Wear 263 (2007) 1243–1248
Short communication
Optimization of steel wool contents in non-asbestos organic (NAO)
friction composites for best combination of thermal conductivity
and tribo-performance
Jayashree Bijwe ∗ , Mukesh Kumar
Industrial Tribology Machine Dynamics and Maintenance Engineering Centre, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110016, India
Received 13 September 2006; received in revised form 4 January 2007; accepted 6 January 2007
Available online 23 May 2007
Abstract
Thermal conductivity (TC) of friction composites plays a vital role in the performance of the composites. Low TC renders the tribo-surface
vulnerable for degradation of organic ingredients and affecting the braking capability adversely. Too high TC, on the other hand, results in adverse
effect on brake-fluid. The optimum TC as a result of tailoring the composite with right contents of metallic fillers is hence desired. However, hardly
any in depth efforts are placed in this direction. Hence, in this work, three friction composites were developed using same ingredients in same
proportion except steel wool and barite which were added in a complementary manner. The compositions containing 4, 8 and 12 wt.% of steel
wool (and inert filler barite in 31, 27, and 23 wt.%) were developed as brake-pads and designated as S1 , S2 and S3 , respectively.
The brake-pads were tested for their friction and wear behavior on a Krauss type RWDC 100, 450 V/50 Hz machine in simulated braking condition
against a cast iron commercial disc and evaluated as per ECR 90 test schedule. It was observed that with increasing metal contents, mechanical
properties decreased, TC increased slowly and other important friction properties including fade resistance improved. Wear properties however,
did not show any correlation with amount of steel wool, TC or mechanical properties. Overall, composite S3 proved to be the best performer. SEM
studies proved helpful in understanding the wear mechanism.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Non-asbestos organic (NAO) friction composites; Thermal conductivity; Fade and recovery; Brake-pad materials
1. Introduction
Non-asbestos organic composite friction materials are being
increasingly used in automotive brake disc pads and clutch
facing applications. They are essentially multi-ingredient systems in order to achieve the desired amalgam of performance
properties [1] amongst which the coefficient of friction and its
stability under various operating conditions, such as temperature, pressure and speed is most important. The four categories
of ingredients, viz. binders, fibers, friction modifiers and fillers
based on the major function they perform apart from contributing towards friction and wear performance are selected. Binders
mainly from phenolics or modified phenolics provide mechanical integrity to the friction material while fibers, such as mineral,
ceramics, organic and metallic types provide mainly the strength.
∗
Corresponding author. Tel.: +91 11 26591280; fax: +91 11 26591280.
E-mail address: [email protected] (J. Bijwe).
0043-1648/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2007.01.125
Friction modifiers, such as abrasives and solid lubricants are used
to achieve the desired range of friction. Fillers the fourth class
is again subdivided as functional fillers (to enhance the specific
function, such as resistance to fade, thermal conductivity, etc.)
and space fillers/inert fillers (just to cut the cost without any
adverse effect on performance properties). The effect of such
ingredients on tribo-performance [2–4] and fade and recovery
behavior [5–11] is extensively studied in the literature. Thermal properties, such as specific heat, thermal conductivity (TC),
diffusivity, thermal expansion, etc. also play a vital role in performance of brake-pads, especially when braking is severe. Low
TC renders the tribo-surface vulnerable due to accumulation of
frictional heat which leads to the degradation of organic ingredients and affects the braking capability adversely. Too high TC on
the other hand, results in adverse effect on brake-fluid because
of conduction of heat towards piston and brake-fluid via back
plate. The optimum TC as a result of tailoring the composite
with right contents of metallic fillers is hence desired. In spite
of this fact, literature pertaining either to the ideal range of these
1244
J. Bijwe, M. Kumar / Wear 263 (2007) 1243–1248
parameters or influence of such ingredients on these parameters is not available. Moreover, it is also important to note that
an ingredient added to tailor a specific property of a composite
may influence other important performance properties in a beneficial or undesirable manner and hence has to be monitored.
Any research papers could not be available on the optimization of thermal conductivity and tribo-performance of friction
composites. Hence, in the present work, three NAO composite
were developed by varying steel wool and barite (inert filler) in
compensatory manner keeping apparent composition unaltered.
Since influence of thermal conductivity is best reflected in severe
operating conditions, fade and recovery mode of performance
testing was selected in the present work. The comparative studies
along with worn surface analysis are reported in this paper. Fade
indicates the loss in braking effectiveness of a material at elevated temperature (typically in the range 300–400 ◦ C) because
of a reduction in the kinetic friction coefficient (μ). High fade
is undesirable since it affects the deterioration in efficiency and
reliability of brakings in severe operating conditions. The recovery term is used to quantify the revival of μ when the pad gets
cooled after severe brakings. Since fade and recovery behavior indicates the potential of composites in very harsh operating
conditions, it was selected for this research study.
