Wear 268 (2010) 784–789

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Tribological study of Fe–Cu–Cr–graphite alloy and cast iron railway brake shoes
by pin-on-disc technique
C. Ferrer, M. Pascual, D. Busquets, E. Rayón ∗
Instituto de Tecnología de Materiales, Universidad Politécnica de Valencia, Edif. 5E, Camí de Vera s/n, E-46022, Valencia, Spain

a r t i c l e i n f o a b s t r a c t

Article history: A new class of materials is being installed in railway brake blocks to substitute classic cast iron in order to
Received 2 September 2008 reduce the rolling noise produced by the roughness of the tread-wheel surface. The tribological properties
Received in revised form of cast iron and Fe–Cu–Cr–graphite sintered alloy brake shoes were analyzed. Kinetic friction coefficient
25 November 2009
() and wear were monitored by means of a pin-on-disc technique. The sintered alloy brake showed an
Accepted 7 December 2009
increase in  at higher braking velocities while the cast iron brake exhibited a decrease in . Wear was
Available online 30 December 2009
greater on the sintered alloy, explained by its low shear strength which decreased due to its low thermal
conductivity. The roughness produced by the sintered brake shoes in wheel-tread surface was 10 times
Keywords:
Sliding friction
lower than that produced by cast iron.
Surface topography © 2009 Elsevier B.V. All rights reserved.
Cast iron
Sintered alloy
Brakes
Pin-on-disc

1. Introduction cast iron brake shoes make the wheel surface much rougher than
a similar product made with a composite material. Replacing the
Noise elimination is an issue that is gaining importance on rail- cast iron shoes with a synthetic product can therefore substantially
ways. At present, an estimate of one million people in Europe has reduce the wagon noise emissions, for example, by about 10 dB
to be protected from railway noise, by noise barriers and by noise- for a 100 km/h freight train [6]. Nevertheless, previous experiences
insulated buildings [1,2]. One of the main sources of railway noise using a polymeric composite brake demonstrated that the organic
is rolling noise arising from the contact between the wheels and material can be burned, coating the wheel-tread with an organic
the tracks. Rolling noise is the dominant source of noise at speeds film. This film diminishes the friction between the wheel tread and
between 60 km/h and 200–250 km/h [3,4]. Rough wheels and rough rail or brake shoe [7,8]. Other sintered alloys have been used as
tracks increase the noise emission. Thus, the rougher the wheel sur- an Fe–Cu–Cr–Sn–graphite alloy [9]. These brake shoes exhibited a
face, the greater the noise produced. Other sources of noise such as better braking behavior with less wear on the tread wheel. How-
aerodynamic friction are only important when high velocities are ever, premature wear has been observed in this material, being less
reached (more than 300 km/h). cost-effective.
The simplest way to reduce the velocity of a railway in motion is The aim of this work is to study and compare the tribologi-
through the friction produced between a brake shoe and the wag- cal properties of a gray cast iron brake and a composite sintered
ons’ wheel tread. Cast iron (usually as a gray cast iron) has been alloy brake (Fe–Cu–Cr–Sn–graphite). A pin-on-disc technique is
the most used material in brake shoes [5]. Cast iron brake shoes proposed to analyze comparatively the friction coefficient of both
have the disadvantage of making the wheel-tread surface rougher materials. Other mechanical analysis and photothermal techniques
during braking [6]. Thus, a reduction in noise should be achievable were performed to discuss the wear phenomena.
through the substitution of cast iron brake shoes by other material
leading to lower wheel surface roughness.
2. Experimental set-up
In fact, over the last years, cast iron shoes have been replaced
by composite synthetic brake shoes. It has been demonstrated that
The samples used in this work were extracted directly from the
commercial brake shoes of a locomotive train. Microstructure and
chemical composition were analyzed by standard metallographic
∗ Corresponding author. Tel.: +34 667238511; fax: +34 963877629. and energy dispersive X-ray analysis (EDX) on a SEM respectively.
E-mail address: emraen@upvnet.upv.es (E. Rayón). The sintered alloy sample showed a composition of 38% Cu, 34%

0043-1648/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2009.12.014
 C. Ferrer et al. / Wear 268 (2010) 784–789 785

