100% found this document useful (1 vote)
131 views17 pages

Towards A Better Understanding of The Reaction Between Metal Powders and The Solid Lubricant Sb2S3 in A..

This document discusses reactions between metal powders (Cu, Fe, Sn, Zn) and the solid lubricant Sb2S3 that occur in low-metallic brake pads during high-temperature braking. Thermal treatments of mixtures containing these metals and Sb2S3 were analyzed using techniques like DSC-TGA and XRD. The results found that the reaction products and features depend on temperature, material combination, and atmosphere. In general, the products are a combination of an intermetallic (metallic antimonide) and a metal sulphide, found both on the surface and inside the brake pad. Understanding these reactions could help develop alternative lubricants to replace Sb2S3 due to

Uploaded by

DianiTz MendOza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
100% found this document useful (1 vote)
131 views17 pages

Towards A Better Understanding of The Reaction Between Metal Powders and The Solid Lubricant Sb2S3 in A..

This document discusses reactions between metal powders (Cu, Fe, Sn, Zn) and the solid lubricant Sb2S3 that occur in low-metallic brake pads during high-temperature braking. Thermal treatments of mixtures containing these metals and Sb2S3 were analyzed using techniques like DSC-TGA and XRD. The results found that the reaction products and features depend on temperature, material combination, and atmosphere. In general, the products are a combination of an intermetallic (metallic antimonide) and a metal sulphide, found both on the surface and inside the brake pad. Understanding these reactions could help develop alternative lubricants to replace Sb2S3 due to

Uploaded by

DianiTz MendOza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/285216651

Towards a better understanding of the reaction


between metal powders and the solid lubricant
Sb2S3 in a...
Article November 2015
DOI: 10.1016/j.wear.2015.11.014

CITATIONS

READS

18

2 authors, including:
AneMaite Martinez
Ecole Centrale de Lyon
12 PUBLICATIONS 0 CITATIONS
SEE PROFILE

Available from: AneMaite Martinez


Retrieved on: 11 November 2016

Wear 348-349 (2016) 2742

Contents lists available at ScienceDirect

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

Towards a better understanding of the reaction between metal


powders and the solid lubricant Sb2S3 in a low-metallic brake pad
at high temperature
Ane Maite Martinez, Jon Echeberria n
CEIT and TECNUN (University of Navarra), 20018 San Sebastian, Spain

art ic l e i nf o

a b s t r a c t

Article history:
Received 24 June 2015
Received in revised form
2 November 2015
Accepted 5 November 2015
Available online 29 November 2015

During braking, high temperatures can be generated on the contact surfaces and inside the brake pads.
Solid lubricants, i.e., metal sulphides, play an important role in brake performance, but they can react,
oxidise or decompose with exposure to temperature, modifying their tribological properties. This paper
investigates the chemical reactions between different metal powders (Cu, Fe, Sn and Zn) and Sb2S3 that
take place in a low-metallic brake pad during braking in conditions were high temperature is generated.
The reaction products were characterised by SEMEDS and the glancing XRD technique. Moreover,
thermal treatments (in air and argon) of binary mixtures, containing the same weight ratio of these metal
powders and Sb2S3 contained in the brake pad, were carried out and analysed via DSCTGA and XRD
techniques.
The paper nds that the reaction features and reaction products depend on the temperature generated during braking, the specic combination of metal/lubricant and the atmosphere conditions
(oxidant or inert). In general, the reaction products are a combination of an intermetallic (metallic
antimonide) and a metal sulphide (from the metal powder), which can be found both on the surface and
inside the brake pad.
& 2015 Elsevier B.V. All rights reserved.

Keywords:
Friction
Solid lubricant
Low-metallic brake pad
Transfer layer
DSCTGA
Glancing XRD

1. Introduction
In order to achieve the desired friction and wear performance,
automotive brake friction materials are complex composites consisting of a combination of more than 10 ingredients [1]. These
ingredients are frequently classied into four core classes: binder
resins, friction modiers, reinforcements and llers [2]. The type
and amount of these ingredients are determined mostly based on
experience, empirical observations [1,3,4] or a trial and error
method when making new formulations [4]. Among the friction
modiers used to produce brake pads solid lubricants play crucial
role in reducing the wear of both pad and disc, optimising friction
levels and controlling brake-induced vibrations [5].
While the temperature of the brake pad and disc increases
during braking due to the transformation of the kinetic energy of
the rotor into thermal energy, the degradation of the phenolic
resin takes place, leading to a dramatic decrease in the friction
coefcient and the increase in wear, the so-called fade (loss of
brake effectiveness at high temperature due to frictional heat). At
n

Corresponding author.
E-mail address: [email protected] (J. Echeberria).

http://dx.doi.org/10.1016/j.wear.2015.11.014
0043-1648/& 2015 Elsevier B.V. All rights reserved.

this point, where high temperatures can reach  600 C, selecting


a proper solid lubricant becomes an important issue since they
decompose at elevated temperatures, resulting in unexpected
friction characteristics and therefore affecting the maintenance of
the friction lm at the friction zone. In fact, in order to achieve
friction stability for a wide range of temperatures, commercial
friction materials often contain two or more solid lubricants,
anticipating that each one will act as an effective solid lubricant in
different temperature ranges [6]. Indeed, the effectiveness of the
solid lubricants is strongly affected by temperature, pressure,
speed, environmental condition, etc. Therefore, more than one
lubricant is often used since they play complementary roles in
brake performance in various braking conditions.
Solid lubricants are added in relatively small amounts because
their type and amount could strongly affect aspects of brake performance such as wear resistance, stopping distance or friction
stability and torque variation [4,7,8]. Typical solid lubricants used
for commercial friction materials include graphite, metal sulphides
such as antimony trisulphide, tin sulphide, molybdenum disulphide, or soft metals. These solid lubricants should form a
tenacious lm that allows for constant regeneration and protection of the surfaces from excessive wear and reduce the noise and
vibration induced by stick-slip at the friction interface. Numerous

28

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

benecial properties of metallic sulphides have been reported by a


number of groups and they all arrive at the same conclusion
regarding the function of metallic sulphides as a lubricating agent:
metallic sulphides must be chemically active in the tribosystem to
achieve the desired results [9].
Among the solid lubricants used for commercial products,
Sb2S3, which has a melting point of 550 C and a hardness of 2 on
the Mohs scale, has been used to improve fade resistance and wear
resistance in friction material for many decades. Its effect on
friction-wear performance is described in number of research
works. The improvement of fade resistance by adding Sb2S3 was
attributed to the presence of Sb2O3 at the sliding interface, which
was observed to act as a high-temperature solid lubricant in [6]
and [10]. However, in a recent work [11] the expected stabilisation
of m due to the formation of antimony oxides as reported in [6] was
not observed in a semi-metallic brake pad. On the other hand, in
[8] it was observed that a brake NAO lining with a mixture of
graphite and Sb2S3 as a solid lubricant had better friction stability
due to the complementary role of the two lubricants than a formulation with just one of the lubricants. Finally, in a subsequent
research work, Lee et al. [5] investigated the effect of Sb2S3 on the
fade resistance and stick-slip propensity of brake friction materials, and they concluded that it improved fade resistance by causing
a transfer lm to develop on the disc surface, which sustained
friction levels at elevated temperature. They attributed the
improvement in fade resistance to the lubricating capability of
Sb2S3 at high temperature, which was more effective than the fade
resistance obtained with graphite as a solid lubricant. They also
proved that there is a low stick-slip propensity in the material
with Sb2S3 compared to that with graphite and therefore reducing
creep groan during braking application, a nding that was also
reported by Jang et al. [12]. A later work by the sterle group [13]
also studied the use of a combination of Sb2S3 and MoS2 as a solid
lubricant and showed good friction performance for the friction
material.
Due to the complex composition of the brake linings it is hard
to describe all the tribochemical reactions that take place on the
friction surface of the brake pad at high temperature during the
braking process. However, some recent works focused on the
reaction between the lubricant and the metal ingredients in the
brake pad material, e.g. in [11], Matejka et al. described the Sb2S3
transformation on the surface of the semi-metallic composites
dsence of iron was the main reason for the origination of elemental antimony at temperatures lower thanuring friction with
the help of a mixture of iron and Sb2S3. As a result they were able
to prove that the pre 360 C, while at higher temperatures the
metallic antimony alloyed powder iron. In [14] the decomposition
of Sb2S3 and alloying of Cu by Sb was described during friction
process and the presence of CuSb and Cu2S was determined.
Recently, brake wear emissions have been cited as a potentially
important source of Sb in the environment, being Sb in form of
Sb2S3. Sb2S3 can be oxidized to Sb2O3 during the braking process
which has been categorized as a potentially carcinogenic substance and has been shown to be partially soluble in physiological
uids [15]. So that, concern about the toxicity of antimony compounds is beginning to force tribological composite designers,
especially in the US and Europe, towards solid lubricants that are
free of antimony compounds [9,16]. Indeed the US legislature
banned Sb2S3 in commercial brake friction materials due to it
being a possible hazard to health. At this point, a detailed understanding of the mechanism for improving fade resistance with
Sb2S3 is necessary in order to nd alternative material to replace it.
In the present work, the intention is to establish a relation
between the friction performance of the brake pads and the
reactivity between Sb2S3 and metal powders present in a lowmetallic formulation (Cu, Sn, sponge Fe and Zn) at high

