

Amorphous structure discovered from both the wear debris and the friction layer of the carbon/carbon
composites after the braking tests was attributed to the breakdown of crystallites in the graphitized bulk
material. Graphite layers produced through interlayer shear deformation were found to be the basic units
of the pyrocarbon during sliding while carbon ber tended to be worn into small fragments along the
ber axis in the primary stage of rubbing. The transition of graphitic pyrocarbon in the friction process
was described as four stages. Nano-scale carbon particles of onion-like structure were found in the ction
layer, which were likely to participate in the formation of friction layer and lubricate the friction process.
& 2016 Elsevier Ltd. All rights reserved.

Keywords:
Carbon/carbon composites
Brake discs
Debris analysis
Friction transitions

1. Introduction
Carbon/carbon (C/C) composites have been widely used as the
brake material for aircrafts due to their light weight, exceptional
friction performance, excellent thermal properties and dimensional stability at severe environment. The frictional properties
and general wear mechanisms of C/C composites were investigated for decades [14].
Because of the complex microstructure, serious deformation as
well as the atomic level interactions between the contact surfaces,
the phenomenon of wear was dened as a dynamic process which
was difcult to model [5]. Kasem et al. [6] presented that both
physico-chemical action and mechanical interaction were
responsible for the tribological transitions between the rst body
surfaces. In addition, the wear debris (third body) formation was
mainly due to the abrasion and fatigue of the rst bodies.
Researches about the wear debris of C/C composites showed
that the debris comprised a much lower structural order phase
compared with the well-graphitized parent matrix [7], which was
revealed through X-ray diffraction (XRD) study [8,9] as well as
transmission electron microscopy (TEM) analysis [1012]. Friction
layer of amorphous structure was also nonuniformly formed in the
dynamic wear process [13]. It was widely accepted that disordered
carbon layer formed on the friction surface protected the bulk
composites from further wearing and degradation [12], however,
how the disordered carbon in the friction layer was transformed in
n

Corresponding author.
E-mail address: yimaozhong@126.com (M. Yi).

http://dx.doi.org/10.1016/j.triboint.2016.06.018
0301-679X/& 2016 Elsevier Ltd. All rights reserved.

the braking process was barely reported. Molecular dynamics


simulation of nanoscratching on multilayer graphene showed that
graphene would partly turn to amorphous structure due to
numerous cross-linking between neighbor-layer graphenes
[14,15]. The transformation of 3D graphite into a 2D turbostratic
phase constituting high degree of stacking interlayer disorder was
considered to increase the friction coefcient of crystalline graphite at high loading range [16]. In general, the tribological
property of carbon material was surely correlative with the
structural transformation during the dynamic friction process.
This paper presented detailed microstructure information of
the wear debris and friction layers of the C/C composites for aircraft brake discs after braking tests. The structural transformations
from graphitic structure to amorphous type structure were
explained and modeled. The wear-induced structural transition
was briey discussed.

2. Material and methods


2.1. Material preparation
The C/C composites used in this study were manufactured from
quasi-three dimensionally needled polyacrylonitrile (PAN)-based
carbon ber preform which was fabricated by alternatively staked
nonwoven ber cloth and chopped ber felt by a needle-punching
technique. The preform was densied to 1.60 g/cm3 by chemical
vapor inltration in an atmosphere containing propylene as precursor and hydrogen as carrier gas followed by furan resin
impregnation-carbonization process to the density of 1.80 g/

498

S. Wu et al. / Tribology International 102 (2016) 497506

cm3. The nal heat treatment temperature of the composites was


2300 C.
2.2. Material characterization
The optical microstructure of the C/C composites was observed
using a polarized microscope. The graphitization degree was
obtained by XRD analysis (Rigaku-3014). Compressive strength
was tested on specimens (10 mm  20 mm) in the direction
perpendicular ( ? ) to the friction surface. Flexural strength was
tested on bar specimens (55 mm  10 mm  4 mm) using the
three-point bending method with a span of 40 mm. Thermal diffusivity was measured according to ASTM E1461-11 (Standard Test
Method for Thermal Diffusivity by the Flash Method) by a ash
laser technique on a JR-3 property tester with samples
(10 mm  4 mm) at room temperature. Thermal conductivity
was calculated according to

Table 1
Braking parameters of normal landing and overload landing conditions.
Test conditions

Braking pressure
(MPa)

Normal landing
8
Overload landing

Kinetic energy
(MJ)

Braking speed
(km/h)

9.14
11.43

240
250

418:68C p
where -thermal conductivity, W/(m K); -density of the material,
g/cm3; -thermal diffusivity, cm2/s; Cp-the specic heat capacity.
Usually, the value of Cp at room temperature is considered to be
0.171 cal/(g K).
2.3. Braking tests
Fig. 1. Worn surface of the C/C brake disc after braking tests.