Fe, 3% Cr, 15% C, 4% Sn, 4% O and 2% Al. Graphite, bronze and ␣- value of Ra = 0.015 ␮m was achieved on both sides. One side of the
alumina were detected. The cast iron brake shoe showed a gray plate was used for the cast iron pin test and the other side was used
cast iron structure with graphite sheets in a perlite matrix. Stan- to perform the sintered alloy pin test.
dard compression testing showed a stress limit at 350 MPa, which The parameters selected for our experiments were loads of
corresponds to EN-GJL-300 standard. The Brinell hardness value 10 N (0.2 MPa) and 20 N (0.4 MPa) and velocities of 0.31 ms−1 and
measured for this material was 223 HB. On the other hand, a Brinell 1.05 ms−1 .
hardness value of 39 HB and compressive strength of 45 MPa was
obtained for the sintered alloy. All measures were carried out at
2.2. Thermal analysis
room temperature.
The thermal study of the brake materials is interesting from
2.1. Friction testing the point of view of the power dissipation capacity of each
material. Thermal studies for both materials were performed
Although pin-on-disc technique does not describe the contact using a Flir Systems ThermaCAMTM S65 infrared camera. Again,
geometry, applied normal pressure and asperity distribution of samples were extracted from a commercial brake shoe. Pieces
wheel-tread, this technique is useful in the comparative tribological 11 mm × 11 mm × 40 mm in size were then installed on a cylinder
study of both materials. In fact, this technique allows the measure- with large mass made of steel at 25 ◦ C. With the large mass of the
ment of the kinetic friction and wear when a plate (the disc) of cylinder, we assumed the break’s holder to have good thermal dis-
one material rolls in contact with a sample (the pin) of another sipation. A cylindrical steel piece preheated to 500 ◦ C as a means to
material which can provide valuable information on the tribologi- simulate the heat released during breaking was placed on top of the
cal behavior of both systems. Therefore, the aim of this study is the sample at the same moment that the images began to be recorded
comparative study of two materials rather than replication of a real every 5 s. The initial temperature conditions were the same for all
braking condition. of the samples and experiments.
The test apparatus and testing technique is described as follows.
MicroTest® pin-on-disc equipment was used to measure the wear
3. Results
and friction coefficient between the brake sample (cast iron or sin-
tered alloy pin) and the steel plate (as the train wheel). A load
3.1. Tribological results
of between 10 N and 500 N and angular velocities ranging from
10 rpm to 500 rpm can be programmed by a computer. The fric-
The friction coefficient  were recorded for both brake materials
tion is then calculated and recorded in the computer by means of
sliding above the steel plate. Fig. 2A shows the  for cast iron and
several sensors installed on the rotary’s axis plate.
Fig. 2B shows the  for sintered alloy. Cast iron curves showed that
Fig. 1 shows the Microtest® pin-on-disc used. The following can
 increased during the experiments regardless of the load condi-
be distinguished in the photograph: the plate (1), which moved
tions and velocities studied, unlike sintered alloy, which exhibited
the wheel sample (2) against the pin (3). Weights (4), which apply
a more stable coefficient.
the contact force and a magnetic sensor (5) to measure the vertical
Braking pressures of 0.2 MPa and 0.4 MPa and velocities of
movement of the arm, thus obtaining the wear of the pin.
0.31 ms−1 and 1.03 ms−1 were tested. The average friction coef-
The wheel sample was composed of the same material as the
ficients for these conditions are summarized in Fig. 3 for clearer
steel wheel railway with a composition of: 0.56% C, 0.4% Si, 0.8%
presentation.
Mn, 0.035% P, 0.035% S, 0.3% Cr, 0.3% Cu, 0.08% Mo, 0.3% Ni, 0.05%
The cast iron results showed that when the braking pressure
V and 0.4% Ti. The plate was polished until an average roughness
was increased from 0.2 MPa to 0.4 MPa, the friction coefficients
decreased from 0.55 to 0.25 at 1.03 ms−1 and from 0.7 to 0.4 at
0.3 ms−1 . Moreover, when the velocity was increased from 0.3 ms−1
to 1.03 ms−1 , the friction coefficient decreased from 0.7 to 0.55
under pressures of 0.2 MPa and from 0.4 to 0.25 under pressures
of 0.4 MPa. Thus, the cast iron brake shoe sample presented two
specific characteristics: (i) the higher the speed, the lower the
friction coefficient and (ii) the higher the contact force of the
brake shoe on the wheel, the lower the friction coefficient. These
results agree with previously published reports [10,11]. In fact,
both of these characteristics led to specific brake system designs
for vehicles braked with cast iron shoes. This included reducing
the brake cylinder pressure at low speeds to prevent the wheels
from locking, and equipping vehicles with twin or even triple brake
blocks to prevent the specific application force from becoming
too great and, as a result, the friction coefficient from becoming
too low.
Kinetic friction results for sintered alloy showed a very stable
 coefficient along the 500 m distance tested. Average  was 80%
higher than that of the cast iron results. The  behavior under the
pressures tested was similar to that of the cast iron: when the pres-
sure was increased from 0.2 MPa to 0.4 MPa,  decreased from 0.9 to
0.7 at 0.3 ms−1 and from 1.3 to 1.1 at 1 ms−1 . Nevertheless, at higher
Fig. 1. Photography of Microtest® pin-on-disc technique used. (1) Disc which rolled velocities  was higher: from 0.3 ms−1 to 1.03 ms−1 ,  increased
at programmed velocity. (2) Steel disc sample extracted from locomotive wheel. (3)
from 0.9 to 1.3 under 0.2 MPa and from 0.7 to 1.1 under 0.4 MPa.
Sintered alloy or cast iron pin extracted from locomotive brake shoes. (4) Weights
used to apply the braking contact force and (5) magnetic sensor to measure the
These results must be taken into account when designing the brake
wear. system in railway transport because, in this way, equivalent speeds
786 C. Ferrer et al. / Wear 268 (2010) 784–789