temperature. For this purpose, different mixtures with Sb2S3 as a


solid lubricant were prepared. Brake pads were characterized
before and after dynamometer testing in accordance with the
procedure in SAE J2522 to study the changes in the microstructure
and the development of the chemical reaction throughout the
depth of the worn friction material. Moreover, in order to understand the chemical reaction and transformation during different
temperature ranges, binary powder mixtures comprising Sb2S3
and metal powders were also studied and carefully analysed using
SEMEDS, DSCTGA and X-ray diffraction techniques.

2. Material and methods


2.1. Friction material (sample preparation)
Fig. 1 shows the morphology of the metal powders and the
lubricant used in this study, while Table 1 includes their specications. The copper powder used was an electrolytic powder
(Fig. 1a), tin powder is a gas atomised powder (Fig. 1b) and sponge
iron has an internal porosity, as shown in Fig. 1c). Zinc powders of
two different particle sizes were used: coarse) and ultrane.
Regarding the Sb2S3 lubricant, impurities such as black particles of
SiO2 were found, as shown in Fig. 1f. The oxygen content of Sb2S3
was 6.55 70.02% O2 (by LECO).
The fabrication of the brake pads carried out in Gama and
containing 13 ingredients was based on keeping the parent composition of 12 ingredients (90 wt%) constant and varying the metal
powders used (copper, tin, sponge iron and zinc) in each composition, as shown in Table 2. The last two formulations were prepared with zinc powders with different particle sizes in order to
study the effect of zinc particle size on the reaction with the
lubricant. In all cases the lubricant used was Sb2S3 (6 wt%).
All constituents were dry-mixed for 3 min in a mixer and
subsequently hot pressed at 170 C and 35 MPa for 4 minutes.
During hot pressing, the mould was released several times to
eliminate the gas entrapment during curing of the phenolic resin.
Then, the friction pads were post-cured for 6 h at 200 C. After the
post-curing process, the surfaces of the pads were ground to attain
the desired thickness and then they were submitted to a scorching
process in order to attain a proper surface nish. The brake pads
measured approximately 11  5  0.9 cm.
With the aim of studying the possible reactions between the
solid lubricant used (Sb2S3) and the metal powders present in the
mixture, binary mixtures of 62.5 wt% of metal powders and
37.5 wt% of Sb2S3 (the same relation between ingredients as in the
brake pad mixtures) were prepared by hand homogenisation of
the powders. The mixtures were called M1 (Cu), M2 (Sn), M3 (Fe),
M4 (coarse Zn) and M5 (ultrane Zn), following the same order as
for the powder blends used for the brake pad fabrication. These
binary mixtures were subjected to different thermal treatments at
200 C, 400 C and 650 C, both in argon and air atmosphere, for
1 h. Differential scanning calorimetry and thermogravimetry
(DSCTGA) measurements of the binary mixtures were carried out
in argon and air from room temperature up to 550 C, using an
automated thermal analyser (Setaram Setsys Evolution 1618).
5 mm diameter and 8 mm deep cylindrical alumina crucibles were
used for these analyses. Approximately 10 mg of sample was
heated at a heating rate of 10 C min  1.
2.2. Dynamometer test
The brake pads were tested on a single-ended inertia full-scale
brake dynamometer (dyno) supplied by TecSA S.R.L. in GAMA,
where a grey cast iron disc was used as rotor. A standard friction
program AK-Master (SAE J2522), detailed in Table 3, was applied in

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

29

Fig. 1. SEM images of the metal powders (a) Cu, (b) Sn, (c) Sponge Fe, (d) Zn coarse, (e) Zn ultrane, and (f) lubricant Sb2S3 (black particles are impurities of SiO2).

(worn surface and cross-sections) so that the reactions between


the metal powders and the lubricant could be studied.

Table 1
Details of the selected metal powders and the solid lubricant.
Powder

Cu
Sn

Supplier

Makin Metal Powders LTD-350#


Pometon Powder106
Pometon Powder

Ave. Particle
size

Purity (%) Ap. density (g/


cm3)

o 45 mm

2.83.2

o 45 mm

499.7

3.8

475 mm
(69%)
Zn coarse
Pometon Powder-S 445 mm
(66%)
Zn ultrane COMETOX
34.5
QUARTZ-Anty65
o 45 mm
Sb2S3
Sponge Fe

2.15
499.7

3.03

98.5

77.2
4.6

Table 2
Relative content (wt%) of raw materials in the brake pads' friction materials.
Ingredients (wt%)/Designation
Parent composition
Sb2S3
Copper
Sn
Sponge iron
Zn (coarse)
Zn (ultrane)
Carbon Fibres (CF)

F1

F2

F3

F4

F5

84
6
10

84
6

84
6

84
6

84
6

2.3. Characterisation methods


The microstructure of the surfaces of the different brake pads
before and after dyno testing were analysed using a Philips
XL30CP Scanning Electron Microscope (SEM) with EDS by EDAX.
This technique was also used to identify the phases of the binary
mixtures during heat treatment. Different cross-sections of the
brake pads before and after dyno testing were prepared by cutting
and subsequently polishing them in order to analyse the microstructure of the sections using Quanta 3D from FEI with EDS
Oxford INCA450.
The phase identication of the transfer layer on the worn pad
surfaces and of the heat-treated binary mixtures was carried out
by X-ray diffraction using a Philips 1 PW 1825 diffractometer
equipped with ICDD software for phase characterisation. The
irradiation used was Cu K, (40 kV and 40 mA). To characterise the
worn surface the glancing technique was used.

10
10

3. Results and discussion

10
6
4

a
Binder (Phenolic resin) 6.5 wt%; friction modiers (alumina, magnesia, graphite, coke, iron oxide) 28 wt%; reinforcements (steel bre, aramid bre) 26 wt%;
llers (NBR powder, barite) 23.5 wt%.

this study to test the friction materials. This test was divided into
two different sections, cold and hot. In the cold section sensitivity
to speed and pressure changes were studied in the friction
material, while in the hot section the temperature in the interface
reached 500 C, simulating extreme braking conditions. Most of
the results from the AK-Master test were reported in the Eurobrake congress EB2014-DF-002 paper [17]. These dyno tests were
used mainly for obtaining the worn brake pads to be characterised

3.1. Characterisation of brake pads


In Fig. 2 the initial microstructure of the surfaces of the different brake pads is shown. In the SEM images with BSE contrast,
copper and its replacements in each brake pad are indicated by
crosses so their distribution can be clearly seen, while the solid
lubricant Sb2S3 is indicated by circles. On all pad surfaces, different
components of the formulation can be distinguished, such as steel
bres (grey contrast), barite particles (white contrast), or graphite
black particles. All constituents are well distributed showing a
homogenous microstructure. In general Sb2S3 particles are not
very well distinguished from barite as they show a similar contrast; however, it is because of their smaller particle size that it is
possible to identify them. The F2 pad with Sn is more complicated,