Braking tests were performed on the HJDS-II simulation tester


with actual size airplane brake discs as well as actual aircraft brake
operations according to the National Aerospace Standards of
China: HB 5434.4-2004 (Test methods for aircraft wheel friction
materials-Part4: Dynamometer test method for brake performance). The dimensions of the rotator discs were 336 mm and
212 mm in outer and inner radius respectively, and 20 mm in
thickness. The dimensions of the stator discs were 312 mm and
188 mm in outer and inner radius respectively, and 22 mm in
thickness. After running in, braking tests were conducted under
aircraft normal landing and overload landing conditions. Braking
parameters of normal landing and overload landing conditions
were shown in Table 1. Temperature at the middle of the stator
was measured through a thermocouple located approximately
10 mm below the friction surface. All the braking tests were
exposed to atmosphere with a 40% relative humidity.
2.4. Wear debris and friction layer observation
Wear debris was collected from the friction surfaces of the
rotor and stator discs (Fig. 1) after all the braking tests, of which
the morphologies were observed by scanning electron microscopy
(SEM, Helios Nanolab 600i) and TEM (Titan G2 60-300 with image
corrector). Furthermore, friction layers for TEM observation were
prepared according to the method described in the literature [12].

3. Results and discussions


3.1. Microstructure and properties of the C/C composites
Fig. 2a showed the optical microscopy micrograph of the C/C
composites. The pyrocarbon (PC) deposited around the carbon
ber (CF) was mainly rough laminar texture and resin-derived
carbon (RC) was lled in the interspace. There existed cracks
(defects) between RC and PC, which was due to their different
thermal expansion coefcients during the high temperature
treatment process of the composites. The PC displayed a welldeveloped and continuous graphitic structure (Fig. 2b) while the
CF had a partially crystalline structure with graphene layers

aligned basically along the ber axis direction (Fig. 2c). The RC
showed a glass-like carbon structure of entangled graphene layers
(Fig. 2d) which was the less-ordered phase in the C/C composites.
In the real usage of braking, the discs are subjected to shear
forces as well as impact loads. Resistance to failure under such
conditions is very important for the stability of the brake material.
The physical properties of the C/C composites in this study were
listed in Table 2.
Fig. 3 showed the coefcient of friction (COF) and the temperature measured by thermocouple during every normal landing
test. The COF stabilized at around 0.30 and the average measured
temperature at the stator disc of all the normal landing tests was
570 C. As listed in Table 2, the thermal conductivity perpendicular
to the friction surface was lower than that parallel to the friction
surface. Therefore, before the frictional heat was transferred to the
measure point (10 mm below the friction surface), part of the heat
was already conducted into the air along the direction parallel to
the friction surface. Generally, the measured temperature was
much lower than the true temperature of the friction interface.
The overload landing tests were arranged after the 10th, 18th
and 26th normal landing tests. The average COF of overload
landing tests was 0.32 and the average measured temperature was
680 C. The average sliding distance of normal landing tests and
overload landing were 440 m and 525 m, respectively.
Researches showed that higher speed would produce higher
frictional heat which might inuence the formation of friction lm
and the wear mechanism[17,18]. Overload landing tests of higher
braking speed and kinetic energy produced higher frictional work
and increased the temperature of the disc, which inuenced the
thermal stress distribution as well as the formation of friction
layer at the contact surfaces. The structure in the friction layer
might be rearranged and the thickness and size of the friction
layer were changed under the overload landing, leading to a
higher COF (0.32). It suggested that the steady wear surface
formed through normal landing was severely damaged after the
overload landing test. What is more, the structural damage of the
wear surface also increased the COF of the normal landing tests
following overload landing. As was seen from Fig. 3, it was noted