Fig. 2. Kinetic friction coefficients () at velocities of 0.3 ms−1 /1 ms−1 and under 0.2 MPa/0.4 MPa contact pressures. (A) Cast iron results and (B) sintered alloy
Fe–Cu–Cr–graphite results.

must produce equivalent stopping distances. The rules which drive constant behavior after testing 100 m. Temperature was also con-
these behaviors will be discussed further below. stant and reached 110 ◦ C. However, in the sintered alloy sample,
During friction analysis, wear and temperature were also mea- the temperature increased very quickly after 100 m and the wear
sured and recorded. Wear was measured as vertical movement of process developed exponentially when it reached 170 ◦ C. However,
the tribometer arm as a result of groove formation and thus repre- one must bear in mind that temperatures measured are not nec-
sented in millimeters. The temperature reached in the sample was essarily the same on the contact surface. After 500 m of testing for
monitored by a thermocouple installed 10 mm from the sample the sintered alloy, the wear level reached 4.5 mm with respect to
contact surface. Although it was impossible to measure the tem- the 2.25 mm measured for the cast iron.
perature of the surface friction with a thermocouple, temperature In order to establish the roughness generated on the wheel sur-
curves gave an idea of the moment in which the power generated face by each shoe material, the plate was initially polished down to
by the friction changed and therefore, also provided the mechanism 0.145 ␮m (Ra ), as measured by a perthometer M2 roughness tester.
which governed the friction and wear process. One side was used as sliding contact with the cast iron and the
No wear was detected at a sliding velocity of 0.3 ms−1 for both opposite side with the sintered alloy sample. The resulting rough-
pressures and materials. In these cases the noise level recorded was ness produced by wear was measured after all of the experiments
very high compared to the wear signal. In addition, it was difficult for each material. Results showed a plate tread surface roughness of
to detect wear by the loss of weight method after the experiments 1.635 ␮m for cast iron. However, the side used in contact with the
made at 0.3 ms−1 due to the low wear produced. It was only possi- sintered alloy practically did not show any changes with respect
ble to monitor a clear wear when a velocity of 1 ms−1 was tested. to the parent roughness, with a measured average roughness of
These observations indicated that sliding velocity was determinant 0.155 ␮m. Although Ra gives no indication of change in surface
in the wear process. Fig. 4 shows the wear and temperature curves topography (which is expected to have taken place), as differences
recorded for both shoe materials when the plate was rolling at in roughness of one order of magnitude have been measured the
1 ms−1 . Wear in the cast iron sample followed a linear and thus a higher roughness produced with using cast iron pads as mentioned

Fig. 3. Coefficient of friction () versus braking pressure at 1 ms−1 and 0.3 ms−1 for Fig. 4. Wear and temperature results for cast iron and a sintered alloy versus dis-
the cast iron and sintered alloy pin-on-disc results. tance tested.
 C. Ferrer et al. / Wear 268 (2010) 784–789 787

Fig. 5. Thermographic images captured by infrared camera. (A) Sintered alloy sample. (B) Sample of a cast iron. The heating was performed from the top side by a preheated
steel mass at 500 ◦ C. The dimensions of the sample were 11 cm × 11 cm × 40 cm.