30

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

Table 3
SAE J2522 AK-Master dynamometer test procedure.
Section

Description

100 stops, speed:80-30 km/h, initial brake temperature (IBT):100 C


8 stops, speed:40-5 km/h, IBT:100 C, pressure:10,20..,80 bar
8 stops, speed: 80-40 km/h, IBT:100 C, pressure:10,20..,80 bar
8 stops, speed:120-80 km/h, IBT:100 C, pressure:10,20..,80 bar
8 stops, speed:160-130 km/h, IBT:100 C, pressure:10,20..,80 bar
8 stops, speed:200-170 km/h,IBT:100 C, pressure:10,20..,80 bar
3.Cold braking check
1 stop, speed: 40-5 km/h, IBT:40 C, pressure:30 bar
4.Motorway braking check #1
1 stop, speed: 100-5 km/h, IBT:50 C, deceleration: 0.6 g
5. Motorway braking check #2
1 stop, speed: 0.9-0.5vmax, IBT:50 C, deceleration:0.6 g
6. Characteristic check (after 1, 2, 3, 5 sections)
6 stops, speed:80-30 km/h,IBT:100 C, pressure: 30 bar
7. First Fade (max 160 bar)
15 stops, speed:100-5 km/h,IBT:100500 C, deceleration: 0.4 g
8. Recovery 1
18 stops, speed:80-30 km/h,IBT:100 C, pressure:30 bar
9. Pressure sensitivity
8 stops, speed:80-30 km/h,IBT:100 C, pressure:10,20..,80 bar
10. Increasing temperature sensitivity (500 C/300 C) 9 stops, speed:80-30 km/h,IBT:100 C, 150,,500 C pressure 30 bar
11. Pressure sensitivity (500 C)
8 stops, speed:80-30 km/h,IBT:500 C, pressure:10,20..,80 bar
12. Recovery 2
18 stops, speed:80-30 km/h,IBT:100 C, pressure:30 bar
13.Second fade (maximun 160 bar)
15stops, speed:100-5 km/h,IBT:100500 C, deceleration:0.4 g
14. Characteristic check
18 stops, speed:80-30 km/h,IBT:100 C, pressure:30 bar
1. Burnish/bedding
2. Speed/pressure sensitivity

as tin, barite and Sb2S3 show similar contrast but again different
particle size. In the F1 mixture with Cu powder, the steel bres and
Cu powders show a similar contrast, but while copper powder
show rounded and smaller particles, steel bres are longer.
On the other hand, the open structure of the sponge iron powder
used in F3 formulation is clearly observed, and it is very different
from the steel bre morphology. Finally, the most signicant feature
is the difference in the size and morphology of the Zn particles on
the surface of formulations F4 and F5. While in F4 large particles of
zinc are well distributed around the surface, in F5 the zinc ultrane
particles are in general observed as agglomerates.
3.2. Friction performance at high temperature
The friction behaviour of the different brake pads during Fade
1 and Recovery 1 sections is shown in Fig. 3, where the effect of
the temperature on m is studied. F5 is not included to its similarity
with F4. It could be observed how formulations F1 and F4, with Cu
and Zn powder respectively, show the highest fade resistance with
a smother reduction of the m, more obvious in the rst brakings at
temperature 4500 C. Formulation with Sn powder (F2) appears
to have the lowest fade resistance with the highest decrease of m,
while F3 with sponge Fe show an intermediate behaviour.
3.3. Study of the worn brake pads
In Fig. 4, the surfaces of the worn brake pads for formulations
F1 (Cu), F2 (Sn), F3 (Fe) and F5 (Zn ultrane) are shown. F4 with Zn
coarse powder is not included due to its similarity to F5. These
SEM micrographs of the friction surfaces show how a friction layer
formed on the surface of all brake pads tested and that it covers a
large part of the worn surface (in a grey contrast). In all cases the
presence of some of the initial constituents of the brake pad formulations (graphite particles and steel bres) can also be distinguished in a dark and lighter grey contrast, respectively. In
addition, some white contrast zones can be observed on these
bres, which apparently indicate areas where the steel bre had
reacted during dyno testing with some of the constituents in the
formulation (Sb2S3 and/or the different metal powders).
The extent of this reaction depends on the metal powder present in the pad formulation. While in F1(Cu) the steel bres nearly
maintain their morphology and chemical composition (only minor
amounts of Cu and Sb have been found on their surface), in F2(Sn)
these bres have a very white contrast, indicating that they have
reacted with Sb2S3 and/or with Sn to a higher degree. Fig. 5 shows

detail of these bres in F2, where the presence of Sb and Sn is


clearly seen from the X-ray elemental mapping of the whole area.
However, no S is detected in zones where Sb is present, thus
indicating that Sb2S3 reacts with Sn and/or Fe to form intermetallic
compounds (SnSb or FeSb) and metallic sulphides. In the F3
formulation, where no Sn is present in the mixture, the presence
of Sb on the surface of the steel bres was also observed, which
would indicate a FeSb intermetallic formation, since again, no S
was observed. In F5 and F4 the white contrast related to Sb on the
steel bres was less evident, similar to F1.
The semi-quantitative EDS analyses carried out on the darker
areas on the surface of the worn pads, show the presence of the
metal powders and the solid lubricant, among the other components used, in the transfer layers. Due to the afnity of Sn to react
with steel bre, as observed in Fig. 5, the content of Sn in the
transfer layer is lower than for the other metals.
Different authors have also observed the presence of antimony
on the surface of steel bre from worn pad surface characterisation, which was also determined to be part of the transfer layer
when a blend of MoS2 and Sb2S3 was used as a solid lubricant in
[13]. In [18] antimony was also reported as being part of the wear
debris along with other components such as Cu, Zn or Sn, although
the authors do not show the chemical composition of the original
composites.
The glancing angle X-ray diffraction technique (GAXRD) is used
for the characterisation of thin polycrystalline structures due to
the low depth of penetration of the X-ray beam relative to its low
incidence angle. Therefore, this technique was used in this work to
characterise the compounds formed on the worn surface during
braking. For this purpose an incident angle of 5.5 was selected,
permitting only the constituents belonging to the transfer layer to
be determined. In Fig. 6 the X-ray patterns obtained on the surface
of the brake pads before (dotted lines) and after (continuous lines)
dyno testing are shown for the different formulations, making it
possible to compare the compounds formed.
The X-ray patterns show that in general the main constituent of
the transfer layer is magnetite (Fe3O4), except for F5 with Zn
ultrane and to a minor degree for F4 with coarse Zn (not included
due to its similarity to F5). This fact was also observed in previous
works [1823]. The main source of the iron oxide, magnetite and
hematite present in the transfer layer is reported to be the tribooxidation of the cast iron. At the same time, the tribo-oxidation of
the steel bre present in the pad could also inuence the iron
oxide amount [20]. Magnetite is an important friction stabiliser in
the transfer layer, which can smooth the behaviour of the friction

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

31

Fig. 2. SEM micrographs of brake pad surfaces prior to dyno testing.

coefcient [23,24]. Even though the structure and the chemical


composition of the friction layer strongly depends on applied
friction conditions [14], it was observed that Sb2S3 was responsible
for the formation of this transfer layer, which at the same time
sustained friction levels at elevated temperature [5].
On the other hand, apart from magnetite, other compounds
have been also identied; their nature and composition depend on
the particular metal powder included in each formulation, such as
Cu2O, observed in the F1 formulation due to the oxidation of the
Cu powder. In this formulation no evidence of reaction products
was observed by the glancing X-ray technique, only some
remaining Cu was determined. For the F2 formulation, SnO2 was
clearly identied and some minor peaks belonging to FeO were

also observed. Sn is completely oxidised, since no Sn was found.


However, no FeSn or SnSb intermetallic compounds were identied, as expected after the EDS analyses of the steel bres, shown
in Fig. 5. Additionally, in F3 with sponge iron powder, FeO was also
identied and it showed a higher concentration than in F2,
therefore indicating that this sponge iron is possibly oxidising to
FeO, in contrast to the cast iron that oxidises to Fe3O4. In F3
intermetallic FeSb was identied, which coincides with the SEMEDS characterisation of the white zones of the steel bres on the
worn surface in Fig. 4. The presence of this intermetallic compound would conrm that a reaction between Fe and Sb2S3 took
place during the dyno test. Finally, as mentioned before, in F4 the
constituents of the transfer layer are very different from the rest of

32

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

0,5
F1

F2

F3

0,5

F4

F1

0,4

F2

F3

F4

0,4

0,3

0,3

0,2

n of braking (Fade 1)

0,2

n of braking (Recovery 1)

Fig. 3. Fade 1 and Recovery 1 sections of the friction test AK-Master.

Fig. 4. SEM micrographs of the worn brake pad surfaces after dyno testing.

formulations. In this case it seems that Zn powder is easily oxidised and it transfers to the surface of the cast iron disc, protecting
it from the tribo-oxidation as it is known to work as a solid
lubricant at elevated temperature [18]. Furthermore, the identication of zinc sulde, ZnS, in the X-ray pattern would corroborate
the reaction between Zn and Sb2S3.
Most of the peaks belonging to barite could be distinguished in
the transfer layer (see Fig. 6) even after dyno testing, but no Sb2S3
was detected. A unique property of copper and other soft constituents, such as graphite, barite [25], MgO and certain metal
sulphurs (Sb2S3, MoS2, SnS) is that they seem to transform into an
equiaxed nanocrystalline structure with that easily incorporate
into the transfer layer, which basically formed by magnetite [26].

