You are on page 1of 8


The effect of the composition and microstructure of gray cast

iron on preferential wear during parasitic drag and on intrinsic
damping capacity
Ho Jang, Jang Hyuk Yoon, Seong Jin Kim
Department of Materials Science and Engineering, Korea University

Jae Young Lee, Hyun Dahl Park

R&D Division for Hyundai Motor Company and Kia Motors Corporation

Copyright © 2003 SAE International

ABSTRACT Gray iron brake disks, however, have attracted much

attention since they were considered as a possible cause
Propensity of cold judder was studied by investigating the of brake judder due to uneven disk wear and thermal
correlation between the microstructure of gray iron brake distortion. Brake judder occurs when disk wear results in
disks and friction properties of commercial brake linings. disk thickness variation (DTV) in a circumferential direc-
Based on a brake disk for a mid-size passenger car, gray tion and generates friction force oscillation during brake
iron disks with 6 different microstructures were manu- applications. Thermal distortion of a disk also pro-duces
factured by changing the carbon equivalent (C.E.) and DTV from uneven wear due to non-uniform contact
cooling speed in a commercial manufacturing facility. pressure on the disk surface during sliding. Many efforts
Graphite morphology of the gray iron changed propor- have been made to prevent the DTV increase during
tionally according to the C.E. and cooling speeds, service by controlling a microstructure of gray iron,
exhibiting longer graphite flakes with high C.E. at slow designing a thermally stable disk geometry to avoid
cooling speeds. After screening tests of 23 commercial warping at elevated temperatures, and by relieving
brake linings, 4 different brake linings (two non-steel and residual stress developed during manufacturing [6-9].
two low-steel linings with high µ and low µ) were selected However, the lack of information on the physicochemical
for parasitic drag tests. Results showed that the prefer- process at the sliding interface makes detailed under-
ential disk wear was pronounced in the case of using low standing of uneven wear mechanism difficult.
steel linings and the trend was marked with the disks
containing short graphite flakes. Examination of the disks While the type of wear mechanism is normally classified
after the drag tests revealed that the initial DTV (disk as abrasive wear, adhesive wear, corrosive wear, or
thickness variation) pattern was preserved during the surface fatigue wear, the wear mechanism at the braking
wear tests while the intensity of the DTV pattern was interface is so complicated that it is not possible to
increased. Sonic resonance tests were also performed to approach with one type of wear mechanism. This is main-
investigate the effect of gray iron microstructures on ly because the brake lining is a multi-component com-
vibration damping capacity. Results showed that the gra- posite containing disparate ingredients such as polymers,
phite morphology strongly affected the vibration damping metals, and ceramics. In addition to the complicated
capacity. material system of brake linings, the microstructure of
gray iron, which consists of graphite flakes in the pearlite
matrix, makes the fundamental understanding of wear
mechanism more difficult.
Wear of gray iron disks is dependent on the
Gray iron has been used for brake disks since the early microstructure of gray iron and the tribological properties
stage of brake system development. This is because the of the brake linings. It is known that the type of the
gray iron has good wear resistance with high thermal graphite morphology determines the thermal conductivity
conductivity and the production cost is low compared to of gray iron and they are directly related with carbon
other possible candidates for brake disks such as Al-MMC, equivalent and cooling speed from melt during casting
carbon composites, and ceramic based composites [1-5]. [10,11]. In a previous publication, Cho et al. [12] reported
that the gray iron structure affected the coefficient of
friction and the effect was pronounced in the case of using
brake linings containing steel fibers. However, the influ-
ence of the gray iron microstructure and the ingredients of Wear tests
brake lining on preferential wear of a disk have not been
reported. A wear tester was used to simulate the parasitic drag.
Full size disk was used for the test and two rectangular
In this work, we investigated the propensity of DTV linings were used for the sliding and total area of apparent
increase from the 6 different gray iron disks having contacts was 8 cm2. The DTV during the sliding test was
different microstructures. The disks were manufactured recorded in-situ using non-contacting distance probes
by changing composition and cooling speed during providing data to a PC-based data acquisition system (Fig.
casting. From 23 commercial brake linings, 4 brake lining 1). The normal load was given by pneumatic pressure
specimens were selected concerning the coefficient of and the pressure during the test was set at 7 kgf/cm2 (0.69
friction and the type of a lining. Sliding tests simulating MPa). The test carried out for 48 hrs at 200 rpm (2.56
parasitic drag were carried out using the 6 different gray m/sec) for each set of a friction couple and repeated 3
iron disks and 4 brake linings to examine the effect of the times for reliability of the test results. During the test the
graphite morphology of gray iron disks and lining type on disk temperature was maintained below 100°C.
the preferential wear of gray iron disks.