before is corroborated. Sintered alloy, softer and with a graphite ficult to observe a clear image. The colored background was due to
flakes that act as lubricant, made less erosion on the tread sur- the irradiation heat flux of the heated sample. Images showed that
face, maintaining a high friction coefficient value. This decrease in the map of temperatures was very similar during the first 45 s of
roughness on the tread surface will translate into a lower rolling the heating process. Only after 65 s did the cast iron sample show
noise. a higher (more than 210 ◦ C) in-depth heating. Sintered alloy was
colder in-depth for the same time reached.
3.2. Thermal results In fact, the thermal coefficient of conductivity of cast iron varies
between 58 and 47 (W m−1 K−1 ) from 100 ◦ C to 500 ◦ C. That of the
Friction dissipates energy in the form of heat. The combination sintered alloy can be approximated by considering the different
of braking effort through friction between the brake shoe and the constituents and the presence of porosity. In fact, sintered alloy has
wheel-tread surface produces thermal and contact stresses on the a high amount of porosity and therefore reduces the effective cross-
wheel, which often result in wheel damage. High thermal and con- section of the sample, and therefore heat is transmitted through
tact stresses are known to initiate microcracks in the wheel tread, the neck between the sintered particles. A complex mechanism of
eventually leading to a defect condition known as shelling. In addi- convection heat transmission occurs in the pores, but with lower
tion, the heat accumulated in the break shoe decreases the shear heat transmission. In addition, the heterogeneous microcomposi-
strength, causing premature wear due to delamination of the brake tion diminished the thermal conductivity. The measured density
surface. Around 25% to 30% of the friction heat produced during values of the cast iron was c = 7.71 g cm−3 , while sintered alloy had
braking is currently dissipated by the brake shoe [12]; the rest by a s = 6.05 g cm−3 . Considering the thermal conductivity of the dif-
the wheel. The wheels must therefore be able to deal with this ferent constituents of the sintered alloy, their proportions and the
greater proportion of energy. For all of these reasons, the thermal relative density (to take porosity into account) and following calcu-
dissipation produced in the braking material must be analyzed. lations as from Ref. [13] a thermal conductivity variation of about
Two types of thermal measurements were performed, (i) by 73–43 (W m−1 K−1 ) again from 100 ◦ C to 500 ◦ C for the sintered
means of a thermocouple inserted in the sample during pin-on-disc alloy is obtained. Thus, lower temperature dissipation at higher
testing and (ii) by means of an infrared camera when the sample temperatures could be expected for the alloy as measured by the
was heated with a preheated mass. In the first case, the temperature thermal experiments by infrared imaging.
observed (Fig. 4) when the wheel was sliding on the brake sample Furthermore, in order to study the effect of temperature on the
reached 110 ◦ C in the gray cast iron at the maximum velocity and shear strength value, a standard shear test procedure was carried
pressure tested (1 ms−1 and 0.4 MPa). Under the same conditions, out at room temperature and at 500 ◦ C. Shear strength of a material
the sintered alloy reached 190 ◦ C. At lower velocities and forces, the may be an estimation of the probability of a possible failure when
temperature did not reach 50 ◦ C for any material. These results gave this material is working under shear loads, as in the case of shoe
a first indication that velocity is the most influential parameter on brake. The maximum friction force value would be reached at the
the maximum temperature reached. Thermal conductivity of the shear strength value of the material. Experiment results showed
alloy was lower than that of the cast iron as shown in Fig. 5 by an a sintered alloy shear strength of  s = 35 MPa at 25 ◦ C, whereas at
infrared image sequence of (A) sintered alloy and (B) gray cast iron 500 ◦ C,  s decreased to 17 MPa. However, gray cast iron showed
samples when a preheated steel mass was installed on top of the 160 MPa at 25 ◦ C and 100 MPa at 500 ◦ C. From these results, it can
samples. Initial conditions have not been represented due to the be predicted that sintered alloy will wear by shear cutting before
homogeneous blue color of the image at 35 ◦ C, which made it dif- cast iron when high pressures and temperatures are reached.
788 C. Ferrer et al. / Wear 268 (2010) 784–789

Table 1
Flash temperatures as calculated from Archard’s equations.