From the characterisation of the worn surface by SEM and Xray diffraction, it was observed that the Sb2S3 lubricant had reacted with the metallic constituents of the brake pad during braking.
In order to study the extent of these reactions, analyses underneath the worn brake surface were performed. For this purpose
cross-section of the worn pads were prepared and analysed by
SEM and EDS up to a maximum depth of 9 mm (bottom of
the pad).
Although all the brake pad formulations were analysed, Fig. 7
shows as an example the SEM micrographs taken at different
depths (from the surface up to 9 mm in depth) for copper content
formulation F1. The analysis of these micrographs reveals that Cu
and Sb2S3 reacted during braking tests since no individual Cu and

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

33

Relative intensity

Fig. 5. X-ray mapping of the worn surface of brake pad F2.

20

25

30

35

40

45

2 (deg)
Fig. 6. GAXRD patterns taken at glancing angle from the brake pad surfaces. The
dotted line represents pre-dyno and the solid line represents post-dyno testing;
where
is barite and
is Fe3O4..

Sb2S3 particles could be found under the worn surface up to 6 mm.


In these micrographs the reaction products are marked with an
x. Only at a depth of 9 mm was almost no reaction observed and
unreacted Sb2S3 particles were observed. Taking to account that
this reaction can only take place at high temperatures, it is
believed that during breaking high temperatures were attained in
the worn pad even to a signicant depth, encouraging the presence of the reaction throughout the depth of the worn pad.
Additionally, Fig. 8 shows the details of the reaction products
between Sb2S3 and the different metallic powders found in the
cross-sections of the different formulations.
In Fig. 8a, a SEM micrograph (BSE contrast) and two EDS analyses of the reaction products between Cu and Sb2S3 inside a worn
F1pad are presented. The particle shown in this micrograph, primarily indicated by a light grey contrast, corresponds to copper

sulphide (Cu2S), and the white zones correspond to the intermetallic Cu2Sb. Moreover, the darker grey zones correspond to
copper-zinc sulphide (CuZnS). Therefore, the reaction between Cu
and Sb2S3 results in copper sulphide and copper antimonide. The
presence of copper zinc sulphide can be explained by the presence
of Zn in the parent formulation. It is important to note that similar
behaviour was found for the rest of the formulations.
In Fig. 8b a particle of Sn in the cross-section of a worn F2 pad
that had reacted with Sb2S3 is shown. In this case the white and
grey zones were identied as intermetallic tin antimonide (SnSb)
and tin sulphide (SnS), respectively. This reaction was observed up
to a depth of 6 mm, similar to the cross-section of F3 containing
sponge iron powder, whose reaction details are shown in Fig. 8c.
The reaction between sponge iron and Sb2S3 is shown at the edge
of an iron sponge particle where iron sulphide (FeS) (grey areas)
and iron antimonide (FeSb) (white areas) could be distinguished.
On the other hand, the characteristics of this reaction for the
formulations containing zinc powder (F4 and F5) differ from the
difference in the particle size of the Zn powder used in each case.
Fig. 8d shows that for F4 the reaction between coarse Zn and Sb2S3
produces zinc sulphide (ZnS), which appears in the dark grey
contrast surrounding the zinc particle, and zinc antimonide
(ZnSb), which is revealed with a white contrast. Meanwhile, Fig. 8e
shows how this reaction and the reaction products found in F4 are
limited by the presence of Zn agglomerates. These agglomerates
were partially oxidised during the dyno test, forming a thin layer
of ZnO around the particles. It seems that the presence of such an
oxide layer around the Zn particles prevented the reaction
between Zn and Sb2S3, and only Zn particles that had not been
oxidised or located at the border of the Zn agglomerates and were
in contact with Sb2S3 had reacted, forming zinc sulphide (ZnS) and
zinc antimonide (ZnSb).
In summary, analyses carried out on cross-sections of the brake
pads revealed that Sb2S3 reacts with Cu, Sn, Fe and coarse Zn to
produce metal antimonides and metal sulphides up to different

34

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

Fig. 7. Evolution of the reaction between Cu and Sb2S3 (marked with x) along pad depth. Only at 9 mm can unreacted Sb2S3 be found.

depths, Cu being the most reactive powder in these conditions in


that it reacts with Sb2S3 up to 9 mm in depth (see Table 4). On the
other hand, it was observed that in F5 (with ultrane Zn powder)
the reaction was less favoured due to the presence of Zn
agglomerates.
Finally, no antimony oxide was found on the surface of the
worn brake pads, so it is believed that when the brake pad and
rotor enter in contact, the friction composite tends to absorb the
oxygen present in the air and this reacts with the Sb2S3 present in
the friction material, forming antimony trioxide; this reacts further
with the iron to form iron oxide and metallic antimony [11]. It is
likely that the presence of alloyed antimony iron was observed on
the surface of the steel bres because of this phenomenon.
3.4. Chemical reactions between Sb2S3 and metal powders
Solid lubricants change their chemical states at elevated temperature and exhibit different tribological properties compared to
those at room temperature. Due to the heat generated during
braking, pad and disc temperature can increase up to 550 C
(maximum temperatures reached during dyno testing). Therefore,
in this work the study of the reaction between the powder metals
and Sb2S3 has been studied via DSC/TGA and XRD analyses until up
to temperature. For that purpose the thermal treatments applied
to the binary mixtures (M1-M5) were carried out under air or
argon in order to try to reproduce the atmosphere conditions on
the brake pad surface (oxidising atmosphere) or inside the brake
pad (inert atmosphere). However, it should be taking into account
that no pressure was applied in these treatments.
3.4.1. Characterisation of the Sb2S3 lubricant
DSCTGA analyses were carried out for the Sb2S3 powder in air
or in argon up to 600 C in order to study its thermal stability in
conditions similar to which the brake pads were exposed. These
analyses will serve as a reference for the study of the binary
mixtures. Fig. 9 compares the DSC (continuous line) and TGA
(discontinuous line) curves obtained in both atmospheres. First of
all, in both cases, between 200 C and 250 C a slight decrease in
the TGA curve could be observed, which is probably related to the
decomposition of some volatile species.
Taking into account that the melting point of Sb2S3 is  550 C,
the main difference between air and argon atmospheres takes
place below this temperature. Whereas in air three clear reactions
associated with exothermic peaks (13) are observed, in argon
only the melting point of Sb2S3 is identied from endothermic
peak 4 of the DSC curve. Therefore, the peaks detected in air must
be related to the oxidation of Sb2S3. In fact, [5,6,27] report that
Sb2S3 oxidation takes place in the temperature range of 300
430 C, where Sb2S3 is converted to Sb2O5 and Sb2O3 and then to

Sb2O4 at 570 C. This last component was reported to form at


approx. 500 C in an oxygen rich environment [11]. However, Ingo
et al. [28] found that Sb2S3 was rst supercially oxidised to Sb2O4
and then to the stable form Sb2O3 because of the temperature
reached during braking. Therefore, peaks (1) and (3) in the DSC
curve in air must be related to the oxidation of the solid lubricant,
since a mass change is also observed in the TGA curve. In contrast,
peak 2, where there is no associated mass change, could be related
to a phase transition. Hence, the oxidation of Sb2S3 takes place in
different steps as the temperature increases.
In order to identify the oxide products formed during the oxidation of Sb2S3 at high temperature and associate them with the
DSC/TGA peaks marked in Fig. 9, XRD analyses were carried out on
the powder heated at 400 C and 650 C, respectively, in air for 1 h.
Fig. 10 shows the XRD patterns of the Sb2S3 powder and the
powder heated at these temperatures (200 C is not included since
no reaction products were found at this temperature, the same is
true of all the binary mixtures analysed). As observed in the pattern obtained from the powder heated at 400 C, Sb2S3 is completed oxidised to Sb2O3 at this temperature. Therefore, peak 1 in
the DSC curve (Fig. 9a) corresponds to the oxidation of Sb2S3 to
Sb2O3 with a mass loss of 10%, according to Reaction (1) (liberation
of SO2). This is also the reason why no Sb2S3 melting peak is
observed in the DSC performed in air, compared to the treatment
in Ar.
2Sb2S3 9O2-2Sb2O3 6SO2

(1)

According to the XRD chart in Fig. 10, -Sb2O3 is the main


polymorph identied, with a residual amount of -Sb2O3. The XRD
pattern of the Sb2S3 powder heated at 650 C indicates that Sb2O3
is subsequently oxidised to Sb2O4 (Reaction (2)). This reaction
must be related to peak 3 on the DSC curve and correlated to the
mass increase observed in the TGA curve. Therefore, peak 2 in the
DSC curve may be associated with the phase transition of
Sb2O3, where no mass change is observed.
Sb2O3 1/2O2 -Sb2O4

(2)

On the other hand, a small amount of -Sb2O3 and -Sb2O5 was


identied by XRD in the powder treated at 400 C and 650 C,
respectively, in argon atmosphere. This would indicate that some
residual oxygen was present in the furnace where the heat treatment was performed, since no mass change was determined at
these temperatures in the DSCTGA analyses.
Finally, even though the presence of SiO2 (quartz) was identied in all XRD patterns, since it is an impurity of the primary Sb2S3
powder, for simplicity its peaks will not be marked in the following gures.

