10.

Impact of Corrosion on Fretting Damage

of Electrical Contacts
T. Liskiewicz, A. Neville S. Achanta
School of Mechanical Engineering Department of Metallurgy and Materials Engineering
University of Leeds Katholieke Universiteit Leuven
Leeds, United Kingdom Leuven, Belgium
[email protected] [email protected]

Abstract—Electrical contacts are used in a large number of resistance faults. The process can be accelerated by corrosive
industrial applications, this includes all sorts of modern gas attack (typical for under-hood conditions) and increased
transportation: airplanes, trains and automobiles. Mechanical relative humidity. It has been estimated that up to 60% of all
assemblies are subjected to vibrations and micro-displacements car electric problems relate to degradation of electrical contacts
between mating surfaces are observed leading to fretting wear. by fretting [2].
Mechanical degradation can additionally be accelerated by a
corrosive factor caused by variable humidity, temperature and Current research in the field of electrical contacts durability
corrosive gas attack. Fretting-corrosion leads to an increase of is mainly focused on the following issues: three body tribo-
contact resistance or intermittent contact resistance faults as contact modeling with two rubbing members and insulating
corrosion products change the nature of the interface primary wear debris layer [3]; physical characteristics of vibration
through a range of film formation processes. In this work the conditions [4]; monitoring of fretting sliding regime as contact
impact of a corrosion product film formed on copper and gold damage is mainly related to gross slip condition [5]; impact of
surfaces on the electrical contact fretting behavior is shown. It variable environmental conditions: relative humidity and
has been observed that modification of the interface by the temperature [6]; fretting-corrosion resistance of connectors in a
formation of the surface layer can surprisingly lead to increase of DC 42 V electrical system [7]. A major scientific contribution
the electrical contact durability. expected from fretting corrosion research is to assure equal
durability of the electrical contacts with other mechanical parts
Keywords-connectors, fretting corrosion, sliding regimes,
of a system. For the automotive industry, it means that all
passive layer
connections must maintain low contact loop resistance as long
as the life of the vehicle to avoid power losses.
I. INTRODUCTION
In this work, the effect of corrosion on mechanical
Safety, but often also life, of vehicles’ users depends on degradation of electrical contacts is shown. Copper and gold
electrical contacts reliability. They are extensively used in all samples have been pre-treated in a corrosive media to build a
sorts of modern transportation: airplanes, trains and corrosion product layer on the surface. On gold, this is very
automobiles. The global tendency to equip all units with new thin layer and does not comprise what would normally be
electronic and computer control, steadily increases the number referred to as corrosion product. Unique behavior of the treated
of electrical contacts in a single vehicle. Some high surfaces has been revealed and analyzed.
specification vehicles have hundreds of meters of wires and
more than 400 connectors with 3000 individual terminals on
board [1]. Safety and comfort issues strictly rely on electronics II. METHODS
with systems like antilock braking, traction control, airbags or
navigation, so the integrity of electrical contacts is often A. Materials
crucial. The design of a typical electrical contact defines one Two types of materials have been selected for this study.
characteristic contradiction: connections must be detachable for Hard Drawn High Conductivity copper (HDHC, 99.9 wt.% Cu)
an installation and service, however during normal usage the is widely used in general electrical contact applications as it
separation of two parts is undesirable. Vibrations as well as combines excellent contact performance with good mechanical
thermal cycling in all transportation applications are common properties. A series of 10mm long samples has been cut off
and cannot be excluded, which causes small relative from the 5mm diameter rod. As a second material, male
displacements, typically tens or hundreds of micrometers components of the Lemo Company Series F electrical
between mated components of a connector. Two damage connectors have been used (Ø 1.5mm). Series F connectors are
mechanisms take place: (i) the top layer of a connector wears originally designed for demanding motorsport market, single
away and (ii) a passive film is formed at the interface. In the terminals are made of brass (UNS C 34500) and coated with
case of (i), pure mechanical damage is observed, while in the copper (0,5µm), nickel (3µm) and top layer of gold coating
case of (ii) also intensive corrosion processes are involved. The (1µm). Copper and gold surfaces have been tested in two
overall degradation process is called fretting-corrosion and states: (i) in the as-delivered clean state and (ii) after pre-
leads to an increase of contact resistivity or intermittent contact exposure to modify the near surface. The pre-exposure was