Gray iron disks

Gray iron brake disks were fabricated using a commercial

manufacturing facility. The brake disk was ventilated type
and was designed for a front brake system of a mid-size
passenger car. The microstructure of the gray iron was
controlled by changing carbon equivalent (C.E.) and
cooling speed of melt during casting. The cooling speed
was controlled by changing the thickness of the cast and
the final microstructure of the disks showed different
graphite morphology. Table 1 shows the C.E. and thick-
ness of the casts used in this experiment. The final thick-
ness of the outboard plate is 9 mm.
Fig. 1. Wear tester used in this work for parasitic prefer-
Table 1. The amount of C.E. and cast thickness (in mm) of
ential wear tests. Two non-contacting distance probes
the outboard side plate used in this experiment
were installed for in-situ measurement of DTV.

Thickness C.E. = 4.6 C.E. = 4.3 C.E. = 4.0

15 mm A B C
11 mm D E F
Gray iron microstructure

Brake linings Fig. 2 shows the microstructure of the 6 gray iron disks.
Gray iron microstructures from different C.E. and cooling
Commercial brake linings were used for the parasitic wear speeds showed a systematic trend in the average graphite
test. Total 23 brake linings were employed (13 non-steel length. The average graphite length was longer when the
linings and 10 low-steel linings) for the screening gray iron contained high C.E. and solidified at slow cooling
procedure. In this work, the coefficient of friction (COF) speeds. Using an image analyzer the maximum graphite
and surface hardness were measured to select can- length and the area fraction of the graphite in the cast iron
didates since aggressive linings against brake disks surface were also obtained (Fig. 3). The quantitative ana-
tended to have high COF values and surface hardness is lysis of the graphite in the gray iron was carried out from
often issued for cold DTV generation. The COF was mea- two different locations; a location near the gate, which was
sured using a small scale friction tester [13]. The friction an inlet for melt during casting practice, and an opposite
tester uses two pieces of linings in the size of 2cm x 2cm, location from the gate. The figure shows that the content
pressed against rotating gray iron disk (φ =12 cm) using a of graphite is higher in the case of containing higher C.E.
hydraulic pressure. The COF for each lining was meas- and at slower cooling speeds. The figure also shows that
ured at 100°C. Surface hardness of the brake linings was graphite morphology is slightly different according to the
measured using a Rockwell Hardness tester in S scale. location of a disk. However, the amounts of deviation acc-
The hardness value was obtained by averaging the values ording to the location of a disk appear within the range of
from 10 points from the surface. measuring error.
0.2 mm 0.2 mm 0.2 mm

(A) (B) (C)

0.2 mm 0.2 mm 0.2 mm

(D) (E) (F)

Fig. 2. Microstructures of gray iron disks obtained by changing carbon equivalent (C.E.) and cooling speed. See the
Table 1 for C.E., casting thickness, and specimen names; A through F.