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

35

Fig. 8. Detail of metal particles that reacted with Sb2S3 lubricant in worn pad cross-sections: (a) F1 (Cu) pad, (b) F2 (Sn) pad, (c) F3 (Fe) pad, (d) F4 (Zn coarse) pad and (e) F5
(Zn ultrane) pad.

3.4.2. Binary mixtures


3.4.2.1. M1 (Cu Sb2S3). In the DSC (a) and TGA (b) curves for the
M1 binary mixture, the primary Cu and Sb2S3 powders obtained
during heating up to 550 C in air are shown in Fig. 11. First of all,
as a reference it can be seen that on the curves corresponding to
the primary Cu powder only a slight oxidation of the metal could
be observed, without the presence of any specic peak. In contrast,
for the M1 mixture three exothermic peaks and a continuous mass
gain can be observed on the DSC and TGA curves, respectively. The
presence of such transitions indicates that some reactions

Table 4
Reaction products and the depth up to where these reactions were observed on the
worn brake pad cross-section.
Meta powder

Reaction products

Depth of the reaction [mm]

Cu
Sn
Sponge Fe
Zn (coarse)
Zn (ultrane)

Cu2Sb, Cu2S
SnSb, SnS
FeSb, FeS
ZnS, ZnSb
ZnS, ZnO, ZnSb

9
6
6
6
6

36

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

Fig. 9. DSCTGA curves for Sb2S3 (a) in air atmosphere (b) in argon atmosphere.

between Sb2S3 and Cu take place during heating. First, in the DSC
curve peak (1) arises between 300 and 450 C, peak (2) between
450 and 500 C, and peak (3) starts at 500 C. XRD patterns from
the M1 mixture and the mixture heat-treated at 400 C and 650 C
are shown in Fig. 11c to assist in the identication of the reactions
involved in the DSC peaks.
The XRD pattern for M1 heated at 400 C may help to explain
the reaction or reactions related to DSC peak (1). Four different
compounds were identied from this pattern: Sb2O3 (mainly ,
with some residual ), Cu, Cu2O and Cu2S. Firstly, as happened
with the primary Sb2S3 powder, the Sb2S3 present in the M1
mixture is completely oxidised to Sb2O3. However, from the TGA
curve of the M1 mixture (Fig. 11b), this reaction is combined with
a mass gain of about 5%, differing from the mass loss observed for
primary Sb2S3 (Fig. 9). Therefore, the SO2 produced from the oxidation of Sb2S3 according to Reaction (1) is not liberating, and it
may react with Cu according to Reaction (3) to form copper sulphide, Cu2S, and copper oxide, Cu2O. Furthermore, some Cu2O can
be formed from the direct oxidation of Cu, which according to the
XRD pattern it is not completed oxidised after 1 hour at 400 C.
6Cu SO2-Cu2S2Cu2O

(3)

Additionally, exothermic peaks 2 and 3 in the DSC curve


(Fig. 11a) are also accompanied by a continuous mass gain.
According to the XRD pattern of M1 heat treated at 650 C in
Fig. 11c, at least two reactions may have taken place between 400
and 650 C, which justies the presence of the two new compounds identied in the XRD pattern: copper antimony oxide,
CuSb2O6, and copper (II) oxysulphate, Cu2O(SO4). Regarding the
copper antimony oxide, Scarlat [29] and Stan et al. [30] reported
the thermodynamic properties of CuSb2O6 and the Cu-Sb-O phase
equilibrium, and prepared this compound by using a mixture of
Sb2O3 and CuO heated up to 1000 C. They reported that this
reaction takes place at 750 C, but it is only possible if Sb2O3 is
previously oxidised to Sb2O4 (Reaction (2)), which is reported to
occur at 500 C. Once Sb2O4 is formed, it reacts with CuO to form
CuSb2O6. Therefore, in mixture M2 CuSb2O6 is formed by a similar
mechanism (Reaction (5)), which implies that Cu2O should have
also oxidised previously according to Reaction (4). In this case,
Reaction (5) occurs at a lower temperature than the one reported
by these authors.
Cu2O 1/2O2-2CuO

(4)

CuO Sb2O4 O2-CuSb2O6

(5)

Although it is difcult to ascertain, these reactions could be


related to peaks 2 and 3 of the DSC curve (Fig. 11a) and the
resultant mass gain determined by TGA (Fig. 11b). This would
indicate that the oxidation temperature of Sb2O3 in M1 mixture
(450550 C) is lower than that determined for primary Sb2S3
(580 C, Fig. 9(a)).
The formation of Cu2O(SO4) seems to be more complicated. Firstly,
Cu2S should be oxidised to CuO. Ramakrishna et al. [31] studied the
kinetics of the oxidation of Cu2S, nding that it oxidises between 750
and 900 C, rst to Cu2O and subsequently to CuO (liberating SO2
during this process), which is a highly exothermic reaction. They
propose a CuSO phase diagram as a function of the PO2 and PSO2. In
such a phase diagram, Cu2O(SO4) is stable in a specic PO2PSO2 range.
Thus, in the present conditions SO2 is not liberated, which increases
the PSO2 and favours the formation of Cu2O(SO4). Ginting et al. [32]
synthesised this compound from the oxidation of anhydrous CuSiO4 at
725 C. Therefore, in any case the formation of Cu2O(SO4) implies a
mass gain and consequently the reaction or reactions associated with
this formation could also be associated with peaks 2 and 3 of the DSC
curve (Fig. 11a).
On the other hand, Fig. 12a shows the DSC curve for M1 in
argon compared to that of primary Sb2S3 and the XRD analyses of
M1 before and after the heat treatments in argon at 200, 400 and
650 C (Fig. 12b). As shown in Fig. 12a, when argon atmosphere
was used, no exothermic peaks are detected, which would indicate
that any reaction happens between Sb2S3 and Cu. Only the endothermic peak corresponding to the melting of Sb2S3 at 525 C is
observed, although the very low intensity of this peak could suggest that some Sb2S3 has reacted below this temperature. The last
observation can be conrmed by the analysis of the XRD patterns
presented in Fig. 12b, where different reaction products between
Cu and Sb2S3 can be identied, even at 200 C. The presence of
copper antimony sulphide, CuSbS2, in the M1 mixture heat-treated
at 200 C reveals that the reaction between Sb2S3 and Cu starts to
take place at a very low temperature.
Additionally, at 400 C the reaction is complete, whereas the
reaction products are different than those found at 200 C: copper
antimonide, Cu2Sb, and copper sulphide, Cu2S. The fact that no
exothermic peak related to this reaction has been detected by DSC
may indicate that the reaction kinetic is very slow, the enthalpy
heat realised is very low, or the heat realised is used to balance the
stoichiometry of the compounds. This is in line with the presence
of the very low intense peak of the Sb2S3 melting, which means
that during heating (10 C/min) the reaction has not been completed. Only a long dwell time (60 min) at a sufciently high
temperature, i.e. 400 C, allows this reaction to complete. There is

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

no further change in the heat treatment carried out at higher


temperatures (up to 650 C), thus denoting that both compounds
are stable in inert atmosphere, in contrast to that observed in air.
The analyses done by DSC/TGA and XRD from the heat treatments of binary mixture M1 in air and in inert atmosphere can be
related to the characterisation of the worn brake pads and consequently the reaction that takes place between Cu powder and
the lubricant Sb2S3 during braking can be understood. The compounds that were determined when argon atmosphere was used
in those analyses are the same as those found in the cross-section
of the worn pad, whereas those found on the worn surface are
more comparable to those determined in the heat treatments
performed in air.