257
1-4244-0581-5/06/$20.00 © 2006 IEEE
conducted by immersion of samples in sea water (3.5% wt. • gross slip regime: with still higher displacement
NaCl) three times for 24 hours with 24 hour intervals in air. amplitude, the stick zone no longer occurs and the
entire contact area is subjected to sliding. In this
B. Test configuration regime, the wear mechanism takes place by debris
Fretting experiments have been carried out using the formation (by creating and breaking the adhesive
MUST tester manufactured by Falex [8]. A reciprocating junctions). The debris can also remain within the
sliding between samples arranged in crossed cylinders contact area as abrasive particles. The maximal
geometry has been applied (Fig. 1). The test program has been tangential force does not depend upon the sliding
carried out under 0.5N and 1N normal load (N), 10µm to 25µm amplitude and can be described by the classic
displacement amplitude (δ), and at the 1.2Hz to 2.5Hz Amonton’s friction law: Q=µP. The evolution of the
frequency (f). The normal load has been kept constant, while fretting log in this regime takes more quadratic shape
the friction force (Q) and displacement amplitude have been (Fig. 2c).
recorded. Tests have been conducted in a room environment at The above approach was then developed [11] by defining
the 20-22oC temperature and 45-50% relative humidity. The three fretting regimes for the running conditions recorded
four-wire measurement technique has been applied as a during the fretting test and the Running Condition Fretting
particularly accurate way to measure small resistances. The Map was proposed (Fig. 3):
value of 500mΩ has been assumed as a representative limit of
electrical resistance for electrical contacts used in the • partial slip regime (PSR) corresponding to the
automotive industry [9]. condition when stick-slip sliding regime is maintained
during the running test, observed for the smallest
C. Sliding conditions under fretting amplitudes;
Different damage mechanisms involved in fretting are • gross slip regime (GSR) corresponding to the condition
related to the sliding conditions at the interface and the when gross slip sliding regime is maintained during
following fretting regimes were distinguished [10]: running test, observed for the largest amplitudes;
• stick regime: the interfacial sliding between two bodies • mixed fretting regime (MFR) corresponding to the
does not occur as displacement is accommodated by condition when the gross slip and partial slip regimes
elastic deformations. The stick domain is maintained are present in one fretting test, observed for
by locked asperities that can be plastically shared in the intermediate amplitudes.
direction of micro-movement. Low fretting fatigue
damage can originate mainly by crack nucleation and III. RESULTS
propagation. For this regime, the evolution Q=f(δ),
called fretting log, takes form of closed linear relation Evolution of the contact resistance as a function of fretting
(Fig. 2a); time for different loading conditions, materials, with and
without the pre-exposure surface layer is depicted in Fig. 4.
• stick-slip regime: even though sliding does occur, the Because of space limitations, only selected sets of data are
stick zone is a dominating area in the contact. Surface presented here. For each experiment, representative shapes of
degradation is characterized by cracking as a result of the fretting loops from the beginning and end of the test are
contact fatigue particularly close to the stick-slip included to indicate the fretting regime of the tribo-couple.
boundary. If rough surfaces are subjected to contact,
The first row of graphs (Fig. 4a-c) gives results for
the stick zone can be spread for a number of single
copper/copper contact with clean surfaces and all three
contacts. The Q - δ curve takes a characteristic experiments are maintained within the gross slip fretting
elliptical shape hysteresis (Fig. 2b); regime. It is shown that increase of normal load from 0.5N to
1N (under equal displacement amplitude) is related to the
multimeter increase of electrical contact durability. At the same time, an
increase of displacement amplitude from 15µm to 25µm (under
equal normal load), significantly decreased the durability of the
V contact.

Cu/Cu Au/Au a) b) c)
Q Q Q
δ (t)

fretting contact δ δ δ
fixed

Figure 1. Schematic of experimental fretting assembly. Figure 2. Fretting log characteristics for different fretting regimes: (a) stick
regime; (b) stick-slip regime; (c) gross slip regime.

258
 The second row of graphs (Fig. 4d-f) presents contact
resistance curves for the same loading conditions as discussed
above, however with the passive layer formed on the copper
samples’ surface. It can be seen that modification of the
interface caused both extended durability of the contacts and
passage (in two cases) from the gross slip to partial slip fretting
regime. It is interesting, that after initial high resistance related
to the presence of the passive film followed by low resistance
of the contact after passive film removal, characteristic increase
of the contact resistance is not observed.
Figure 3. Running Condition Fretting Map [11].

(a) (b) (c)

0.5 0.5 0.5
contact resistance [Ω]

Cu/Cu, 0.5N, 15µm Cu/Cu, 1N, 15µm Cu/Cu, 1N, 25µm

contact resistance [Ω]

contact resistance [Ω]

0.4 clean surface 0.4 clean surface 0.4 clean surface

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0 0 0
0 2000 4000 6000 0 2000 4000 6000 0 500 1000 1500 2000 2500
fretting time [s] fretting time [s] fretting time [s]
(d) (e) (f)
0.5 0.5 0.5
Cu/Cu, 0.5N, 15µm Cu/Cu, 1N, 15µm Cu/Cu, 1N, 25µm

contact resistance [Ω]

contact resistance [Ω]

contact resistance [Ω]

0.4 passive layer 0.4 passive layer 0.4 passive layer

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0 0 0
0 3000 6000 9000 0 3000 6000 9000 0 3000 6000 9000
fretting time [s] fretting time [s] (i) fretting time [s]
(g) (h)
0.5 0.5 0.5
Au/Au, 1N, 25µm Au/Au, 0.5N, 25µm
contact resistance [Ω]
contact resistance [Ω]
contact resistance [Ω]

Au/Au, 0,5N, 25µm

0.4 clean surface 0.4 clean surface 0.4 passive layer

0.3 0.3 0.3

0.2 0.2 0.2

0.1 0.1 0.1

0 0 0
0 10000 20000 30000 0 10000 20000 30000 0 10000 20000 30000
fretting time [s] fretting time [s] fretting time [s]

Figure 4. Evolution of the contact resistance as a function of fretting time with the fretting loops corresponding to the beginning and the end of each test.

259
 The last row of graphs (Fig. 4g-i) depicts data for gold/gold TABLE I. SLIDING RATIO FOR CU/CU CONTACT WITHOUT
PASSIVE LAYER.
fretting contact with clean and passivated surface. In that case,
in spite of much longer test durations, building of the passive Displacement Sliding ratio e
layer does not take place independently of the loading amplitude δ(µm) P=0.5N P=1N
conditions and state of the sample surface. Increase of normal
10 0.43 0.34
load from 0.5N to 1N (under equal displacement amplitude)
caused a shift from gross slip to partial slip fretting regime. 15 0.65 0.52

It has been observed that presence of the passive layer was 20 0.87 0.69
conducive to the higher coefficient of friction (COF) values. 25 1.09 0.86
Comparison of mean COF for several fretting tests on clean
surfaces and coated with the passive layer is shown in Fig. 5.
Generally the gold/gold interface is characterized by higher (e>1) when all the surface of mutual contact is exposed to the
friction. It can also be noticed that in the case of copper, an atmosphere. The sliding ratio can be also applied to indicate the
increase of displacement amplitude is related to higher COF unexposed surface of the fretting contact area (Fig. 6). The
values. lower value of the e coefficient the bigger is the unexposed
fretted surface. Exposure of the tribo-contact to the atmosphere
IV. DISCUSSION has a key meaning in the oxidation process, as there is limited
air access to the contact surface layer as well as wear debris
A. Fretting contact exposure surface within the limits of unexposed surface. Sliding ratio values for
Bibliographical sources quote different values of copper contact under applied normal loads are listed in Table 1.
displacement amplitudes as a border between fretting and
reciprocating sliding motion. This value is variously interpreted B. Impact of loading conditions
and contained within a wide range of amplitudes between 50 One of the ways to increase the durability of the electrical
and 300µm [12-14]. This issue can be clarified by introducing contact is to limit the sliding at the interface and maintain the
the “e” coefficient [15], which is defined as a sliding ratio: stick-slip regime [5]. According to the Running Condition
Fretting Map theory, it can be achieved by increasing the
normal load. Higher normal load induces higher tangential
e=δ/a (1) force and changes the shape of the fretting loop by shortening
the sliding part of the hysteresis (Fig. 7). In fact, the actual
where: sliding amplitude is shorter under higher loads than the
δ – sliding amplitude, which is different from the displacement nominal one. Shorter sliding distance means that the unexposed
amplitude due to the contact and testing device compliance; surface is larger. In case of Cu/Cu tribo-couple under 15µm
a – Hertzian contact radius. displacement amplitude (Fig. 4a, b), the durability of the
The tribo-system remains in the fretting regime (e<1) when electrical contact under 0.5N normal load (e=0.65) is much
the unexposed surface is maintained at the centre of the fretted shorter than under 1N (e=0.52). For gold contact, increase of
surface. The system goes into the reciprocating sliding regime applied normal load caused even shift from gross slip to partial
slip regime for the same displacement amplitude (Fig. 4g, h).
0.7 Increase of displacement amplitude, as expected, leads to
reduction of the unexposed surface area and, as a result, shorter
mean coefficient of friction