V = 0.3 m/s V = 1 m/s

F = 10 N F = 20 N F = 10 N F = 20 N

Gray cast iron 121 ◦ C 98 ◦ C 375 ◦ C 273 ◦ C

Sintered alloy 58 ◦ C 73 ◦ C 350 ◦ C 456 ◦ C

calculated temperatures refer to the flash temperatures at the irreg-

ularities in the contact area whereas measured temperatures are
about 10 mm away of this area in the pin.
A wide range of experiments must be performed to establish the
critical force at which plastic deformation occurs, to determine the
point at which  begins to decrease in brake shoes. This should be
matter of further research work.
Fig. 6. Cushioned function of the friction force when an increase in braking force is
applied to the contact surface and , result of this function. The average increase of  at higher velocities in the sintered
alloy is thought to be due to the cohesive behavior of Cu rich phase.
When temperature increases on the contact surface at high speeds,
4. Discussion
as inferred both by the measured temperatures and the flash tem-
peratures worked out, this phase is more cohesive with the wheel.
Cast iron kinetic friction coefficient curves showed that lower
At low temperatures, graphite is responsible for maintaining the
friction was measured for higher speeds and forces. Also, for this
lubrication to prevent welding between the brake and wheel. How-
material, friction depended on the brake distance. However, the
ever, when graphite is released by increased temperature, due to
brake wear was higher in the sintered alloy. A  loss measured
frictional power, lubrication diminishes, increasing  and temper-
when the force was increased, was observed in both shoe materials.
ature. Heat is not well dissipated through the shoe as shown in the
Our experience in railway brake systems showed us that, in the
infrared images. Thus, the shear strength limit of the brake surface
range of pressures applied, friction force, Fc, shows a linear increase
decreases. As the shear strength limit is not very high for this mate-
until a value at which the curve is cushioned as shown in Fig. 6.
rial ( s = 35 MPa at 500 ◦ C) metal is then released and wear in the
Under low contact pressures, the coefficient of friction is constant
brake pad is produced. This is the suspected reason for which sin-
(Pb < Pb 0), and follows Eq. (1):
tered alloy presented a high level of wear at high velocities. Indeed,
Fb cast iron, which presented high shear strength values and good
= = tg˛ (1)
F thermal conductivity, showed less wear.
Regarding the reduced friction at higher sliding velocities in the
However, when the pressure reaches Pb 0, a cushioned behavior is
case of gray cast iron, this behavior can be explained in terms of
observed. At this pressure range:
reduction of the ploughing term of friction [16–24] for increased
Fb = Ci · Panp · S (2) velocities.
On the other hand, the higher wheel roughness observed when
np being a coefficient of pressure with a value less than one. Ci is a the cast iron sample was tested can be explained by the scratching
constant of friction which depends on the topography of the sur- of hard particles detached from the brake contact area. Usually,
faces in contact. Thus, the friction coefficient is a cushioned function gray cast iron brake shoes contain titanium carbide particles in
guided by: order to reduce wear [25]. These hard particles, together with other
Fb Ci · Pbnp · S hard released material, are the cause of the major erosion pro-
= = = Ci · Pb −(1−np) (3) duced in the wheel, and are therefore the cause of major rolling
F Pb · S
noise.
The mechanisms occurring on the contact area when a cush-
ioned function is observed are due to the plastic deformation of
the surface. The crests which form the roughness are deformed, 5. Conclusions
thus changing the real contact surface. In fact, calculation of plas-
ticity indexes as proposed by Greenwood and Williamson [14] Pin-on-disc has been a useful technique for the measurement of
yielded results well above the lower limit for considering plastic the coefficient friction, , for one brake shoe made with cast iron
behavior. In order to calculate the local temperature increase in and another made with a sintered copper alloy in contact with a
the irregularities, Archard’s approach is followed [15]. According standard railway wheel material. Although technique did not rep-
to this, to calculate flash temperature of the sliding pairs, apart resent the actual conditions of the wheel-brake systems, the results
to know whether the materials behave elastically or plastic, slow obtained showed interesting different behavior of both materials
and fast moving contacts have to be distinguished. For this pur- studied. These differences were resumed as:
pose a non-dimensional parameter L considering speed, contact
radius and properties of the materials is calculated. Unfortunately 1.- Sintered alloy has a constant and 80% higher friction coefficient
for the parameters and materials used in this work the results fall compared to the cast iron shoe.
in between both regimes (L values vary from 0.32 to 1.55 for both 2.- Friction coefficient decreases with sliding friction velocity in
materials, being in between the maximum value of 0.1 to be con- cast iron shoes but increases this value in the sintered alloy.
sidered low speed, and the minimum value of 5 to be considered This behavior must be taken into account when high velocity
high speed. Nevertheless the temperatures are calculated consid- train designs are being considered.
ering low speed and plastic conditions as a first approximation. The 3.- The premature wear of sintered alloy occurred at high veloci-
results are presented in Table 1: ties due to the decrease in shear strength, an effect increased
These results are well in agreement with the temperatures mea- by the low thermal conductivity. However, the softer particles
sured by the thermocouple, taking into account that these the abraded from the sintered shoe produced a lower level of rough-
 C. Ferrer et al. / Wear 268 (2010) 784–789 789