Sb2S3_650C Air

Sb2S3_400C Air

Sb2S3

20

25

30

35

40

45

3.4.2.2. M2 (Sn Sb2S3). Fig. 13 shows the DSC and TGA curves in
air (a) and in argon (b) for Sn and the binary mixture M2. In
addition, Figs. 13c and 13d also shows the XRD patterns for M2
heat treated at 400 C and 650 C in air and argon respectively. For
primary Sn powder, only an endothermic peak (1) associated with
its melting point at 225 C is observed (Fig. 13a), without any mass
change throughout the whole heating cycle. Evidently, this peak is
also observed in the M2 mixture. For this mixture two additional
exothermic peaks are clearly seen (2 and 3), the rst one occurring

50

2 (deg)
Fig. 10. XRD patterns of Sb2S3 solid lubricant without treatment, heating up to
400 C and 650 C in air where:
Sb2S3,
SiO2,
-Sb2O3,
Sb2O3 and

37

-Sb2O4..

20%

M1

15%

Sb S

2
10%

m (%)

Heat Flow (mW/mg)

5%

Cu

0%

2
Cu

Sb S

-5%

1
M1

-10%

0
-1
100

150

200

250

300

350

400

450

500

-15%
100

550

150

200

250

Temperature (C)

300

350

400

450

500

550

Temperature (C)

M1_650C Air

M1_400C Air

M1

20

25

30

35

40

45

50

2 (deg)
Fig. 11. (a) DSC and (b) TGA curves in air for Sb2S3, Cu and M1 mixture; (c) XRD patterns of M1 initial mechanical mixture, M1 after heating up to 400 C and 650 C in air
where:
Sb2S3,
Cu,
Cu2O,
-Sb2O3,
-Sb2O3,
Cu2S,
Cu (Sb2O6), and
Cu2O(SO4).

38

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

M1_650C Ar

7
6
5

Heat flow (W/m0)

M1_400C Ar

M1_200C Ar

2
1

M1
Sb S

M1

0
-1
100

150

200

250

300

350

400

450

500

550

20

25

30

35

40

45

Fig. 12. (a) DSC curves in argon for Sb2S3 and M1 mixture; (b) XRD patterns of M1, M1 after heating up to 200 C, 400 C and 650 C in argon where:
CuSbS2,

Cu2Sb and

4%

6
M2

2
3

0%

2
Sn

3%
2%
M2

1%
0%

2
-1%

M2

Sb S
-2%

-3%

-1

-2
200

M2

150

4%

-2%

100

Cu,

250

300

350

400

450

500

-4%
550

m (%)

2%

Heat Flow (mW/mg)

m(%)

Heat Flow (mW/mg)

-1

Sb2S3,

Cu2S.

50

2 (deg)

Temperature (C)

-2
100

150

200

250

Temperature (C)

300

350

400

450

500

-4%
550

Temperature (C)

M2_650C Ar

M2_650C Air

M2_400C Air

M2_400C Ar

M2_200C Ar
M2

20

25

30

35

40

45

50

20

25

30

35

40

45

50

2 (deg)

2 (deg)

Fig. 13. DSC and TGA curves in air (a)and in argon(b) for Sn and M2 mixture; XRD patterns of M2 and M2 heat treated at 400 C and 650 C in air (c) and argon(d) where:
Sb2S3,

Sn,

SnO2,

SnS,

SbSn,

-Sb2O3 and

-Sb2O4.

between 340 and 440 C and related to a mass gain of  2%, and
the second one occurring between 450550 C and with no
apparent mass change. The XRD pattern obtained from M2 heattreated at 400 C (Fig. 13c) shows that the reaction between Sn (in
liquid state) and Sb2S3 that yields tin sulphide, SnS, and tin

antimonide, SbSn, takes place (according to Reaction (6)). Taking


into account that this reaction does not involve a mass change, it
may be associated with peak 3 on the DSC curve. Therefore, in
contrast to what was found in the M1 mixture, in this case a
reaction between the metal and the lubricant takes place in air.

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

seen, with no related mass change in the TGA curve. At the same
time, the lack of a melting point endothermic peak related to the
melting point of Sb2S3, could indicate that the reaction between
the lubricant and Sn takes place in the temperature range where
the low intense DSC peaks are shown. The XRD patterns (Fig. 13d)
conrm this observation. Although at 200 C no reaction is
observed in M2 at this temperature, at 400 C the formation of SnS
and SbSn from the reaction between Sn and Sb2S3 (Reaction (6)) is
determined, with some residual Sn and a certain amount of SnO2,
which means than some oxygen is remaining in the atmosphere.
However, no Sb2S3 is found, in agreement with the lack of its
melting point peak in the DSC curve. The only difference between
the XRD patterns at 400 C and 650 C is that the peaks belonging
to Sn disappeared at 650 C.
For this mixture the correlation between the analyses performed in argon also coincides with the characterisation of the
cross-section of the worn brake pads. However, the correlation
between the heat treatments in air and the surface of the worn
pad is more complicated. In both cases SnO2 is found, but SnS and
SbSn are barely found on the worn surface.

Additionally, Sb2O3 and SnO2 are also identied by XRD in the


mixture heat-treated at 400 C (with some remaining Sn) formed
by the oxidation of Sn to SnO2 (Reaction (7)) and the oxidation of
Sb2S3 to Sb2O3 (Reaction (1)), which is related to peak 2 of the DSC
curve. These reactions have an associated mass gain of 13.4% and
mass loss of 9.4%, respectively. The combination of both oxidation reactions, taking into account the initial Sn/Sb2S3 ratio in the
binary mixture and the presence of some unoxidised Sn, as
determined in the XRD pattern, may result in a mass gain, which
should be similar to the one determined from the TGA curve
( 2%) at the temperature range where peak 2 occurs. It must be
pointed out that it seems that the oxidation of Sb2S3 promotes the
oxidation of Sn, since, as mentioned previously, primary Sn alone
does not oxidise.
Sb2S3 5Sn-2SbSn 3SnS

(6)

Sn 1/2 O2-SnO2

(7)

The XRD pattern obtained from the mixture heat-treated at


650 C, reveals that Sb2O3 is further oxidised to Sb2O4, and Sn is
completely oxidised. The presence of SnSb and SnS is also conrmed, without producing any other reaction.
On the other hand, Fig. 13b shows the behaviour of M2 in
argon. In this case, apart from the melting point of Sn, no welldened peaks can be distinguished in the DSC curve for the M2
mixture. Only some slight peaks between 400 and 500 C can be

3.4.2.3. M3 (FeSb2S3). For this mixture three exothermic peaks


can be clearly seen in the DSC analysis carried out in air (Fig. 14a).
Remarkably, the DSC curve for M3 is very similar to that obtained
for M2 except for the Sn melting peak. Nonetheless, the TGA
curves for M2 and M3 are very different.

2%

7
6

6
Fe

1%

1
4

2 3
0%

2
1
M3

m (%)

Heat Flow (mW/mg)

Heat Flow (mW/mg)

39

4
3
2
M3

-1%

0
Sb S

M3

-1

-1

-2

-2

-2%
100

150

200

250

300

350

400

450

500

100

550

150

200

250

300

350

400

450

500

550

Temperature (C)

Temperature (C)

M3_650C Air
M3_650C Ar

M3_400C Ar
M3_400C Air

M3
M3_200C Ar

20

25

30

35

40

45

50

20

25

30

35

40

45

50

2 (deg)

2 (deg)

Fig. 14. DSC and TGA curves in air (a)and in argon(b) for Sn and M3 mixture; XRD patterns of M2 and M3 heat treated at 400 C and 650 C in air (c) and argon(d) where:
Sb2S3,

Fe,

Sb2O3,

Sb2O3,

Fe2O3,

Sb2O4 and

Fe3O4.

