Fretting corrosion.pdf

Tribology International 32 (1999) 587–596

www.elsevier.com/locate/triboint

Fretting corrosion of materials for orthopaedic implants: a study of

a metal/polymer contact in an artiﬁcial physiological medium

Bruno Tritschler * , Bernard Forest, Jean Rieu

D´partement M´canique des Biomat´riaux et Traitements de Surface, Ecole Nationale Sup´rieure des Mines de Saint-Etienne, 158, Cours

Fauriel, 42023 Saint-Etienne Cedex 2, France

Received 20 May 1999; received in revised form 15 November 1999; accepted 16 November 1999

Abstract

The fretting corrosion behaviour of a 316L SS ﬂat against a PMMA counterface has been investigated in an artiﬁcial physiological

medium. A speciﬁc device has been used to visualize the in situ degradation at the contact interface. Simultaneous analysis of the

coefﬁcient of friction and free corrosion potential has shown four distinct stages during fretting experiments. An energy-oriented

approach to the fretting process was conducted in tandem with measurement of wear. This method has shown a linear progression

in the wear volume of the samples as a function of the interfacial energy dissipated during fretting. The presence of chlorides

contributes to a considerable acceleration of the degradation of the stainless steel surface. This process was explained by a mechanism

related to crevice corrosion activated by friction.

Keywords: Fretting corrosion; Wear; 316L SS; Orthopaedic implants

1. Introduction

trochemical reactions with the environment predominate,

are known as fretting corrosion [8,9].

The biomedical ﬁeld is not immune to these phenom-

ena. Thus, the mechanical devices used in orthopaedic

surgery are subject to fretting [10,11]. These mech-

anisms have been used for many years in assemblies

such as screw/plate connections [12,13]. In association

with this phenomenon, problems of fretting corrosion in

total hip joint prostheses have been observed for several

years [14–16]. The most frequently employed method

for ﬁxation is by cementing the femoral stem into the

bone using a polymer cement, poly(methylmethacrylate)

(PMMA). Fig. 1 shows the mode of implantation of

such prostheses.

Three groups of materials are widely used at present:

316L austenitic stainless steels, cobalt–chromium alloys

and Ti–6Al–4V titanium alloys, much used since the

1970s because of their good mechanical properties com-

bined with high biocompatibility [17].

Small movements develop either immediately on

implantation or after a few years of service. These small

amplitude movements are all the more dangerous for

occurring in an environment rich in chlorides. Instances

of loosening of the prosthesis can be explained in part

by the great difference in elastic modulus between the

When two surfaces in contact are subjected to an

oscillatory movement of low amplitude, surface deterio-

ration generally appears [1–4]. The phenomena of wear

induced by friction over small displacements, often

designated by the term fretting , can develop in any

mechanical device subjected to alternating movements,

such as vibrations or oscillations. Examples include

bolted assemblies, cables [5], rotors, riveted joints or

electrical contacts.

Damage by fretting involves all kinds of materials:

metals, ceramics [6] or polymers [7]. The most important

consequence is a drastic reduction in the lifetime of

mechanical systems, due to jamming of the parts in con-

tact or, in contrast, the appearance of play.

Environmental effects are of overriding importance in

the phenomena of fretting. Some corrosive media, such

as chlorides, contribute to a considerable acceleration of

wear. The processes involved, in which chemical or elec-

* Corresponding author. Tel.:

+33-4-77-42-02-51; fax:

+33-4-77-

42-00-00.

E-mail address: tritschl@sms.emse.fr (B. Tritschler).

0301-679X/99/$ - see front matter

PII: S 0301-679X(99)00099-7

588

B. Tritschler et al. / Tribology International 32 (1999) 587–596

a UHMWPE hip prosthesis cup was fretted against a

316L SS sphere in Ringer’s solution, in a conﬁguration

similar to the in vivo relationship between the spherical

head and the acetabular cup, a phenomenon of crevice

corrosion activated by the friction developed at the sur-

face of the metal, and considerable abrasion of the poly-

mer was observed.