S. Wu et al. / Tribology International 102 (2016) 497506

499

Fig. 2. Microstructure of the investigated C/C composites: (a) optical microscopy micrograph of the C/C composites; (b) HRTEM and the fast Fourier transform image of
pyrocarbon; (c) HRTEM and the fast Fourier transform image of carbon ber along the ber axis; (d) HRTEM and the fast Fourier transform image of resin-derived carbon.
Table 2
The properties of the C/C composites in this study.
Physical property
Graphitization degree
Compressive strength ( ? )
Flexural strength ( ? )
Thermal conductivity coefcient (room temperature)

?
//

63.2%
190.2 7 8 MPa
186.17 11 MPa
49.9 73 W/(m K)
128.0 78 W/(m K)

that the normal landing tests after overload landing, the 11th, 19th
and 27th, exhibited extremely higher COF at about 0.35 and
completely stop in a shorter time. It could be explained that the
badly worn wear surface at overload landing test reduced the
contact area between the discs, causing the temporary increase of
COF at the following normal landing test. When the braking

condition was changed, the friction layer would be destroyed and


then restored repeatedly. As long as the destructionrecovery
process of the friction layer reached a dynamic balance state, the
COF would be stable. The damaged surfaces could be repaired by
forming stable friction layer under the normal landing tests and
the COF gradually decreased and remained around 0.28. The
results also proved that the COF of C/C composites was correlated
with the wear surface morphology and the wear surface characteristics were sensitive to the test conditions [18,19].
3.2. SEM characterization of the wear debris
C/C composites with rough laminar pyrocarbon were reported
to have nearly constant COF and stable friction behavior under
various braking conditions [20], which was related to the wear
process of the highly ordered texture of pyrocarbon.

500

S. Wu et al. / Tribology International 102 (2016) 497506

Fig. 3. The coefcient of friction and the measured temperature of every normal landing test.

Fig. 4. (a) SEM morphology of the detached particle debris; (b) high magnication SEM morphology of pyrocarbon debris.

The wear debris obtained at the friction interface of the rotor


and stator discs was actually components of newly produced,
slightly worn, severely worn debris and deciduous friction layers.
Fig. 4a showed a detached particle debris which appeared to be
newly produced from the bulk material because of the shear stress
during friction. The microstructure of the worn section presented
separated thin layers of graphitic pyrocarbon with tens of nanometers in thickness (Fig. 4b), which implied that the shear
deformation happened between the graphene layers of
pyrocarbon.
Fig. 5a showed the morphology of debris formed during the
dynamic wear at the friction surface, consisting mainly of carbon
ber fragments, matrix particles and thin aky pieces. The thin
aky pieces (Fig. 5b and Fig. 5c) produced from the shear deformation of graphitic pyrocarbon indicated that the basic unit of the
ordered pyrocarbon at the sliding process was the graphite layers.
3.3. Microstructure of the wear debris characterized by TEM
Due to the shear behavior of the pyrocarbon in C/C composites,
plenty of exfoliated graphite layers were found among the wear
debris. As was shown in Fig. 6a, when newly produced by interlayer slipping of the pyrocarbon, the detached graphite sheets of
several layers still remained relatively ordered and intact structure. During the rst period of rubbing, the graphitic structure
fractured into small basal planes. The morphology of the stacking
graphite sheets at position was shown in Fig. 6b. The HRTEM

image of the marked area in Fig. 6b and its reconstructed image


produced by using a mask applied to the 2-D fast Fourier transform were shown in Fig. 6c and d. It revealed that few hexagonal
crystal lattices could be found and defects were induced within
the graphene planes due to the mechanical rubbing and interaction among the debris. Fig. 6e showed the morphology at position
and the marked area indicated that thin graphite sheets were
easily folded within the graphene planes. The HRTEM image and
its fast Fourier transform image of the marked area in Fig. 6e were
shown in Fig. 6f and g. The HRTEM image of the folded area displayed an ordered stacking of nearly 50 graphene layers, which
indicated that the aky wear debris that was produced from the
pyrocarbon had at least 50 graphene layers. Fig. 6g showed (002)
diffraction arc, (100) diffraction ring, (004) diffraction arc and
(110) diffraction ring, successively. Strong (002) and (004) diffraction arc implied that the graphene layers were still orderly
stacked with highly textured microstructure in the newly produced wear debris. The (100) and (110) diffraction rings indicated
that defects were induced within the graphene planes.
Such aky debris of different morphologies could be considered as the intermediate phases in the transition from ordered
graphitic structure to disordered structure. When produced from
the graphitic pyrocarbon, the thin graphite sheets of several layers
with rolled-up edges (Fig. 7a) were slightly deformed both at the
interlaminar direction and within the basal planes. With further
rolling, defects and broken holes were induced within the basal
planes (Fig. 7b), which might enhance the disruption of graphite