40

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

Whereas in M3 a slight mass loss is observed at the same


temperature range where the three DSC peaks are observed, in M2
there was a mass gain. This would indicate that in M3 the oxidation of Sb2S3 is the main reaction. This is conrmed by the XRD
pattern corresponding to M3 heat-treated at 400 C (Fig. 14c),
where the main phase identied is -Sb2O3. Apart from Fe and a
small amount of Fe2O3, this pattern is very similar to that of primary Sb2S3 treated at 400 C. Consequently, peak 1 in the DSC
curve corresponds to the oxidation of Sb2S3 to Sb2O3. Peaks 2 and
3 are related to the - Sb2O3 transformation, since no mass
change takes place. In contrast to the reactions found in M1 and
M2 mixtures, there is no reaction between the lubricant (or its
oxide) and the Fe powder. The XRD pattern from the mixture
treated at 650 C in air denotes that Sb2O3 does not react with Fe
nor with iron oxide, even at this high temperature. Besides, iron is
completely oxidised to Fe3O4.
In contrast, when the heat treatment in M3 is performed in
argon, a reaction between Sb2S3 and Fe takes place, producing iron
sulphide and iron antimonide. In the DSC curve for M3 (Fig. 13b)
almost no evidence of this reaction is revealed. Once again there is
an indication that this reaction releases a very low amount of heat.
The XRD patterns obtained from the mixture heat-treated at 400
and 650 C (Fig. 14d) indicate than the reaction stoichiometry and
the stoichiometry of the iron antimonide produced depends on
the temperature of the heat treatment. Whereas at 400 C the
reaction products are FeS and FeSb2, at 650 C they are FeS and

FeSb. The latter reaction products have also been observed by


Matjka et al. [11] in a mixture of iron powder and Sb2S3 subjected
to a heat treatment at 500 C in argon.
Sb2S3 4Fe-3FeS FeSb2

(8)

Sb2S3 5Fe 3FeS 2FeSb

(9)

3.4.2.4. M4 (Zn coarse Sb2S3) and M5 (Zn ultrane Sb2S3). Fig. 15


shows the DSCTGA curves and XRD patterns for the binary
mixtures with Zn coarse (M4) and Zn ultrane (M5) heat treated
in air. For both Zn primary powders, apart from the DSC endothermic peak corresponding to the melting point of Zn, which is
observed at 420 C in both cases, Zn ultrane oxidises once the
powder melts, whereas Zn coarse is not oxidised.
Comparing the DSC/TGA curves of the M4 and M5 mixtures, a
clear inuence of the Zn particle size on the temperature at which
the mean reaction between Zn and Sb2S3 takes place is revealed,
although in both cases the temperature at which it starts is the
same (325 C). For M4 this reaction starts very slowly with initial
low heat liberation and subsequently a very intense exothermic
peak just after the Zn melts. In contrast, for the M5 the reaction
starts faster and the main liberation of heat occurs before the
melting temperature of Zn. Therefore, in M5 this reaction takes
place in solid state, whereas in M4 it starts in solid state and nishes in liquid state. Moreover, the same tendency is observed in

35

5%
M4

30

4%
M5

3%
20

m (%)

Heat Flow (mW/mg)

25

15

M4

2%

M5

10

Zn ultrafine

1%

Zn ultrafine

Zn coarse

5
0%
0

Zn coarse

-1%

-5
100

150

200

250

300

350

400

450

500

100

550

150

200

250

300

350

400

450

500

550

Temperature (C)

Temperature (C)

M4_600C Air
M5_600C Air

M4_400C Air

M5_400C Air

M4
M5

20

25

30

35

40

45

50

20

25

30

Fig. 15. DSC (a), TGA (b) curves and XRD patterns in air for M4 (c) and M5 (d) mixtures. Where:
ZnSb and

ZnSb2O4.

35

40

45

50

2 (deg)

2 (deg)
Sb2S3,

Zn,

ZnO,

-Sb2O3,

ZnS

Zn4Sb3,

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

the TGA curves; it seems that the mass gain, which is related to the
oxidation of Zn, starts after the reaction between Zn and Sb2S3 is
almost completed, at 350 and 460 C for M5 and M4, respectively.
The XRD diffraction analyses carried out on the mixtures heattreated in air at 400 and 650 C, also corroborated the effect of
particle size on the reactivity of Zn with Sb2S3. Thus, even though
in both binary mixtures ZnS is formed at 400 C, the stoichiometry
of the intermetallic formed in each case is different, -Zn4Sb3 and
-ZnSb for M4 and M5 respectively. In both mixtures a certain
amount of Zn is unreacted (the molar relation of the binary mixture is 9Zn/Sb2S3, meaning there is an excess of Zn). Additionally,
in line with the different mass gains determined in the TGA analyses, the ultrane Zn powder in M5 is more oxidised, since the
intensity of the ZnO peaks in the XRD pattern is higher than in M4.
The different surface area of the Zn powders, which also implies a
different oxidation tendency, may be responsible for the different
products formed. At 400 C only traces of Sb2O3 were determined
by XRD, which combined with the mass gain of 3% determined by
TGA, this proves that the SO2 produced from oxidation according
to Reaction (1) is not liberating and may react with Zn to form ZnS
and ZnO (as it was the case for the M1 mixture).
In addition, at 650 C minor differences are observed between
M4 and M5, where Zn is fully reacted or oxidised, At this temperature, the antimonide formed on both is the same: ZnSb, with a
large amount of ZnO. ZnSb2O4 in M4 is identied, which would be
due to the reaction between ZnO and Sb2O4 that formed during
the oxidation of Sb2O3 (Reaction (2)), meaning that the reaction
mechanism would be the same as in the case of M1 with Cu.
Fig. 16 shows the DSC curves and the XRD patterns for M4 and
M5 for the mixture heat-treated in argon. The TGA curves are not
included since no mass change was observed. It is important to
mention that in contrast to the rest of the mixtures, broad peaks
are observed in the DSC curve (Fig. 16a). Once again, the particle
size of Zn affects the reaction with Sb2S3. For M4 the exothermic
reaction starts at a similar temperature as in air, 350 C, although
in argon it seems to nish just before the melting temperature of
Zn, whose peak is also noted. Otherwise, for M5 two exothermic
peaks are observed, the rst one at very low temperatures (200
325 C) and the second one is similar to what happens in M4, but
at a lower temperature ( o30 C). The Zn melting transitions also
take place after the last reaction nishes.
The XRD patterns obtained from the mixtures heat-treated at
200, 400 and 650 C (Fig. 16b) do not show the differences found
in the DSC curves for both mixtures. The patterns from M4 and M5
mixtures heat-treated in argon at 400 and 650 C are the same.

41

Apart from ZnS, Zn4Sb3 and ZnSb are found at 400 and 650 C,
respectively. Therefore, the stoichiometry of the intermetallic
formed depends on the temperature at which the reaction
takes place.
For M4 and M5 binary mixtures, with regard to the reaction
products, few differences between the heat treatments carried out
it air or in argon have been found. Apart from the higher amount
of ZnO formed in air, in both cases zinc sulphide and zinc antimonide are formed. Since no agglomerates are present in binary
mixture M5 with Zn ultrane, the reaction is complete in comparison to the reaction found in the cross-section of the worn
F5 pad.
3.5. Relation between the friction performance of the brake pads and
the interfacial chemistry
Generally, it was observed that the friction transfer layer
formed on the brake pad and disc surfaces was not a homogenous
lm and that its composition depends largely on the tribooxidation of the disc, on the initial composition of the brake
pads and, also, on the reaction between the solid lubricant and the
metal constituents (powders and bres).
The presence of Cu and Cu2O in the transfer layer of the formulation F1 may act as lubricants at high temperatures and,
therefore, contribute to the lower pad and disc wear during
braking tests helping to obtain a steady friction coefcient at high
temperatures. Similar friction behaviour was observed in F4 formulation (Zn). In this case a high concentration of Zn was determined in the transfer layer. It is believed that the Zn powder is
easily oxidised protecting the cast iron disc from the tribooxidation. Additionally, ZnO and ZnS, present in the transfer
layer, could be acting as lubricants, as is reported by [18], so that
supporting this friction behaviour.
On the other hand, F2 formulation with Sn powder seems to
have the lowest fade resistance with the highest decrease of m. In
this case, the steel bres have actively been reacted with the soli
lubricant and with the Sn powder, as it was determined on the
worn surface of the pad, affecting on its friction behaviour.
Moreover, the presence of Fe3Sn and SnO2 compounds on the
worn disc surface may also be impairing this worse friction
behaviour.
Finally, sponge iron powder in F3 modies the tribo-oxidation
behaviour of the friction disc. It could be noticed that it is the only
formulation in which the intermetallic FeSb and a high amount of
FeO are found in the disc transfer layer. Therefore, both

5
M4_650C Ar

Heat Flow (mW/mg)

M4

3
M4_400C Ar

2
M5

1
Sb S
M4_200C Ar

-1
100

150

200

250

300

350

400

450

500

550

20

25

30

Temperatura (C)
Fig. 16. (a) DSC curves and (b) XRD patterns for M4 (coarse Zn) and M5 (Zn ultrane) after heat treatments in argon. Where:
ZnSb,

Zn4Sb3.

35

40

45

50

2 (deg)
Sb2S3,

Zn,

ZnO,

ZnS,

42

A.M. Martinez, J. Echeberria / Wear 348-349 (2016) 2742

compounds should be related to the presence of such metal


powder and its reaction with Sb2S3.

4. Conclusions
The chemical reaction that occurs between the different metal
powders (Cu, Sn, Fe, Zn) and the solid lubricant (Sb2S3) in the
brake pad during braking depends on the nature and morphology
of the metal powder, the temperature attained during braking and
the atmosphere conditions. In general, inside the brake pad a
combination of a metal sulphide and a metal antimonide is formed
in all formulations, whereas the transfer layer on the worn surface
is composed of a mixture of metal oxides and, depending on the
combination of the powder metal and Sb2S3, some reaction products can also be found.
The low angle glancing X-ray technique was successfully used
to determine the constituents of the transfer layer on the worn
brake pad surface. In all cases, magnetite, Fe3O4, is the main
constituent, expect for the Zn content formulations, in which Zn is
easily oxidized to ZnO, which protects the cast iron disc from the
tribo-oxidation.
It is possible to explain the chemical reactions between the
metal powders and the lubricant included in the brake pad formulation by the preparation of the binary mixtures and the use of
DSCTGA and X-ray diffraction techniques to determine the reaction products formed during thermal treatments in argon and air
atmospheres. Whereas the thermal treatments carried out in air
could be used to reproduce the braking conditions on the pad
surface (except pressure), the thermal treatments performed in
argon could reproduce the reactions that take place inside the
brake pad.
The reaction between the metal powders and the lubricant
starts at temperatures between 350 and 400 C, except for the Cu
Sb2S3 system, where it starts at a lower temperature (200 C). This
observation could be used to estimate the maximum temperature
inside the brake pad during the complete braking dyno test.
According to these data, 400 C can be attained at a depth of 6 mm.

Acknowledgements
The authors acknowledge POMETON S.p.A. (130018/11) for its
nancial support while carrying out this work and GAMA S.p.A. for
providing the friction material and dynamometer test results.

References
[1] S.J. Kim, H. Jang, Friction and wear of friction materials containing two different phenolic resins reinforced with aramid pulp, Tribol. Int. 33 (2000) 477.
[2] P.W. Lee, P. Filip, Friction and wear of Cu-free and Sb-free environmental
friendly automotive brake materials, Wear 302 (2013) 14041413.
[3] M.H. Cho, S.J. Kim, D. Kim, H. Jang, Effects of ingredients on tribological
characteristics of a brake lining: an experimental case study, Wear 258 (2005)
1682 258 (2005).
[4] L. Gudmand-Hyer, A. Bach, G.T. Nielsen, P. Morgen, Tribological properties of
automotive disc brakes with solid lubricants, Wear 232 (1999) 168.

[5] W.K. Lee, T.H. Rhee, H.S. Kim, H. Jang, Effects of Antimony Trisulde (Sb2S3) on
sliding friction of automotive brake friction materials, Met. Mater. Int. 19
(2013) 1101.
[6] M.H. Cho, J. Ju, S.J. Kim, H. Jang, Tribological properties of solid lubricants
(Graphite, Sb2S3, MoS2) for automotive brake friction materials, Wear 260
(2006) 855.
[7] E. Rukiye, Y. Nurettin, The effect of graphite, coke and ZnS on the tribological
and surface characteristics of automotive brake friction materials, Ind. Lubr.
Tribol. 63 (4) (2011) 245.
[8] S.J. Kim, M. Hyung Cho, K. Hyung Cho, H. Jang, Complementary effects of solid
lubricants in the automotive brake lining, Tribol. Int. 40 (2007) 15.
[9] R. Cox, Engineered Tribological Composites: The Art of Friction Material
Development, SAE International, USA, 2011.
[10] H. Jang, S.J. Kim, The effects of Antimony Trisulde (Sb2S3) and Zirconium
Silicate (ZrSiO4) in the automotive brake friction material on friction characteristics, Wear 239 (2000) 229.
[11] V. Matjka, Y. Lu, P. Matjkov, B. Smetana, J. Kukutschov, M. Vaculk, et al.,
Possible stibnite transformation at the friction surface of the semi-metallic
friction composites designed for car brake linings, Appl. Surf. Sci. 258 (2011)
1862.
[12] H. Jang, J.S. Lee, J.W. Fash, Compositional effects of the brake friction material
on creep groan phenomena, Wear 251 (2001) 1477.
[13] W. sterle, H. Klo, I. Urban, A.I. Dmitriev, Towards a better understanding of
brake friction materials, Wear 263 (2007) 1189.
[14] P. Filip, Z. Weiss, D. Rafaja, On friction layer formation in polymer matrix
composite materials for brake applications, Wear 252 (2002) 189.
[15] D. Varrica, F. Bardelli, G. Dongarr, E. Tamburo, Speciation of Sb in airborne
particulate matter, vehicle brake linings, and brake pad wear residues, Atmos.
Environ. 64 (2013) 1824.
[16] A. Iijima, K. Sato, K. Yano, M. Kato, K. Kozawa, N. Furuta, Emission factor for
antimony in brake abrasion dusts as one of the major atmospheric antimony
sources, Environ. Sci. Technol. 42 (2008) 29372942.
[17] A.M. Martinez, Echeberria J., Di Loreto A., Zanon M., Rampin I, EuroBrake,
EB2014-DF-002, 2014.
[18] W. sterle, I. Urban, Friction layers and friction lms on PMC brake pads, Wear
257 (2004) 215.
[19] W. sterle, A.I. Dmitriev, Functionality of conventional brake friction materials
perceptions from ndings observed at different length scales, Wear 271
(2011) 2198.
[20] R. Hinrichs, M.A.Z. Vasconcellos, W. sterle, C. Prietzel, A. TEM, Snapshot of
magnetite formation in brakes: the role of the disc's cast iron graphite
lamellae in third body formation, Wear 270 (2011) 365.
[21] W. sterle, I. Drfel, C. Prietzel, H. Rooch, A.- Cristol-Bulth, G. Degallaix, et al.,
A comprehensive microscopic study of third body formation at the interface
between a brake pad and brake disc during the nal stage of a pin-on-disc
test, Wear 267 (2009) 781.
[22] W. sterle, I. Urban, Third body formation on brake pads and rotors, Tribol. Int.
39 (2006) 401.
[23] R. Hinrichs, M.R.F. Soares, R.G. Lamb, M.R.F. Soares, M.A.Z. Vasconcellos, Phase
characterization of debris generated in brake pad coefcient of friction tests,
Wear 270 (2011) 515.
[24] P.J. Blau, Wear of truck brake lining materials using three different test
methods, Wear 259 (2005) 10221030.
[25] W. sterle, M. Griepentrog, T. Gross, I. Urban, Chemical and microstructural
changes induced by friction and wear of brakes, Wear 251 (2001) 1469.
[26] W. sterle, C. Prietzel, H. Klo, A.I. Dmitriev, On the role of copper in brake
friction materials, Tribol. Int. 43 (2010) 2317.
[27] P.W. Centers, F.D. Price, Tribological performance of MoS2 compacts containing sulfur, Sb2S3 Or Sb2S4, Wear 129 (1989) 205.
[28] G.M. Ingo, M. DUfzi, G. Falso, G. Bultrini, G. Padeletti, Thermal and microchemical investigation of automotive brake pad wear residues, Thermochim.
Acta 418 (2004) 61.
[29] O. Scarlat, Ph.d thesis: Thermal Phase-Equilibria in Oxidic Systems: SnO2Sb2O8-MexOy, 2002.
[30] M. Stan, S. Mihaiu, D. Crisan, M. Zaharescu, Subsolidus phase equilibrium in
the CuSbO System, Eur. J. Solid State Inorg. Chem. 35 (1998) 243.
[31] V. Ramakrishna Rao, K.P. Abraham, Kinetics of oxidation of copper sulde,
Metall. Mater. Trans. 2 (1971) 2463.
[32] A.R. Ginting, B.H. O'Connor, J.G. Dunn, X-Ray powder data for synthetic
Dolerophanite, Copper(II) Oxysulphate [Cu2O(SO4)], Powder Diffr. 9 (1994) 21.

You might also like