The purpose of the present work was to study the

phenomena of wear by fretting corrosion of a metal,

316L stainless steel, in contact with a polymer,

poly(methylmethacrylate) (PMMA). The tests, per-

formed in an artiﬁcial physiological medium using a

device for in situ visualization of the contact surface,

led to the development of a qualitative and quantitative

approach to fretting corrosion.

Fig. 1. Diagram showing cemented implantation of a total hip joint

prosthesis.

2. Materials and methods

polymer and the metal femoral stem, and in part by

imperfect initial ﬁxation resulting from the shape of the

prosthesis [18]. Fretting corrosion then causes the forma-

tion of debris in the form of a mixture of particles of

metallic oxides and polymer. The direct consequences

are inﬂammatory reactions of the surrounding bony

tissues and the necessity for a second surgery. Examples

are metallosis cases which are due not only to the cor-

rosion of the prosthesis in service, but more likely to the

emission and the effects of wear particles produced by

fretting [14]. The analysis of prostheses which have been

removed does not always reveal the actual reasons for

their failure. It was therefore necessary to develop test

methods to study the fretting behaviour of materials used

for implants. One solution consists in reproducing as

faithfully as possible the functioning of a hip joint pros-

thesis by means of walk simulators. Such devices, based

on the average behaviour of the patients, are difﬁcult to

use and costly to create. Thus the need to develop tech-

niques of in vitro investigation led to the design of

devices capable of producing wear faces similar to those

observed on explanted prostheses, by means of simple,

rapid and reproducible tests. Many previous studies have

indeed examined the problems of deterioration of ortho-

paedic implants in general: experimental techniques have

been created to reproduce it and surface treatments

devised to remedy it [19–22]. Conﬁnement appears to

be an essential factor in the deterioration. In the case of

large amplitude displacements, to which the term “fret-

ting” cannot be applied, the degree of deterioration is

observed to vary according to the degree of conﬁnement.

Earlier studies were carried out on a pin made of

UHMWPE and a disc of 316L SS in an artiﬁcial physio-

logical solution. These showed almost no damage to the

surface of the metal [21], while slight polishing of the

surface of the polymer had occurred. In contrast, when

Cylindrical polymer specimens were rubbed against

stainless steel ﬂats. Samples of PMMA were taken from

cylindrical bars of 20 mm diameter. They took the form

of blocks with a rectangular base of 9 mm

15 mm, with

10 mm. The PMMA

used was optically transparent (Altuglass

). The plane

surface opposite the contact interface was polished to

enable observation of the polymer/metal interface

through the transparent block. PMMA exhibits a glassy

transition at T g =399 K and a sub- T g (‘b’) transition at

T b =226 K. At room temperature, it is a hard, brittle

material.

The stainless steel used was the 316L SS (Z2CND17-

12). It is a high purity steel as regards inclusions and

very low in carbon. It was supplied rapid-quenched after

hot rolling. Its chemical composition is shown in Table

1. It is in conformity with the standards for manufacture

of orthopaedic implants (ISO standard 5832-1).

The metal samples were 10 mm´10 mm´20 mm par-

allelepipeds. The friction surface was ﬁrst polished with

diamond paste to 1

Before each test, the metal and polymer samples were

cleaned for 10 min in ethyl alcohol in an ultrasonic cle-

aner and then dried in air. The mechanical characteristics

of the materials are shown in Table 2.

The fretting tests were conducted on a TRIBOM-

INES

m to a roughness of Ra

0.03

device installed on a SCHENCK-PSA 10 servo-

hydraulic traction–compression machine. The cylindrical

sample of PMMA was mounted on a ﬁxed component

provided with a conical aperture through which the con-

tact surface can be observed. The plane sample of 316L

SS was ﬁxed in a ball joint, enabling the parallelism of

the surfaces in contact to be ensured. The length of the

cylinder/plane contact was 15 mm. Once adjusted, the

ball joint was ﬁxed immobile relative to the hydraulic

cylinder by a clamping screw. The whole experimental

set-up is shown in Fig. 2.

one cylindrical surface of radius R

B. Tritschler et al. / Tribology International 32 (1999) 587–596

589

Table 1

Chemical composition of 316L SS stainless steel (% w/w)

17.3

13.7

2.68

0.029

0.63

1.70

0.025

, 0.01

Bal.