0.6 lifetime of the electrical contact. Change of the amplitude from

15µm to 25µm, under 1N normal load (Fig. 4b, c), leads to the
0.5 increase of sliding ratio from 0.52 to 0.86.

unexposed
0.4 surface fretted surface
force normal : P

2δ
0.3 e=1

a fretted surface
0.2 fretting regime 2δ
e<1
0.1
a
reciprocating sliding
0
Cu/Cu Cu/Cu Cu/Cu Au/Au e>1
10µm 15µm 25µm 25µm
displacement amplitude : δ
Figure 5. Mean coefficient of friction for different fretting tests (P=0.5N).
Each first column – clean surface, each second column – surface with passive Figure 6. Definition of sliding ratio e and identification of transition between
layer. fretting regime and reciprocating sliding motion.

260
 Q

δ3N δ
δ0.5N

Figure 7. Identification of the actual sliding amplitude under two different

normal loads for Cu/Cu contact (clean surface, δ=25µm).

C. Impact of pre-exposure and surface film modifications

The sliding interface has been modified by the presence of
the formed passive film and durability of the electrical contacts
has been extended. An increase of the mean coefficient of
friction has been recorded for all samples with passive film.
Higher tangential force is again a factor responsible for change
of sliding conditions at the interface. In some cases, a shift
from gross to partial or from gross to mixed slip fretting
regime has been observed, compared to clean surfaces under
the same loading conditions. The passive film is worn out more
quickly under bigger displacement amplitudes.
The interesting question to pose is why the high resistance
state of the electrical contacts has not been reached after
degradation of the formed passive layer. The crucial factor
seems to be the incubation period of the sliding contact as the Figure 8. Generation and evolution of the third-body
at the contact interface [17].
first few hundred fretting cycles appear to determine the overall
degradation process of the mated surfaces. It has been shown • Durability of the electrical contacts is strictly related to
that metallic materials subjected to alternating sliding, tend to the sliding conditions at the interface. Sliding ratio can
generate a specific transformed layer on the top surface [16]. be applied to indicate the unexposed surface of the
This layer, called tribologically transformed structure (TTS), fretting contact area, which has an impact on intensity
has a particular nanocrystalline structure, corresponding to the of the oxidation process.
chemical composition of the primary material. Under
successive fretting cycles, TTS is fragmented and wear debris • After initial high resistance related to the presence of
is generated (Fig. 8). Wear debris is then subjected to the the passive film followed by low resistance of the
progressive oxidation process and, as a result, sliding surfaces contact after passive film removal, characteristic
are separated by a film of fully oxidized particles. passivation of the contact is not observed.
Clean copper surfaces follow this classical degradation • Higher normal load induces smaller sliding amplitude
mechanism and, under particulate loading conditions, the and increase of contact durability.
resistance of the contact has reached the critical value of
500mΩ. However, in the case of surfaces with a passive layer, • Larger displacement amplitude leads to reduction of
this process has been disturbed. Gold/gold contacts have not the unexposed surface area and, as a result, a shorter
shown the resistance increase and the degradation mechanism lifetime of the electrical contact.
is related in that case to adhesive wear promoted by higher • Increase of mean coefficient of friction has been
friction. established for all samples with passive film.