Physical properties of commercial brake linings

55 20
Graphite length (gate) The COF and surface hardness of 23 commercial brake
Graphite length (far away) 18
Area fraction (gate)
linings were measured (Fig. 4). Fig. 4 shows that low-
Area fraction (far away) steel type linings tend to have higher COF and lower
Area fraction of graphite (%)

surface hardness and vise versa for non-steel linings. We
Max. graphite length (mm)

14 selected 4 different brake linings for disk wear tests (two
from non-steel and two from low-steel linings) that
12 represent the extreme cases considering the COF,
surface hardness, and lining types.
35 10

Preferential wear during parasitic drag

6 Sliding tests under the light pressure condition were

25 performed using the six different brake rotors and four
4 commercial brake linings. Fig. 5 shows the disk thickness
20 as a function of rotation angle measured before and after
the tests. The results suggest that the amount of DTV
15 0
change is strongly affected by the type of the brake lining
F) E) D) C) B) A) and cast iron microstructures. In the case of low-steel
T( T( T( T( T( T(
0 11 0 15 3 11 3 15 6 11 6 15 linings, DTV tended to increase and showed large DTV
4. 4 . 4. 4. 4 . 4 .
CE CE CE CE CE CE increment when the lining was slid against gray iron disks
exhibiting short graphite flakes. On the other hand, non-
Fig. 3. Maximum graphite length and area fraction mea- steel linings showed little DTV change regardless of the
sured from the 6 different gray iron disks (one location microstructure of gray iron disks. Fig. 6 shows the ten-
near gate and one location from opposite to gate). Each dency of DTV change with respect to the microstructure of
data point in the graph was obtained from the average of gray iron and the type of linings.
25 different regions.
Low-steel linings USA 2
0.48 25

Disc thickness (µm)

Non-steel linings Asia 5
0.46 Europe 5 (No.1)
Friction coefficient

Europe 4
0.44 Europe 4

Europe 1 USA 3 USA 4

0.42 Europe 3 Europe 6 15
Europe 2 (No.2)
0.40 USA 1 Asia 6
Asia 1 Asia 7 Asia 8 (No.3)
0.38 Asia 4 USA 5
Asia 9 Initial DTV
0.36 Asia 10 Final DTV
Asia 2 (No.4) Asia 3
USA 6 0
0 180 360 540 720
0.32 Angular position (o )
20 30 40 50 60 70 80 90

Rockwell hardness (S-scale) (b)

Fig. 4. The coefficient of friction and surface hardness of

the 23 commercial brake linings tested in this work. Open
square indicates low-steel linings and open circle
represents non-steel linings. Four different linings (solid 25

Disc thickness (µm)

symbols) were annotated as No.1 – No.4, which were
selected for preferential wear test. 20

The figure clearly indicates that the DTV change is 15

affected by the type of the linings and the graphite
appears beneficial to reduce DTV increase when low-steel 10

linings are used. This is a very interesting result since it

Initial DTV
shed light on the efforts to find a method to prevent DTV 5
Final DTV
increase from parasitic drag. In particular, the fact that
gray iron disks with high content of graphite flakes 0 180 360 540 720
reduced the DTV increase suggests that large amount of Angular position (o )
graphite in the disk can reduce the DTV. The disk surf-
aces were examined using an optical microscope (Fig. 7). (c)
The rubbed gray iron surface with low-steel linings was
shiny due to exposed metallic matrix suggesting
significant surface wear. On the other hand, the disk surf-
ace rubbed with non-steel linings showed remaining 30

machining marks that were produced during fabrication.

In this case, disk surface was covered with friction films 25
Disc thickness (µm)

that were transferred from the linings. The amount of disk

wear was also measured after extended tests using a low-
steel lining (Fig. 8). Fig. 8 clearly suggests that the wear
amount is directly related to the graphite morphology
which can be controlled by the C.E. and cooling speed. 10

Initial DTV
Final DTV
Disc thickness (µm)

0 180 360 540 720
Angular position (o )
15 (d)