Table 2

Mechanical properties of the materials used (suppliers’ data)

Poisson’s ratio n

Young’s modulus E

Yield stress R c

Ultimate tensile strength

(GPa)

(MPa)

PMMA

0.35

3.5

316L SS

0.3

200

280

640

Fig. 2. TribominesÒ

fretting set-up.

m. The selected cyclical frequency

of 1 Hz is comparable to the average walking frequency

of the human body. The shape of the cycles is sinusoidal.

The movements were monitored in the immediate vicin-

ity of the contact by means of inductance sensors situ-

ated on either side of the device. A pre-stressed piezo-

electric load-cell was installed on the ﬁxed component

in order to measure the tangential force F T . The contact

interface between the two samples was displayed on a

40 to

100

monitor screen via a CCD camera linked to an optical

system. The metal sample was connected to a TACUS-

SEL PGP 201 potentiostat and a saturated calomel refer-

ence electrode (SCE) was placed close to the contact in

order to monitor changes in corrosion-free potential dur-

ing the test. A system of data acquisition and processing

was set up to study changes in the coefﬁcient of friction,

to record the F T –displacement cycles (or F T – D cycles),

and to make electrochemical measurements throughout

the test. It must be noted that the shape of the fretting

cycles obtained was, for all the tests conducted, charac-

teristic of a gross slip regime within the contact. The

fretting tests were performed in a Ringer’s solution in

The normal load F N , applied using calibrated masses,

lay within the range 0.7–6 N/mm, in order to remain

in the elastic domain of the polymer. The displacement

amplitude imposed on the metal sample lay within the

range

590

B. Tritschler et al. / Tribology International 32 (1999) 587–596

order to simulate human physiological ﬂuids. This sol-

ution was prepared in the laboratory to the following

formula: NaCl to 8.5 g/l, KCl to 0.25 g/l, CaCl 2 to 0.22

g/l, NaHCO 3 to 0.15 g/l, with pH value 7.8±0.1. The

volume of material removed during the test was evalu-

ated by means of a proﬁlometric technique. After the

test, the two metal and polymer samples were carefully

cleaned in ethanol in an ultrasonic cleaner in order to

remove all traces of debris from the surface. After dry-

ing, a TALYSURF proﬁlometer took the proﬁles at reg-

ular intervals across the entire friction surface. The vol-

ume of material worn away was then calculated by

summing all the elementary volumes thus measured.

Calculation by this method provides a more accurate

assessment of the volume of abrasion than the weighing

method, which would only yield an approximate value

due to the small quantities of material worn off by fret-

ting.

The cell containing the stationary Ringer’s solution

was left open at room temperature in the laboratory

atmosphere. Once the samples were placed in the sol-

ution and after application of the load, a 30 min wait

ensued before the fretting test was started in order to

allow the free potential to stabilise.

dispersion X-ray analyses reveal mainly iron and chro-

mium oxides. The surrounding reddish deposit consists

primarily of Fe 2 O 3 iron oxides.

The PMMA shows an accumulation of ﬁne debris at

the limits of the contact, on both sides of the friction

zone (Fig. 3). Orange-coloured deposits similar to those

observed on the 316L SS sample are also seen all around

the friction surface, although in smaller quantities. These

are the corrosion products resulting from the electro-

chemical process occurring on the surface of the metal

and little by little transferred onto the surface of the

PMMA during the test.

3.2. Evolution in coefﬁcient of friction and free

corrosion potential

Four stages were observed in coefﬁcient of friction or

free corrosion potential changes during the fretting tests.

Fig. 5 shows a typical example of the curves obtained

for the tests performed in Ringer’s solution. For more

clarity, the evolution of both coefﬁcient of friction and

free potential were plotted using a logarithmic abscissa.