2. Experimental
2.1. Fabrication of the composites
The fabrication of composites containing twelve ingredients
was based on keeping parent composition of 10 ingredients
(65%, w/w) constant and varying two ingredients, viz. steel wool
and barite (35%, w/w) in complementary manner as shown in
Table 1 based on a systematic increase in steel wool (4, 8 and
12 wt.%). The parent composition contained straight phenolic
resin (10%), functional fillers, such as alumina, graphite, vermiculite, cashew dust (27%, w/w) and fibers, such as glass, PAN,
Lapinus, Aramid and brass (28%, w/w). The ingredients were
mixed in a plough type shear mixer to ensure the macroscopic
homogeneity using a chopper speed of 2800 rpm and feeder
speed 1440 rpm. The addition of ingredients during mixing was
in a particular sequence. Aramid and glass fibers were mixed
initially for 2 min followed by other pulpy materials and finally
by powdery materials. The mixing schedule was of ten minutes
duration. The mixing sequence and time of mixing of each lot
of ingredients lead to proper uniformity in the mixture. If mixing time is low, proper homogeneity cannot be achieved. If it is
too high, it does not improve the homogeneity further. Hence, it
has to be optimized which was done in the laboratory in earlier
Table 1
lists the varying ingredients in composites
Ingredients
S1 (wt.%)
S2 (wt.%)
S3 (wt.%)
BaSO4
Steel wool
31
4
27
8
23
12
Total
35
35
35
work [12]. The mixture was then placed into a four-cavity mould
supported by the adhesive-coated back plates. Each cavity was
filled with approximately 80 g of the mixture and heat cured
in a compression-moulding machine under a pressure of 8 MPa
for 7–8 min at curing temperature. Three intermittent ‘breathings’ were also allowed during the initiation of curing to expel
volatiles. The pads were then removed and were then post-cured
in an oven at 100 ◦ C for 8 h. The post curing operation is done to
cure the residual resin. As per standard practice in the literature
after optimization, it is done at 100 ◦ C for 8 h to ensure complete
curing of the resin [12]. The surfaces of the pads were then polished with a grinding wheel to attain the desired thickness and
smooth surface. Three composites were designed as S1 , S2 and
S3 accordingly (Table 1).
2.2. Characterization of the composites
For mechanical strength testing, test bars were fabricated
as per ASTM standards. Composites were characterized for
physical (density, water swelling, heat swelling, void contents) (details of the methods are given in Appendix A) and
chemical properties (acetone execration). All the results are
shown in Table 2. Thermal conductivity was measured as
per ASTM-E1461-01 standard on FL-3000 Flash line instrument supplied by Anter Corporation, USA. Square samples of
size 10 mm × 10 mm and thickness 2–3 mm were used for TC
measurement at room temperature. Thermal conductivity of a
commercial brake-pad (composite C) was also measured to compare with that of developed composites. (The Fade & Recovery
study on this material, however, was not done because details of
its ingredients were not available. The data on TC and Fade &
Recovery studies could not be correlated with the composition.)
2.3. Test set-up and procedure for fade and recovery
evaluation by ECR 90 test schedule
The fade and recovery tests were conducted using a Krauss
type RWDC 100 C (450 V/50 Hz) machine discussed elsewhere
[11–12]. A gray cast iron rotor disc of the passenger car (Maruti
Suzuki 1000 cc) with radius 0.12 m and average roughness (Ra )
value 3.28 ␮m was used as a counterface. Details about testing
method are explained elsewhere [13]. The software computes
the various values (as shown in Table 3) as an essence of fade
Table 2
Mechanical properties of the composites evaluated in the laboratory
Properties
S1
S2
S3
Density (g/cc)
Acetone extraction (%)
Water absorption (%)
Void content (%)
Heat swelling (%)
Tensile strength (MPa) ASTM D638
%Strain (ASTM D638)
Flexural strength (MPa) ASTM D790
Flexural modulus (GPa) (ASTM D790)
%Strain (ASTM D790)
2.27
0.169
0.60
37
1.54
22.56
3.10
46.65
11
0.53
2.29
0.141
1.05
39
2.78
20.47
2.97
44.64
10.7
0.49
2.32
0.181
0.70
40
2.94
17.57
2.72
38.74
9
0.49
J. Bijwe, M. Kumar / Wear 263 (2007) 1243–1248
1245
and recovery plot. The various performance parameters derived
from the plots are defined in Appendix B. Wear volume of the
composites was calculated using weight loss and the density of
the materials. The change in roughness of the discs after the test
was also measured. SEM studies on worn surfaces were also
carried out to understand the wear mechanisms.
3. Results and discussion
As seen from Table 2, it was observed that the density of
composites showed increasing trend because of addition of steel
wool which is heavier than the filler barite. Acetone exaction
indicates amount of uncured resin in the composites which was
negligible in all the composites. Void contents increased with
increase in steel wool. The heat swelling also showed the same
trend as anticipated because of higher expansion of steel wool.
All the mechanical properties, (tensile and flexural) decreased
with increase in contents of steel wool and decrease in barite%.
It could be because of the fact that barite being powdery filler
has better binding capacity with the resin rather than steel wool.
3.1. Fade and recovery behavior
Seven steps of friction assessment test are shown in a single
graph. Three such graphs for selected composites (S1 , S2 and S3 )
are shown in Fig. 1(a–d). Following trends were observed from
the seven step friction assessment test graphs for the composite. In composite S2 large fade was observed after fifth braking
during cold cycle, while in S1 and S3 comparatively quite less
amount of fade was observed after fifth and eighth brakings,
respectively. During fade cycle S2 showed maximum fluctuations in μ which is not a desirable feature for friction materials.
In other two composites, after first few braking cycles, fluctuations reduced drastically as seen in plateau regions. Such
behavior is desirable for good friction material. The constant
speed (660 rpm), constant interval (5 s) and constant pressure
(1.82 MPa) R-90 frictional response of the three composites is
summarized in Table 3. It incorporates the performance μ, the
fade μ, the recovery μ, temperature rise in disc and wear volume of pads. A clear trend was observed in all composites with
the increase of steel wool content in most of the properties.
μperformance , μrecovery , μfade and μmax increased with increasing content of steel wool which a desirable feature. Percent of
fade also showed fixed pattern. With increase in steel wool, tendency of fade decreased. In case of %recovery, however, no fixed
trends were seen. The values did not vary much. Temperature
rise in the disc should be as low as possible and termed as one
of the important parameters called counterface friendliness. In
this case, composite S1 (419 ◦ C) performed better than other two
composites S3 (431 ◦ C) and S2 (456 ◦ C).
Interestingly, wear behavior of composite S1 was also far
superior to the other composite. In literature, general trends show
that the friction and wear performance do not go together. If friction behavior is good, wear performance is poor and vice versa
[11]. However, wear performance is of secondary importance
than the friction performance. Based on above-mentioned discussion it was confirmed that increase in metal contents, led to
Fig. 1. Fade and recovery cycles for composites: (a) S1 , (b) S2 and (c) S3 .
improvement in friction performance (μperformance ) and fade tendency. μperformance improved from 0.371 to 0.407 while %fade
decreased from 18 to 11% (Table 3). Thus, S3 proved to be the
best performer followed by S1 . S1 though proved best in wear
performance, it was poor in friction performance. It was also
confirmed that wear and mechanical properties of the composites
Table 3
ECR 90 frictional response of the composites
Properties
S1
S2
S3
μperformance
μrecovery
μfade
μmin
μmax
%Fade
%Recovery
Wear of pads (×10−6 m3 )
Temperature rise of disc (◦ C)
Ra (original-worn sample) (␮m)
0.371
0.370
0.304
0.250
0.446
18
100
2.5
419
2.723
0.386
0.411
0.329
0.193
0.499
14.7
106
6.7
456
2.755
0.407
0.416
0.361
0.246
0.482
11.4
102
3.4
431
2.815
1246
J. Bijwe, M. Kumar / Wear 263 (2007) 1243–1248
Table 4
Thermal properties of the composites S1 , S2 , S3 and C
Properties
(cm2 /s)
Diffusivity
Specific heat (J/kg K)
Conductivity (W/m K)
S1
S2
S3
C
0.0099
1039.2
2.39
0.0104
1116.2
2.43
0.0107
1233.8
2.51
0.0091
1001.3
2.15
did not show any correlation. Mechanical properties deteriorated
with increase in steel wool while wear did not show any trend
in this context.
Table 4 shows the thermal properties of all composites at
ambient temperature. Data on commercial sample in identical
conditions are also included. As expected thermal conductivity
increased with increasing content of steel wool. For commercial
sample TC value was found 2.15 W/m K. For S1 almost 11%
increase in TC value was observed while for S2 and S3 it was 13
and 16%, respectively. Thus, the increase in steel wool led to the
increase in TC and hence more efficient dissipation of frictional
heat. Higher the dissipation lower was the surface temperature
and hence lower was the degradation of organic contents in the
composite and hence lower tendency of fade. Though this was
anticipated while tailoring the composites, effect of increasing
contents of steel wool on other performance properties could
not be anticipated. Present studies brought out interesting features. The increase in steel contents not only increased the TC
and decreased the fade tendency, but also helped to improve
other friction related parameters. The μperformance , μrecovery and
μfade , increased which are desirable features of the good friction
composites. The difference between μmax and μmin (μ) shows
Fig. 2. SEM micrographs of composites S1 , S2 and S3 .
J. Bijwe, M. Kumar / Wear 263 (2007) 1243–1248
1247
Fig. 3. SEM micrographs of corresponding discs D1 , D2 and D3 .
undulations in friction behavior. This was best for S1 (0.196) and
poorest for S2 (0.306). For S3 , it was moderate (0.236). Increase
in temperature of the disc indicates counterface friendliness of
the materials. It should be as low as possible. Table 3 shows that
S3 was moderate in this respect while S1 was best.
Thus, based on the studies, metallic contents up to 12%
proved to be beneficial from most of the angles. However, the
studies on composites with further increase in the amount of
steel wool were not performed, which could lead to optimization of steel wool contents for best combination of friction and
TC values.
3.2. Microscopy and wear mechanism
Wear performance of the composites was in the order,
S1 > S3 > S2 . The reasons for lowest and highest wear of S1 and
S2 , respectively, could be observed in SEM studies of worn surfaces as shown in Fig. 2. Micrographs 2a1 , 2b1 and 2c1 are for
worn surfaces of composites S1 , S2 and S3 at 500× magnification while 2a2 , 2b2 and 2c2 are for S1 , S2 and S3 at 1000×
magnification. When micrographs 2a1 , 2b1 and 2c1 or 2a2 , 2b2
and 2c2 are compared, surface of S2 showed largest number of
secondary plateaus of very big size. (Micrograph 2b2 shows half
the portion filled with single secondary plateau.) Such plateaus
originate from the back transfer of the degraded organic materials from the disc. They do not have load-bearing capacity and
affect the wear performance adversely. These micrographs also
show the primary plateaus which contain the original ingredients
along with wear debris. These contribute in rejuvenating friction
performance and controlling wear. When the micrographs are
compared, extent of secondary and primary plateaus was lowest
and highest, respectively, in S1 supporting its best wear behavior. In case of S2 , the situation was exactly reverse supporting its
poorest wear performance. For S3 , which showed moderate wear
behavior, the surface topography supported this since it showed
moderate size and number of secondary plateaus. Fig. 3(3a1 , 3b1
and 3c1 ) shows the corresponding worn disc surfaces at 500×
magnification. The surface was heavily covered with thick layer
of friction film. In case of D2 worn against S2 which showed
highest wear, the tribo-layer was thickest and patchy supporting
heavy transfer of organic materials from the composite due to
severe operating conditions. For composite S1 (lowest wear),
on the other hand, tribo-layer was very thin and coherent (D1 ).
Worn surface D3 showed totally different features. Hardly any
patchy transfer was seen on the surface. A thin film of uniform
thickness was observed through out the surface. Surface was
covered with scorching marks (white spots) and not the patches
of organic degraded materials. These are the spots appearing on
the surface due to mild scratching followed by discoloration on
micro-level (not apparent by naked eyes) in hot conditions. An
additional micrograph of D3 (Fig. 3c2 ) at lower magnification
(100×) was also included to show the different features on the
disc.
The friction layer/friction film/tribo-layer/secondary
plateaus terms are used by the researchers to describe the
effect of back transfer of organic materials from the disc to the
pad. This back transferred materials is generally responsible
for the deterioration of friction behavior by reducing the μ,
1248
J. Bijwe, M. Kumar / Wear 263 (2007) 1243–1248
which is undesirable feature. It influences wear behavior also in
adverse way. The secondary plateaus are generated on the pad
surface because of back transfer of organic material, such as
resin (charred, degraded and softened), Aramid fibers, cashew
nut shell liquid (CNSL) powder, graphite, etc. These plateaus
do not have adhesion to the pad material as good as that of
original ingredients. Hence, these do not have good load bearing
properties resulting in deterioration in friction and increase in
wear.