V. CONCLUSIONS The undertaken research program will be continued by

investigations directed towards understanding the mechanism
In this work, some preliminary results of the broader of electrical contact surface modification. The answer for the
research program on interaction between corrosion and fretting following question needs to be provided: how to benefit from
damage of electrical contacts have been presented. the fact that chemical treatment of the surface leads to an
Nevertheless, even this introductory study has revealed some increase of contact durability? Future challenges include wear
crucial results and indicated the importance of corrosion in debris analysis, focus on the initial state of the surfaces with
overall fretting-corrosion degradation process. The following characterization of the surface energy and fretting experiments
conclusions can be drawn: under wider spectrum of loading conditions.

261
 Gold-plated electrical contacts show outstanding fretting- [6] N. Saka, M.J. Liou and N.P. Suh, “The role of tribology in electrical
corrosion resistance but due to high costs, their application is contact phenomena”, Wear, 100 (1984), pp. 77-105.
limited in medium and low specification vehicles. With [7] K. Takahashi, O. Sakaguchi, T. Yamamoto and H. Inaoka,
“Investigation of contact materials in 42VDC automotive relay”, Joint
dramatically increasing prices of gold market rates during last 22nd Int. Conf. on Electrical Contacts & 50th IEEE Holm Conf. on
years, manufacturers would appreciate reliable design of Electrical Contacts, Seattle, USA, 2004, pp. 22-27.
cheaper electrical contacts. [8] S. Achanta, D. Drees and J.-P. Celis, “Friction and nanowear of hard
coatings in reciprocating sliding at milli-Newton loads”, Wear, 259
(2005), pp. 719-729.
ACKNOWLEDGMENT
[9] A.K. Rudolphi, et al., “Fretting testing of electrical contacts at small
The authors wish to thank the Falex Tribology NV displacement amplitudes – experience from a BriteEuram project”, in:
(Belgium) for providing the MUST tester for this project and Proc. of 20th Int. Conf. on Electrical Contacts, Stockholm, Sweden,
the Lemo Company (Switzerland) for the supply of gold-plated 2000, pp. 471-476.
samples of electrical contacts. [10] O. Vingsbo and S. Soderberg, “On fretting maps”, Wear 126 (1988), pp.
131-147.
[11] L. Vincent, Y. Berthier, M.C. Dubourg and M. Godet, “Mechanisms and
REFERENCES materials in fretting”, Wear 153 (1992a), pp. 135-148.
[12] J.S. Halliday and W. Hirst, “The fretting corrosion of mild steel”, Proc.
[1] G. Bolger, “The selection of automotive connectors”, Senior Thesis, Roy. Soc. A 236 (1956), pp. 411-425.
Coventry Univ., U.K., 1997. [13] N. Ohmae and T. Tzukizoe, “The effect of slip amplitude on fretting”,
[2] J. Swingler and J.W. McBride, “The synergistic relationship of stresses Wear 27 (1974), pp. 281-293.
in the automotive connector”, Proc. 19th Int. Conf. Electric Contact [14] S.C. Gordelier and T.C. Chivers, “A literature review of palliatives for
Phenom., Nuremberg, Germany, 1998, pp. 141-145. fretting fatigue”, Wear 56 (1979), pp. 177-190.
[3] C.E. Heaton and S.L. McCarthy, “High cycle fretting corrosion studies [15] S. Fouvry, Ph. Kapsa and L. Vincent, “Quantification of fretting
on Tin-coated contact materials”, Proc. 47th IEEE Holm Conf. on damage”, Wear 200 (1996), pp. 186-205.
Electrical Contacts, Montreal, Canada, 2001, pp. 209-214. [16] P. Blanchard, “Usure induite en petits débattements: TTS d’alliages de
[4] G.T. Flowers, F. Xie, M.J. Bozack and R.D. Malucci, “Vibration titane”, Thèse Ecole Centrale de Lyon, No 91-32 (1991).
thresholds for fretting corrosion in electrical connectors”, IEEE [17] E. Sauger, S. Fouvry, L. Ponsonnet, Ph. Kapsa, J.M. Martin and L.
Transactions on Components and Packaging Technologies, 27 (2004), Vincent, “Tribologically transformed structure in fretting”, Wear 245
pp. 65-71. (2000), pp. 39-52.
[5] S. Hannel, S. Fouvry, Ph. Kapsa and L. Vincent, “The fretting sliding
transition as a criterion for electrical contacts performance”, Wear, 249
(2001), pp. 761-770.

262

View publication stats