S. Wu et al. / Tribology International 102 (2016) 497506

501

Fig. 5. SEM morphologies of the wear debris: (a) an agglomeration of wear debris; (b) high magnication image of the aky debris at position ; (c) high magnication
image of the aky debris at position .

Fig. 6. TEM analysis of the graphite sheets in wear debris: (a) morphology of the detached graphite sheets of several layers; (b) morphology, (c) HRTEM image and (d) its
reconstructed image (produced by using a mask applied to the 2-D fast Fourier transform) of the graphite sheets at position ; (e) morphology, (f) HRTEM image and (g) the
fast Fourier transform image of folded area at position .

502

S. Wu et al. / Tribology International 102 (2016) 497506

Fig. 7. TEM morphology of the aky debris: (a) detached graphite sheets with distortion and folded edges; (b) morphology and SAED pattern of broken sheets with wearinduced defects or holes; (c) HRTEM image and its fast Fourier transform image of the marked area.

Fig. 8. TEM analysis of the agglomerated debris and amorphous structure debris: (a) morphology, (b) HRTEM image and (c) the fast Fourier transform image of the
agglomerated debris; (d) morphology, (e) HRTEM image and (f) SAED pattern of the amorphous structure debris.

layers. The selected area electron diffraction (SAED) of the broken


sheets exhibited (100) and (110) diffraction rings. It was noticeable
that both the rings were approximately elliptic and discontinuous,
which may be related with the defects or deformations of the
graphene planes. HRTEM image of the marked area were presented in Fig. 7c. The exposed edges as well as holes within the
worn-out sheets produced more active dangling bonds. They could

act as adsorption sites for easier adsorption of vapor (water or


oxygen) with the carbon atoms, which had a lubricative effect on
the sliding surface [6,21]. The porous structure of the broken
sheets also increased the specic surface area of debris [11], which
explained the previous results that the debris was much richer in
oxygen than the parent C/C composites [8,9]. In addition, due to
the graphite layers produced from graphitic structures, debris

S. Wu et al. / Tribology International 102 (2016) 497506

503

Fig. 9. (a) TEM morphology and (b) HRTEM image of the worn carbon ber fragments.

from ordered carbon could contain more carboxylic groups and


larger specic surface areas than that from less ordered carbon
material.
Fig. 8a showed the morphology of agglomerated debris consisting of the damaged plane pieces with varied orientations.
Disordered structure with partial ordered fragments was observed
in Fig. 8b. Strong (100) and (110) diffraction rings and weak (002)
diffraction arc were seen from the fast Fourier transform image of
this region (Fig. 8c). The tearing of the (002) basal planes and the
disbandment of the well-organized texture eventually transformed the partial ordered structure to the amorphous structure
(Fig. 8d) which was the main phase formed at the friction interface. Fig. 8e showed the HRTEM image of the amorphous structure
where no obvious (002) carbon plane could be seen. As expected,
only (100) and (110) diffraction rings were distinctly observed
from the SAED pattern shown in Fig. 8f.
The PAN-based carbon ber served as the reinforcement in C/C
composites with its better hardness and stiffness than rough
laminar pyrocarbon, which was attributed to the microstructure of
carbon ber that graphite basal planes were preferentially oriented along the ber axis direction (Fig. 2c). Fig. 9a shows the
morphology of scraped fragments of the worn carbon ber. It
signied that the wear behavior of carbon ber mainly occurred
along the ber axis. The irregular ber breakage (Fig. 9b) showed
an appearance of twisted and intermingled texture. It appeared
that slippage of the graphite layers might hardly occur in carbon
ber. The fragment was likely to be the basic unit of carbon ber
during the wear process and act as a relatively hard phase during
the friction.
As for resin-derived carbon, which was the hardest but fewest
component in the studied C/C composites, the wear behavior
mainly began at the cracks between pyrocarbon and resin-derived
carbon. Particulate wear debris was produced by the brittle
deformation of resin-derived carbon under shear stress [22].
3.4. Microstructure of the friction layers
In Fig. 1, distinguished morphology was seen in the radical
direction of the disc. The middle area presented a lustrous black
appearance called bright band while the outer area showed a
relatively dumb grey appearance named dark band. The inner