The ﬁrst stage, which is very rapid, corresponds to a

reduction in the coefﬁcient of friction during the early

fretting cycles, associated with a drop in free potential.

This initial fall is followed by a slower decrease of the

coefﬁcient of friction, while the corrosion potential rises

again slightly. The second stage corresponds to an

increase in the coefﬁcient of friction associated with a

decrease of free potential, though this decrease is less

abrupt than that observed at the beginning.

The ﬁrst few cycles of the test cause the collapse of

the layer of oxide initially formed on the surface of the

metal, after which strong interaction is observed between

the two surfaces now laid bare. This marks the beginning

of the damage to the surface of the PMMA and the cre-

ation of the ﬁrst debris which will cover the surface of

the metal. Those particles act as a steady-state lubricant

ﬁlm between the counterfaces since the tangential force

is applied. This pseudo repassivation causes a new rise

in potential with a drop in the coefﬁcient of friction due

to the accommodation of the movement by this new

3. Experimental results

3.1. Observation and analyses of the surfaces

The fretting tests carried out in Ringer’s solution led

to considerable damage to both surfaces. Fig. 3 shows

the appearance of the friction zone on the samples,

observed using a JEOL-840 scanning electron micro-

scope. In both cases, numerous parallel scratches in the

direction of friction are observed. Visual examination of

the friction zone on the 316L SS shows deep traces of

wear, shiny and scratched in appearance. Around the

edge of this zone, a wide orange-red deposit is observed,

revealing considerable corrosion phenomena. Fig. 4

shows debris resulting from wear located at the edge of

the friction zone at the surface of the 316L SS. Energy

Fig. 3. SEM observation of wear surfaces (frequency 1 Hz, displacement

±40

mm, normal force 4 N/mm, duration of test 20,000 cycles, Ringer’s

solution). (a) View of surface of 316L SS. (b) View of surface of PMMA.

B. Tritschler et al. / Tribology International 32 (1999) 587–596

591

m, normal force 4 N/mm, duration

of test 20,000 cycles, Ringer’s solution). Right: EDX analysis of debris; presence of oxides of iron and chromium.

Fig. 5. Coefﬁcient of friction and free corrosion potential as a function of number of cycles. Contact PMMA Ø20/316L SS plane in Ringer’s

solution. Frequency 1 Hz, displacement

freshly formed third body. However, the cleansing action

of the surrounding aqueous medium causes the debris to

be carried away more or less rapidly to the edge of the

friction area. A ﬂow of debris, gradually carried away

from the contact surface, builds up. The common dam-

age to the two surfaces results in a regular decrease in

corrosion potential to values which are more negative

(less “noble”)—showing increased electrochemical

activity of the stainless steel surface—together with an

increase in the coefﬁcient of friction. It is pertinent to

remark here that during the ﬁrst few hundred cycles, any

variation in the coefﬁcient of friction is immediately fol-

lowed by a change in free potential, irrespective of the

operating conditions applied. This linked progression of

both values has already been shown in previous stud-

ies [22].

By in situ observation it is possible to relate the differ-

ent values measured to the phenomena of deterioration

of the polymer/metal interface. Damage to the surfaces

is visible from the very ﬁrst cycles. The consequence of

the formation of PMMA debris is the appearance of a

white band which adheres to the surface of the metal

before gradually disappearing. During the period of rise

of the coefﬁcient of friction, increasing deterioration of

the contact interface is observed, revealed by the appear-

ance of scars parallel to the direction of movement. An

increasing presence of ﬁne, snowy debris at the edge of

the contact zone may also be observed.

The third stage of development corresponds to a slow

decrease in the coefﬁcient of friction. During this period,

the presence in situ of increasingly marked scarring is

seen on the surface of the 316L SS, together with the

Fig. 4. SEM observation of wear debris on the surface of the 316L SS (frequency 1 Hz, displacement

m, normal force 4 N/mm. 1, breakdown of the oxide layer initially formed and pseudo repassivation;

2, ﬂow of debris—deterioration of the contact interface; 3, degradation of the stainless steel surface; 4, equilibrium in the tribosystem.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: