POSSIBILITIES OF FRETTING CORROSION MODEL TESTING FOR CONTACT SURFACES OF AUTOMOTIVE CONNECTORS I. BURESCH Wieland Werke AG, D 89070 Ulm, GERMANY; e-mail: isabell.buresch@wieland.de P. REHBEIN Robert Bosch GmbH, Postfach 106050, D 70049 Stuttgart, GERMANY; e-mail: peter.rehbein1@de.bosch.com D. KLAFFKE BAM, Unter den Eichen 87, D 12200 Berlin, GERMANY; e-mail: dieter.klaffke@bam.de SUMMARY The number of electrical contacts in automotive applications increases steadily . The trends of miniaturisation in all fields of technique and the increasing demands concerning functionality, mechanical and thermal load require the per-manent development of connectors. An improvement of electrical connectors can be achieved by coating of the connec-tor elements. Additionally, the price of improved connectors has to remain as low as possible. One of the most impor-tant degradation processes that limits electrical functionality is fretting corrosion, caused by vibrations. The de-velopment of better coatings requires a check of functionality at a very early state of development. For this purpose different test facilities are available, that be used in order to qualify the improvement. In model tests with simple speci-mens made of original connector materials the electrical degradation mechanisms caused by friction and wear processes can be studied. It is shown for tin- and gold-coatings that in tribometer tests under certain parameters wear appearances and mechanisms can be similar to that in product testing. Thus, model testing methods can help to reduce time-to-market.

Keywords: connector coatings, fretting corrosion, model testing, contact resistance, contact failure, coating

1 INTRODUCTION Fretting corrosion [1], [2] due to microscopic relative motion of the connecting parts is one of the most impor-tant failure mechanism for separable electrical contacts. Especially tin coated contacts show this phenomenon [3]. As a result of these micro movements, the oxides near the contact spot can be permanently broken up and each time fresh material is exposed to the atmosphere. This leads to an add on of oxidized wear debris in the contact zone which eventually causes an intolerable increase of the contact resistance. The trend towards miniaturisation of connectors with a steadily increasing number of contacts in one connector strongly limits the use of high contact forces. Therefore, the possibilities for micro motions in the connector increase and the relationship between contact force and displacement amplitude on fretting corrosion becomes more and more of importance. 2 OBJECTIVES Fretting damage will be influenced by the contact de-sign, atmosphere, contact surface, contact force, dis-placement amplitude, lubrication and frequency. The aim of model testing is to evaluate the number of cycles to failure in dependence of the influencing parameters like contact force, sliding amplitude, atmosphere, tem-perature and contact geometry. Model tests are usually “low cost” tests and can provide preliminary informa-tion about the suitability of new developed materials and/or coatings. The transfer of the results from labora-tory tests to practical applications, however, is always risky and has to be proved by additional field oriented tests.

In this paper the results of different test methods are analysed with respect to their practical relevance.

A field oriented test method that provides informa-tion about the suitability of the connector modifica-tion,

A tribological model test that provides information of the basics of electrical and tribological degrada-tion processes.

In the automotive industry the need for reliable, small connectors for new high-stressed connector applications like CRI-systems (Common-Rail-Injection) is high. In order to develop new connectors by simultaneous engi-neering to save time and money, model testing methods with a good correlation to the product testing results represent a fit tool for the new challenges [4],[5]. 3 EXPERIMENTAL

3.1 Test methods

3.1.1 Model tests In different test-rigs periodic relative displacements between the connecting model parts were applied while the contact resistance and the friction force were moni-tored continuously. The difficulties are the transforma-tion of the model test results into practice. Investiga-tions and comparison on the wear scars (contact spots) from model testing samples and real connectors from the field by SEM (scanning electron microscopy) enable the evaluation of the possibilities and limits of fretting corrosion model testing.

Model tests are performed with cylindrical shaped specimens, made of coated bronze strip material pro-duced by a punch. This contact geometry is similar to a lot of real connectors in car applications with male and curved female contact spots. Two of these specimens are tested in a "crossed cylinder" configuration as shown in Figure 1 [7].

Fn

2d

Figure 1: Crossed cylinder test configuration The couple of specimens is fixed in a test rig, described in more detail in [5], and is loaded with defined values of stroke, frequency and normal force to a number of cycles between 104 and 106 under dry circuit conditions. All parameters are kept constant during the test. In each test the friction force and the electrical contact resis-tance was measured and recorded. The tribometer works inside a climate chamber, the relative humidity can be varied in the range from 5 % R.H. up to 100 % R.H. Furthermore, the chamber can be evacuated and tests can run in vacuum as well as in inert gases [6]. An elec-trical voltage can be applied to the tribo contact, allow-ing the study of the affect of electrical current on tribo-electrical degradation of connectors.

The volumetric wear of the system is determined after the test from the size of the nearly circular wear scars (Figure 2) with Eqn. (1):

Wv = π∗d4/64∗R (1) where d is the mean value of the wear scar diameter, determined at both specimens and R = radius of the bulge (2.5 mm).

3.1.2 Shaker tests Tests on real connectors usually are performed by a shake test system consisting of a electrodynamic excited shaker table. Different types of vibration modes exist: sinus, random and sinus-on-random. Investigations of design resonances can be done by sweeping sinus fre-quencies, whereas sinus plus random is often used to simulate the real field conditions. These tests are very time consuming and expensive. Therefore there is a high need for rapid testing in a first step of releasing a product. The vibration loading is up to accelerations of 100 g (1 g = 9.81 m/s2) and depends on the demands. Changes of temperatures and atmospheres can be super-imposed. During the tests the electrical functionality will be checked continuously to recording both interrup-tions and an increase of the contact resistance. In prac-tice, tests were running to fulfil the car manufacturer's specifications or standards.

3.1.3 Materials and testing conditions The crossed cylinder samples were hot dip tinned (hdt) with 1-3 µm Sn in a usual production line before punch-ing . The gold coated samples were electroplated with 0.6µm to 1.4 µm Au over 1.0 µm to 2.5 µm Ni after punching. This process is typical for the use of hot dipped tin copper alloy strips used in the connector industry and also typical for gold coated connectors which were plated after stamping. 3.2 Results

3.2.1 Model tests The advantage of model tests is the possibility to evalu-ate the influence of individual parameters of the test conditions. As an example concerning the influence of humidity, Figure 2 shows the results from two tests with Sn coated specimens in dry air (5% R.H.) and in moist air (100 % R.H.). Figure 3 shows the evolution of elec-trical contact resistance in three tests with Sn (hdt)/ Sn (hdt), running in dry and moist air, respectively.

200 µm 200 µm 5 % R.H. 100 % R.H.

Figure 2: Optical micrograph of a wear scar on Sn coated sample after tests 2d=10 µm, Fn =2 N, ν = 10

Hz, n=10 000, T=24 °C

The wear scars are in both cases irregularly shaped, pointing a plastic deformation, adhesive processes and oxidation at the surface of the soft coatings. The wear scar in moist air is slightly bigger than in dry air. The evolution of contact resistance (Figure 3), however, shows an increase of number of cycles to an electrical degradation of the contact with increasing humidity.

0

100

200

300

400

500

0 5 10 15 20Number of cycles n [103]

Elec

trica

l res

ista

nce

[m Ω

]

5 % R.H. 50 % R.H.100 % R.H.

Figure 3: Evolution of electrical resistance in tests with hdt/hdt in dry, normal and moist air, Fn=2 N, ν=10 Hz,

n=100 000, T=24 °C Another example demonstrates the influence of coating thickness on the "life time" of a coating in a fretting test. Figure 4 shows the wear scars on two Au coated speci-mens with different thicknesses of Au and Ni underlayer after testing in normal atmosphere.

200 µm

Au: 0.6 µm / Ni: 1.5 µm Au: 1.4 µm / Ni: 2.6 µm

Figure 4: Optical micrograph of a wear scar on Au coated sample after tests with 2d=20 µm, Fn=2 N,

ν=10 Hz, n=1'000'000, T=24 °C, R.H. = 50 %

0

100

200

300

400

500

0 200 400 600 800 1000Number of cycles n [103]

Elec

trica

l res

ista

nce

[m Ω

]

0.6 µm Au- 1.5 µm Ni

1.4 µm Au- 2.6 µm Ni

Figure 5: Evolution of electrical resistance in tests with Au/Au with different coating thickness with 2d = 20 µm,

Fn=2 N, ν=10 Hz, n=10 6, T=2 °C, R.H.=50%

SEM micrograph

Elemental distribution of:

Ni

Cu Sn

Au

Figure 6: SEM micrograph of a wear scar on Au- coat-

ing and elemental distribution The size of the wear scars is similar, however, the deg-radation of resistance, Figure 5 occurs much earlier for the "thin" coating than for the "thick" one. The analysis of the wear scar by EDX in a SEM shows for the Au distribution, that the coating material was perforated in

the middle of the scar, Figure 6, where Cu of the sub-strate is dominating. At the rim of the scar Ni can be detected that was used as an underlayer to improve the adherence of the Au coating, to avoid the diffusion of Cu in the Au-layer and to give a dense coating.

3.2.2 Shaker tests Presented results of connectors are of pin-type 1.5mm (Au) and 2.8mm (Sn) and were tested in a typical auto-motive specification applying first 50 temperature cy-cles (-40°C to 130°C (Sn) / 140°C (Au)) followed by a sinus-on-random at 30 geffective for 3 times 100 hours in each direction in air at 130°C (Sn) and 140°C (Au).

1 mm Figure 7 shows a SEM micrograph of a Sn coated con-

tact (1-3 µm) surface after a shaker test The worn surface is of elliptical shape due to the geometry of the Sn coated connector (Figure 7). Plastic deformation and adhesion processes are the dominating and for tin typical wear mechanisms. At the scar rim more oxidative wear is accumulated and led to darker spots. This corresponds to EDX analysis that proved such dark fields as tin-oxides. In the middle of the wear scar the tin-coating is locally worn through onto the intermetallic phase. The striations are the result from micro motions and adhesive wear for tin in which small wear fragments grow by adhesive removal of a surface layer of tin in the manner of a rolling snowball. During the testing the electrical resistance started to increase slightly.

1 mm Figure 8: SEM micrograph of an Au coated contact (1.4

µm) surface after a shaker test. The ring formation of the wear scar (Figure 8) is typical for partial slip mode, where fretting is not responsible for failure. This wear scar with the sticking zone was electrically stable during the vibration test, because in the centre the coating was not destroyed. This centre is typical for the sticking zone and indicates a kind of mild abrasion like it occurs in polishing wear. At the rim of

the wear scar the coating is locally worn through (EDX) or even cracked off by fatigue and adhesion by the coat-ing of the corresponding pin. The black areas indicate an oxidation of the Ni-underlayer and contamination like debris, oil or dust that can be trapped there. 4 DISCUSSION The wear appearance of the tin specimen tested in air of different humidity as shown in Figure 2 is characterised in the optical micrographs by a black wear area indicat-ing that the free tin is worn through. The accumulation of wear debris (oxidised tin) leads to an increasing of the electrical resistance. Thin quasi-liquid films of ad-sorbed water seems to act as a lubricant.

The comparison with the tin contact tested on a shaker (Figure 7) reveals similarity to the wear mechanisms occurring during the model tests. In both cases plastic deformation, adhesive process and an increasing oxida-tion are the dominating wear mechanisms.

The coatings of the Au specimen tested in the tribome-ter are also worn through onto the Ni-underlayer showed by dark wear zone in the optical micrographs confirmed by the EDX study. For Au, plastic deforma-tion and adhesion are also the controlling processes superimposed by oxidation of the underlayer. The uni-form appearance of the wear area is in accordance with the huge number of 106 cycles and because of the gross-slip micro motions.

At the rim of the wear scars of the shaker tested connec-tor, the location of the micro motions, similar wear processes occurred (Figure 8) and led also to a local worn through and oxidation of the Ni-underlayer and the base material. For wear analysis of contacts this circumstance has to be considered. 5 SUMMARY Typical coatings of automotive connectors were tested and analysed in model tests with model samples. Differ-ent coatings show different wear appearances according to both plastic deformation and wear mechanisms like adhesion, chemical reaction, polishing abrasion and fatigue. Often the critical tribo-oxidation occurs and leads to an increasing contact resistance. In comparison to real connectors the wear mechanism and the wear scar look similar. This shows that during model testing the typical fretting conditions can be simulated in a time-saving manner with test samples like crossed cyl-inders. Additional tests show that also test sample con-

figurations like rider/flat or ball/flat lead to comparable results. An explanation is that in all cases a Hertz con-tact exists with similar radii and that the displacement amplitudes are in a comparable range for both tribome-ter and connector. By means of model testing the influ-ence of individual operational parameters can be studied and critical conditions can be evaluated in a time saving way. 6 ACKNOWLEDGEMENTS The financial support of the work by the European Community in the frame of a Brite/Euram project (BE96-3188) is gratefully acknowledged. Thanks are due Mr. A. Schliessus (MPA Stuttgart) and Mr. T. Klaas (Bosch), Mrs. P. Seipt, Mrs. A. Krause, Mr. M. Hartelt and Mr. J. Schwenzien (all BAM) for technical assis-tance. 7 REFERENCES [1] Antler, M.: Electrical effects of fretting connector contact materials: A Review, Wear 106 (1985) 5-33 [2] Waterhouse, R. B., Lindley, T. C. (Eds.): Fretting Fatigue. ESIS 18, Mechanical Engineering Publications Limited, London (1994), 530 pages [3] Buresch, I. :Thin and low alloyed tin-coatings; 32nd ISATA, Vienna, Austria 1999. [4] R. B. Waterhouse: Fretting Corrosion, Pergamon Press, 1972 [5] Klaffke, D. and M. Hartelt: Investigations on fret-ting performance of connector materials by model tests. Tagungsband 19. Internationale Tagung über elektrische Kontakte, 14.-17. Sept. 1998, VDE-Verlag Berlin und Offenbach, 181-185 [6] Klaffke, D., Hartelt, M. and R. Wäsche Non-oxide ceramics as tribo materials – analysis of wear mecha-nisms under oscillating sliding conditions in various atmospheres. 12th Internat. Colloquium Tribology 2000-Plus, 11.-13.01.2000, (ed.: W. J. Bartz), TA Esslingen, Vol. III, 1755 – 1763 [7] Å. Kassman-Rudolphi, O. Vingsbo, K. White, L. Deneuville, J.-P. Célis, Ph. Kapsa, S. Hannel, S. Fouvry, D. Klaffke, P. van Dijk, J. Horn, I. Buresch, G. Ide, F. Paelinck, P. Rehbein, J. Schoefer, B. Blomberg, G. Liraut: Fretting testing of electrical contacts at small displacement amplitudes – experience from a Brite/Euram project, Proc. ICEC conference 2000, Stockholm, p. 471-476, ISBN 91-7170-585-6