area showed an appearance between bright band and dark


band. The different surface features of the discs after the braking
test can be explained by the different contact pressure as well as
friction layers formed at the friction surfaces. During the braking,
there existed thermal gradient in the radical direction of the discs
due to the non-uniform frictional heat generation which caused
unequal thermal expansion of the contact surface [23]. The temperature eld results simulated by the ANSYS program showed
that the heat ux was proportional to the radius of the disc and
the most friction heat was generated at the outer radius of the
friction surface [24]. The highest thermal stress brought negative
inuence on the formation of the friction layer in the outer area,
leading to a rough surface named dark band without continuous
friction layer coverage. The middle area of the disc, with low
contact pressure and less oxidation, had continuous friction layer
and a bright band appearance. It was reported that both the
outer and inner edges of the discs were quite susceptible to oxidation due to the present atmospheric oxygen during the braking
tests, which lead to oxidative wear of the friction surface [25].This
oxidation brought negative effects on the formation of dense and
continuous friction layer. Due to the oxidative attack of inner
edges of the discs, thinner friction layer with cracks or gaps was
formed at the inner area and an appearance between bright
band and dark band was seen.
Fig. 10 showed TEM micrographs of the friction layers obtained
from the wear surface of the aircraft brake discs after all the
braking tests. As seen in Fig. 10b, the friction layer in the middle
area displayed a dark and dense vision without cracks or gaps,
while friction layers of the outer (Fig. 10a) and inner diameter
(Fig. 10c) showed gray appearances with smaller thicknesses and a
few cracks. When comparing the friction layer of outer diameter to
that of the inner diameter, the latter seemed denser than the
former. In addition, all the obtained friction layers mainly showed
an amorphous structure (Fig. 10d).
It was notable that agglomerated spherical or elliptical nanoscale carbon particles (Fig. 11a) were discovered in the friction
layer from the inner diameter of the brake disc. The average diameter of the particles was around 50 nm and no individual particle
could be identied as the particles were compacted or connected
with each other. The HRTEM image (Fig. 11b) of the particle
showed that partially graphitized layers were stacked

504

S. Wu et al. / Tribology International 102 (2016) 497506

Fig. 10. Morphology of the friction layers obtained at the (a) outer diameter, (b) middle area and (c) inner diameter of the brake discs; (d) HRTEM image and the SAED
pattern of the amorphous structure in the friction layers.

concentrically, forming an onion-like structure. However, the


organization of the layers was less developed and the SAED pattern exhibited weak (002), (100) and (110) diffraction rings. It was
inferred that the nanoparticles were transformed by the wear
debris rolling at the friction surface and participated in the
forming of friction layer. The rolling and compacting of the particles could be considered as solid lubrication effect [7], which
might improve the friction and wear in the sliding contact at some
extent.
3.5. Structure transformation during the wear process of carbon/
carbon composites
Wear-induced amorphization phenomenon was essentially the
breakdown of crystallites in the graphitized parent material,
which was one of the typical structure transformation occurred

during the wear of C/C composites. The breakdown of the graphitic


(sp2 dominant) structure in C/C composites would produce wear
debris of different crystalline structures and worn degrees which
acted as three body abrasives during the braking process. From the
microstructure analyzes in this study, a structure transformation
model could be concluded.
As displayed in Fig. 12, before totally transforming into amorphous structure, the intact graphitic layers in pyrocarbon experienced four major evolutionary stages: (i) detached graphite sheets
of several layers, (ii) graphite sheets of small thickness with distortion and interlayer slipping, (iii) broken sheets with wearinduced defects or holes and (iv) agglomerations of damaged
plane pieces with varied orientations. During the stage (iii), most
graphite sheets would be broken down and the produced carbon
dangling bonds would be passivated by absorbing vapor (water or
oxygen), which lubricated the friction process. What is more, the

S. Wu et al. / Tribology International 102 (2016) 497506

505

Fig. 11. (a) Morphology and (b) HRTEM image of the nano-scale carbon particles observed in the friction layer.