A.3. Heat swelling
Heat Swelling was measured by SAE J 160 JNU80 standard.
Thickness of the sample (10 mm × 10 mm × 4 mm ) from the
brake pad at six different places was measured accurately at
room temperature. The process was repeated after heating the
sample in oven at 200 ± 2 ◦ C for 2 h. The increase in thickness
was recorded as heat swell & percentage was calculated.
Appendix B
4. Conclusions
Definitions of friction coefficient in fade and recovery studies.
Based on studies on three composites containing increased
amount of steel wool (4–12%, w/w) at the cost of inert filler
(barite), it was confirmed that the higher the wool contents, better was the friction performance in fade and recovery mode,
higher was the resistance to fade. However, wear did not show
any correlation with wool contents. SEM studies indicated the
tendency of secondary plateaus formation was maximum in the
composite which showed highest wear and vice versa. Metal
contents (12 wt.%) proved to be most beneficial for tailoring
the tribo-properties. Thermal conductivity played a vital role in
controlling friction and wear behavior. Higher the steel wool
amount, higher was the TC and better was the performance.
Acknowledgements
Authors acknowledge funding by Department of Science and
Technology (Govt. of India) to carry out this work and Mr. B.S.
Tomar (Manager R&D) Allied Nippon Ltd. Sahibabad (India)
for extending the facility of Krauss tester.
Appendix A
μmin : this is lowest coefficient of friction for cold, fade and
recovery cycles.
μmax : this is highest coefficient of friction for cold, fade and
recovery cycles.
μperformance : this is average of coefficient of friction taken after
1 s for fade and recovery cycles at temperature greater than
100 ◦ C. (The requirement of this varies from vehicle to vehicle. For a small sized car, μ in the range 0.4–0.45 is required
while that for racing cars range of 0.5–0.6 is required. For heavier vehicles or medium sized cars, it should be in the range
0.3–0.35.)
μfade : minimum coefficient of friction for the fade cycle taken
after 270 ◦ C.
%Fade: (μperformance − μfade ) × 100/μperformance (lower fade%
is desirable for a good friction composite).
μrecovery : maximum coefficient of friction for the recovery
taken after 100 ◦ C.
%Recovery: μrecovery × 100 /μperformance (desirable 100–120%
for a good friction composite).
Difference in μmax and μmin (μ) should be minimum since
it indicated unsteadiness of braking behavior.
A.1. Method for density measurement
References
It was based on Archimedes principle. The friction composite piece, however, has surface porosity and hence cannot be
dipped in water directly. Hence, it was covered with M-seal
(curing polymer which takes approximately 12 h for complete
curing). This sandwich was then dipped in water and density
was measured.
density(combined) − density(M−seal) × Vf(M−seal)
density =
1 − Vf(M−seal)
where Vf = volume fraction
A.2. Void contents
Void contents give the idea of porosity of the specimen. It
was calculated by the following relation:
%void content =
theoretical density − actual density
× 100
theoretical density
[1] J. Bijwe, Polym. Compos. 18 (3) (1997) 378–396.
[2] F. Dong, F.D. Blum, L.R. Dharani, Polym. Polym. Compos. 4 (3) (1996)
155–161.
[3] P. Gopal, L.R. Dharani, F.D. Blum, Wear 181–183 (1995) 913–921.
[4] T. Kato, A. Magario, Tribol. Trans. 37 (1994) 559–565.
[5] B.K. Friley, B.E. Mcneese, J.T. Trainor, SAE Paper No. 911951, SAE,
1989–1991, pp. 1796–1807.
[6] B.K. Satapathy, J. Bijwe, Wear 257 (5–6 (1–2)) (2004) 573–584.
[7] P.J. Blau, J.C. Mclaughlin, Tribol. Int. 36 (2003) 709–715.
[8] J.M. Herring, SAE Trans., Paper No. 670146, 1967.
[9] B.K. Satapathy, J. Bijwe, J. Reinf. Plast. Compos. 24 (6) (2005) 563–577.
[10] A. Wirth, D. Eggleston, R. Whitaker, Wear 179 (1994) 75–81.
[11] J. Bijwe, Nidhi, N. Majumdar, B.K. Satapathy, Wear 259 (2005)
1068–1078.
[12] B.K. Satapathy, Performance evaluation of non-asbestos fibre reinforced
organic friction materials, Ph.D. Thesis, Indian Institute of Technology
Delhi, India, 2002.
[13] Replacement brake lining assemblies, E.C.E. Regulation No. 90,
INTEREUROPE Regulation Ltd., 1997, UN, 31 March.