Initial DTV Fig. 5. Disk thickness as a function of rotation angle of a

Final DTV disk, measured before and after parasitic drag tests. (a)
and (b) were results tested on disks F and A with low-steel
0 180 360 540 720 lining No.1, respectively, and (c), (d) were results tested
Angular position (o) on disks F and A with low-steel lining No. 2. (continues on
(a) the next page for non-steel linings).
Fig. 5. (Continue) Disk thickness as a function of rotation
30 angle of a disk, measured before and after parasitic drag
tests. (e) and (f) were results tested on disks F and A with
a lining No.3 (non-steel type) and (g), (h) were results
Disc thickness (µm)

tested on a disks F and A with a lining No. 4 (low-steel


10 14

Disk thickness variation (µm)

Graphite length (gate)
Initial DTV Graphite length (far away)
Final DTV 10 Area fraction (gate)
8 Area fraction (far away)
0 180 360 540 720
Angular position (o ) 4

(e) 2
30 0

40 20
Disc thickness (µm)

Area fraction of graphite (%)


Max graphite length (mm)

20 18


10 25

Initial DTV 20
Final DTV

0 15
0 180 360 540 720
Angular position (o )
10 10
(f) CE4.0 CE4.0 CE4.3 CE4.3 CE4.6 CE4.6
11T 15T 11T 15T 11T 15T

(a) Tested with low steel lining No.1

Disc thickness (µm)

Disk thickness variation (µm)

15 12
Graphite length (gate)
Graphite length (far away)
10 Area fraction (gate)
8 Area fraction (far away)

Initial DTV 6
Final DTV

0 2
0 180 360 540 720
o 0
Angular position ()
40 20
(g) Area fraction of graphite (%)
30 35 18
Max graphite length (mm)

Disc thickness (µm)

30 16

25 14

20 12

15 10
Initial DTV
Final DTV
10 8

0 CE4.0 CE4.0 CE4.3 CE4.3 CE4.6 CE4.6

0 180 360 540 720 11T 15T 11T 15T 11T 15T
Angular position (o )
(b) Tested with low steel lining No.2
Disk thickness variation (µm)

Graphite length (gate)
Graphite length (far away)
10 Area fraction (gate)
8 Area fraction (far away)

40 20
0.2 mm

Area fraction of graphite (%)

35 (a)
Max graphite length (mm)






10 10
CE4.0 CE4.0 CE4.3 CE4.3 CE4.6 CE4.6
11T 15T 11T 15T 11T 15T
0.2 mm
(c) Tested with non-steel linings No.1

Disk thickness variation (µm)

Graphite length (gate)
Graphite length (far away)
10 Area fraction (gate)
8 Area fraction (far away)

0.2 mm
40 20

Area fraction of graphite (%)

Max graphite length (mm)






10 10
CE4.0 CE4.0 CE4.3 CE4.3 CE4.6 CE4.6
11T 15T 11T 15T 11T 15T 0.2 mm

(d) Tested with non-steel lining No.2 (d)

Fig. 6. Tendency of DTV change with respect to the

microstructure of gray iron and lining type. The figure Fig. 7. Rubbing surfaces of the gray iron disks A. (a) lining
clearly indicates that the DTV change is affected by the No.1 (low-steel), (b) lining No.2 (low-steel), (c) lining No.3
gray iron microstructure and lining type. (non-steel), (d) lining No.4 (non-steel).
1.0 0.005

Q-1 (Inverse quality factor)

Amount of wear(g)





11 15 11 15 11 15
4.0 4.0 4.3 4.3 4.6 4.6
4.0 11T 4.3 11T 4.5 15T CE CE CE CE CE CE
Fig. 9. Internal friction of the six gray irons.
Fig. 8. Amounts of disk wear after extended wear tests
using low-steel lining No.1.

Vibration damping of gray iron ACKNOWLEDGMENTS

Vibration damping capacity was measured from the gray This works was funded by Hyundai Motor Company.
iron disks with different graphite morphology using a sonic Authors thank for the technical supports from Sungwoo
resonance tester since the length of graphite flakes affects Automotive Co. Ltd. and Korea Flange Co. Ltd. Authors
the vibration damping, which is important in brake perfor- also thank Prof. Youngman Kim of Chonnam National
mance [14]. Damping capacity was measured at room University for the measurement of damping capacity.
temperature and calculated from the internal friction: Q-1.

I 
ln 1  REFERENCES
Q −1 =  2 
πN 1. M. Terhech, R.R. Manory, J.H. Hensler, The Friction
and Wear of Automotive Grey Cast Iron under Dry
where I1 and I2 are amplitudes of vibration intensity after N
Sliding Conditions, Wear 180 (1995) 73-78.
oscillations [15]. Fig. 9 shows the internal friction values
2. J. D. Rainbolt, Effects of Disk Material Selection on
measured from the six gray iron disks. The figure strongly
Disk Brake Rotor Configuration, SAE Technical Paper
suggests that the longer graphite flake is effective to
diminish the vibration. This suggests that the gray iron
disks with longer graphite flakes can absorbed physical 3. M. Kubota, T. Hamabe, Y. Nakazono, M. Fukuda, K
excitation at the friction surface during brake applications. Doi, Development of a Light Brake Disk rotor: a
Design Approach for Achieving an Optimal Thermal,
Vibration and Weight Balance, JSAE Review 21
(2000) 349-355.
CONCLUSION 4. C. Kuylenstierna, T. Storstein, Cost Effective Alumi-
num MMC Brake Discs, SAE Technical Papers 2000-
Parasitic drag tests were performed to investigate the 0102763.
effect of gray iron microstructure and type of brake linings 5. M, Krupka, A. Kienzle, Fiber Reinforced Ceramic Co-
on the preferential wear (DTV generation) of brake disks. mposite for Brake Discs, SAE Technical Paper 2000-
The results show that low-steel linings tended to increase 01-2761.
DTV and not much DTV change was observed when non- 6. A. M. Lang, An Approach to the Solution of Disk Brake
steel linings were used. Another salient feature of the test Vibration Problems, IMechE C37/83 (1983) 223.
result is that the DTV increase from the low-steel linings 7. W. Stringham, P. Tsang, J. Pfeifer, A. Wang, Brake
can be avoided by using a gray iron disks containing high Roughness - Disk Brake Torque Variation, Rotor
graphite contents, although disk wear is significant. Distortion and Vehicle Response, SAE Technical
Damping test results suggest that the gray irons with Paper 930803.
longer graphite flakes are more effective to diminish the 8. A.E. Anderson, R.A. Knapp, Hot Spotting in Automo-
vibrations. tive Friction Systems, Intl. Conf. on Wear of Materials,
Vol.2 (1989) 673-680.
9. K. Lee, J.R. Barber, Friction Excited Thermoelastic 13. S.J. Kim, M.H. Cho, D.-S. Lim, H. Jang, Synergistic
Instability in Automotive Disk Brakes, ASME J. of Effect of Aramid Pulp and Potassium Titanate Whis-
Tribology 115 (1993) 607-614. kers in the Automotive Friction Material, Wear 251
10. R.L. Hecht, R.B. Dinwiddie, W.D. Porter, H. Wang, (2001) 1484-1491.
Thermal Transport Properties of Grey Cast Iron, SAE 14. A.R. Daudi, W. Dickerson, Ultra Q Process, SAE
Technical Paper 962126. Technical Paper 200-01-2760 (2000) 59-65.
11. T. Okamoto, A Kagawa, K. Kamei, H. Matsumoto, Eff- 15. K.H. Cho, Y.M. Kim, Internal Friction Change of Al,
ects of Graphite Shape on Thermal Conductivity and Mg-alloys in Terms of Casting Method, J. Korea. Inst.
Young’s Modulus of Cast Iron below 500°C, J of the Met. & Mater. 38 (2000) 253-259.
Japan Foundaryman’s Society 55 (1983) 32-36.
12. M.H. Cho, S.J. Kim, R.H. Basch, J.W. Fash, H. Jang,
Tribological Study of Gray Cast Iron with Automotive
Brake Linings: The Effect of Rotor Microstructure, CONTACT: Prof. Ho Jang, Korea University, e-mail:
Tribology International 36 (2003) 537-545.