Fig. 12. Structure transformation schematic of pyrocarbon and carbon ber during the wear process of carbon/carbon composites.

energy from bond brokenness as well as bond formation would


also affect the COF.
The carbon ber was preferentially broken along the ber axis
into small fragments while resin-derived carbon mainly produced
particulate debris. The aky sheets produced from the graphitic
pyrocarbon as well as the onion-like nanoparticles transformed
from the wear debris mainly had lubricant effects at the friction
interfaces. Meanwhile, the carbon ber fragments together with
resin-derived carbon particles mainly acted as the grinding media
among all the debris and increased the friction force of the contact
surfaces.
The continuous deformation of the (002) basal planes and the
disbandment of the sp2 dominant texture from the bulk material
happened throughout the wear process. Massive turbostratic or
amorphous carbon would preferentially accumulate in the valleys
of asperities at the rubbing interface, which formed friction layers

on the worn surface and protect the composites from further


wearing. Because of the thermal gradient and contact pressure
distribution at the friction interface, friction layers were not uniform over the whole contact area. As a result of the dynamic
friction process, there would be a balance of the formation and
destruction of friction layers.
The disruption of the intact graphite sheets and the formation
of amorphous structure were explained as the transformation of
carbon atoms from sp2 hybridization state to sp3 hybridization
state. However, the local stress and high temperature at the contact surfaces during braking could lead to the localized transformation from amorphous structure to highly ordered sp2 dominant
structure, which was described as stress-induced graphitization in
previous study [12]. Actually, it was evident to say that the sp3
carbon atoms in the friction lm can be rearranged during the
friction process and partly turns to sp2 hybridization state. What's

506

S. Wu et al. / Tribology International 102 (2016) 497506

more, because of the dynamic wear process, the mutual transformation of sp3 and sp2 was always happening in the friction
interface. Such transformation would also have positive inuences
on the energy balance of the system, contributing to the stable
friction process of C/C composites.
In the braking of C/C composites, three dynamic balances
beneted the stable process. Firstly, from the perspective of
components in the composites, the soft pyrocarbon and relatively
hard carbon ber as well as resin-derived carbon would produce
wear debris of different types and basic units. The stable COF of
the composites resulted from the dynamic balance of interactions
of different carbon phases at the friction interface. Secondly, the
disruption of the sp2 dominant structure and the formation of the
sp3 dominant amorphous structure as well as the stress-induced
graphitization made up the dynamic sp2sp3 mutual transformation, which also contributed to the energy balance in the system.
Thirdly, there would be a dynamic balance of the formation and
destruction of friction layers at certain brake conditions. Continuous friction layer was related to the stable COF and low wear of
the C/C composites. Such structural transformations in the friction
interfaces jointly played positive roles in the stabilized friction
process of C/C composites.

4. Conclusions
From the analyzes of the debris and wear surface produced by
C/C composites for aircraft brake pairs after braking tests, the
following conclusions can be drawn.
(1) The basic unit of the pyrocarbon was the graphite layers
produced by interlayer shear deformation. Flaky wear debris
of graphite sheets experienced four major stages in the transition from ordered graphitic structure to amorphous type
structure, during which adsorption sites were produced in the
broken sheets to enhance the adsorption of vapor (water or
oxygen), leading to a lubricative effect on the sliding surface.
The wear of carbon ber mainly occurred along the ber axis
in the primary stage of rubbing.
(2) Friction layers formed at the worn surface were not uniform
from the outer diameter to the inner diameter and the middle
area was quite plausible to form a dense and continuous
friction layer. Onion-like nanoparticles were observed and
might serve as solid lubrication and participate in the formation of continuous friction layer on the wear surface.
(3) Dynamic balance of interactions of different carbon components in the friction interface, the sp2sp3 mutual transformation of carbon atoms as well as the dynamic balance of the
formation and destruction of friction layers jointly beneted
the stable friction process of C/C composites.

Acknowledgments
This research was nancially supported by the National Basic
Research Program of China (125-GH-116) and the State Key
Laboratory for Powder Metallurgy Foundation (2016), Central
South University, Changsha, China.