ness on the wheel in comparison to cast iron, therefore being a [9] X. Xiang, S. Hong-chao, C. Jie, Y. Ping-ping, Effects of sintering pressure and tem-
better material in the policy to reduce the noise of the railway perature on microstructure and tribological characteristic of Cu-based aircraft
brake material, Trans. Nonferrous Met. Soc. China 7 (2007) 669–675.
transport versus a higher wear. [10] I. Ripley Maurice, Oliver Kirstein, Residual stresses in a cast iron automotive
brake disc rotor, Physica B 385–386 (2006) 604–606.
Acknowledgements [11] G. Cueva, A. Sinatora, W.L. Guesser, A.P. Tschiptschin, Wear resistance of cast
irons used in brake disc rotors, Wear 255 (2003) 1256–1260.
[12] D. Bettge, J. Starcevic, Topographic properties of the contact zones of wear
We would like to thank the R&D&I Linguistic Assistance Office, surfaces in disc brakes, Wear 254 (2003) 195–202.
Universidad Politécnica de Valencia (Spain), for translating this [13] L.J. Gibson, M.F. Ashby, Cellular Solids, 2nd edition, Cambridge University Press,
1997, pp. 284-292.
paper. [14] J.A. Greenwood, J.B.P. Williamson, Contact of nominally flat surfaces, Proc. Roy.
Soc. Lond., Ser. A 295 (1966) 300–319.
[15] J.F. Archard, Wear 2 (6) (1959) 438–455.
References [16] D. Tabor, Tribol. Int. 28 (1) (1995) 7–10.
[17] F.P. Bowden, D. Tabor, The Friction and Lubrication of Solids, Clarendon Press,
[1] A. Lundström, M. Jäcker-Cüppers, P. Hübner, The new policy of the Euro- Oxford, 1953, pp. 98–100.
pean Commission for the abatement of railway noise, J. Sound Vib. 267 (2003) [18] E. Orowan, The calculation of roll pressure in hot and cold flat rolling, Proc. Inst.
397–405. Mech. Eng. 150 (1941) 140–167.
[2] P. De Vos, How the money machine may help to reduce railway noise in Europe, [19] C.M. Edwards, J. Halling, An analysis of the plastic interaction of surface asper-
J. Sound Vib. 267 (2003) 439–445. ities and its relevance to the value of the coefficient of friction, J. Mech. Phys.
[3] Tomaz Sedmak, Bojan Rosi, Vesna Smaka Kincl, Basic tools for controlling rolling Solids 10 (1968) 101–110.
noise in railway traffic, Logist. Sust. Transport 1 (2) (2001). [20] J. McFarlane, Tabor, Relation between friction and adhesion, Proc. Roy. Soc.
[4] M. Kalivoda, U. Danneskiold-Samsøe, F. Krüger, B. Barsikow, J. Sound Vib. 267 202A (1950) 244–253.
(2003) 387–396. [21] D. Tabor, Junction growth in metallic friction: the role of combined stresses and
[5] M. Macnaughta, Cast iron brake discs, a brief history of their development and surface contamination, Proc. Roy. Soc. 251A (1958) 378–393.
metallurgy, Technical Report, Foundryman, 1998, p. 321. [22] A.P. Green, The plastic yielding of metal junctions due to combined shear and
[6] S. Bühler, Methods and results of field testing of a retrofitted freight train with pressure, J. Mech. Phys. Solids 2 (1954) 197–211.
composite brake blocks, J. Sound Vib. 293 (2006) 1041–1050. [23] A.P. Green, Friction between unlubricated metals: a theorical analysis of the
[7] W. Österle, I. Urban, Friction layers and friction films on PMC brake pads, Wear junction model, Proc. Roy. Soc. 228A (1955) 191–204.
257 (2004) 215–226. [24] B. Bhushan, Principles and Applications of Tribology, 1999, 978-0-471-59407-9.
[8] P. Filip, Z. Weiss, D. Rafaja, On friction layer formation in polymer matrix com- [25] B.J. Chapman, G. Mannion, Titanium-bearing cast iron for automotive braking
posite material for brake application, Wear 252 (2002) 189–198. applications, J. Foundry Trade 152 (3732) (1982) 232–234.