References
[1] Hutton TJ, Johnson D, McEnaney B. Effects of bre orientation on the tribology
of a model carboncarbon composite. Wear 2001;249:64755.
[2] Shin H, Lee H, Kim K. Tribological properties of pitch-based 2-D carboncarbon composites. Carbon 2001;39:95970.
[3] Ozcan S, Filip P. Microstructure and wear mechanisms in C/C composites.
Wear 2005;259:64250.
[4] Lee K, Tsai M, Cheng H. Effect of densication parameters on the low-energy
tribological behavior of carbon/carbon composites. Carbon 2010;48:135361.
[5] Zhao S, Hilmas GE, Dharani LR. Numerical simulation of wear in a C/C composite multidisc clutch. Carbon 2009;47:221925.
[6] Kasem H, Bonnamy S, Berthier Y, Dufrnoy P, Jacquemard P. Tribological,
physicochemical and thermal study of the abrupt friction transition during
carbon/carbon composite friction. Wear 2009;267:84652.
[7] Gong Q, Li Z, Zhang Z, Wu B, Zhou X, Huang Q, et al. Tribological properties of
carbon nanotube-doped carbon/carbon composites. Tribol Int 2006;39:937
44.
[8] Pevida C, Jacquemard P, Joly JP. Physicochemical properties of debris ejected
from C/C brakes with different structural orders. Carbon 2008;46:9941002.
[9] Rietsch JC, Dentzer J, Dufour A, Schnell F, Vidal L, Jacquemard P, et al. Characterizations of C/C composites and wear debris after heavy braking demands.
Carbon 2009;47:8593.
[10] Lee KJ, Chern Lin JH, Ju CP. Microstructure study of PAN-pitch carboncarbon
composite lubricative lm. Mater Chem Phys 2003;78:7606.
[11] Franois M, Joly JP, Kapsa P, Jacquemard P. A temperature-programmed desorption and oxidation investigation of wear debris from carbon/carbon composite aircraft brakes. Carbon 2007;45:12431.
[12] Lei B, He L, Yi M, Ran L, Xu H, Ge Y, et al. New insights into the microstructure
of the friction surface layer of C/C composites. Carbon 2011;49:455462.
[13] Murdie N, Ju CP, Don J, Fortunato FA. Microstructure of worn pitch/resin/CVI CC composites. Carbon 1991;29:33542.
[14] Zhang Q, Diao D, Yang L. Dangling bond induced cross-linking model in
nanoscratched graphene layers. Surf Coat Technol 2013;237:2303.
[15] Zhang Q, Diao D, Kubo M. Nanoscratching of multi-layer graphene by molecular dynamics simulations. Tribol Int 2015;88:858.
[16] Kumar N, Kozakov AT, Ravindran TR, Dash S, Tyagi AK. Load dependent friction
coefcient of crystalline graphite and anomalous behavior of wear dimension.
Tribol Int 2015;88:2809.
[17] Gopal P, Dharani LR, Blum FD. Hybrid phenolic friction composites containing
Kevlars pulp: part I. Enhancement of friction and wear performance. Wear
1996;193:199206.
[18] Gopal P, Dharani LR, Blum FD. Hybrid phenolic friction composites containing
Kevlars pulp: part IIwear surface characteristics. Wear 1996;193:1805.
[19] Ju CP, Chern Lin JH, Lee KJ, Kuo HH. Multi-braking tribological behavior of
PAN-pitch, PAN-CVI and pitchresin-CVI carboncarbon composites. Mater
Chem Phys 2000;64:196214.
[20] Xiong X, Huang B, Li J, Xu H. Friction behaviors of carbon/carbon composites
with different pyrolytic carbon textures. Carbon 2006;44:4637.
[21] Nian J, Si Y, Guo Z. Advances in atomic-scale tribological mechanisms of solid
interfaces. Tribol Int 2016;94:113.
[22] Lei B, Yi M, Xu H, Ran L, Ge Y, Peng K. Effect of resin-derived carbon on the
friction behavior of carbon/carbon composites. Tribol Lett 2011;41:3718.
[23] Bing KY, Ishihara T. The surface morphology and structure of carbon-carbon
composites in high-energy sliding contact. Wear 1994;174:1117.
[24] Xu HJ, Huang BY, Yi MZ, Xiong X, Lei BL. Inuence of matrix carbon texture on
the temperature eld of carbon/carbon composites during braking. Tribol Lett
2011;44:1824.
[25] Byrne C. Modern carbon composite brake materials. J Compos Mater
2004;38:183750.