Tribological behaviour of ion-implanted Ti6A14V sliding against polymers

Tribological behaviour of ion-implanted Ti6A14V sliding against polymers

WEAR ELSEVIER Wear 209 (1997) 49-56 Tribological behaviour of ion-implanted Ti6AI4V sliding against polymers H. Schmidt ~'*, A. Schminke a, D.M. Rii...

562KB Sizes 0 Downloads 4 Views

WEAR ELSEVIER

Wear 209 (1997) 49-56

Tribological behaviour of ion-implanted Ti6AI4V sliding against polymers H. Schmidt ~'*, A. Schminke a, D.M. Riick b Facul~' of Materials Science, Departmentof Physh'al Metallurgy, Technical Universityof Darmstadt. Petersenstrasse23. D-64287 Darmstadt, Germany GSI Centerfor Heavy Ion Research. Planckstrasse I. D-64291 Darmstadt. Germany

Received 30 May 1996; aecept~d6 November1996

Abstract The effect of ion implantation of various elements (C, N, O, Y, HI', Pt, Au) with different energy-dose combinations on the tribological behaviour of the alloy Ti6AI4V was investigated. Wear tests were performed with flat alloy specimens sliding against PMMA pins and UHMWPE discs, yielding the following re~*!!s. Th.e we~ resis:~,ce of Ti6AI4V againsl PMMA increases with the microhardness in the implanted region. Wear and friction of Ti6AI4V/UHMWPE couples are decreased by implantation of noble metals as well as nitrogen or carbon. The removal of oxide particles from the alloy surface is greatly reduced, owing to the formation of titanium compound precipitates. The most pronounced wear reduction occurs when the enrichment of implanted atoms starts directly below the oxide film on the titanium surface, followed by a continuous decrease of the concentration of implanted atoms with increasing depth. © 1997 Elsevier Science S.A. Ke)words: Ion implantation;Titanium;Joint prosftheses;Backscaueringspectrometry

1. Introduction

Ti6AI4V alloy is increasingly used for loadbearing components of joint prostheses [ 1,21. The acetabular cup in a hip and the fibial plateau in a knee joint replacement usually consist of ultrahigh molecular weight polyethylene (UHMWPE) [3]. When the sliding surfaces of the femoral components are made from Ti6AI4V, wear can lead to black staining of the surrounding tissue and failure of the joint [46 ]. Small particles of bone cement (polymethylmethacrylate, PMMA) or bone can accelerate wear in joint prostheses [6,7]. The tribologicai behaviour of Ti6AI4V/UHMWPE sliding couples can be improved when the surface of the alloy is modified by ion implantation of nitrogen, as demonstrated in earlier works using pin-on-disc and annulus-on-disc devices [7-15] (Table l) or hip joint simulators [ I 1,15,16l.Carbon or oxygen implantation was shown to have a positive effect as well [ 8,12,13 ]. Implantation at different target temperatures was performed by Corchia et al. [ 15], who indicated that nitrogen ion implantation at 500 °C leads to higher UHMWPE wear than implantation below 250 °C. This shows that ion implantation at low temperatures is promising for wear reduction. According to Mishra and Davidson [ 17 ] and * Correspondingauthor. 0043-1648/97/$17.00 © 1997ElsevierScience S.A. All tights reserved PIISOO43-1648(96)O7456-X

Stretcher et al. [ 18], the wear resistance of Ti6AI4V and Ti6AI7Nb against PMMA is only slightly improved by nitrogen ion implantation. Earlier work by the present authors [ 13] revealed that, in addition to implantation of carbon, nitrogen or oxygen, low dose implantation of noble metals can also improve the tribological behaviour of Ti6AI4V sliding against UHMWPE. A necessary condition for reducing wear is that the layer em'iched in noble metals is in contact with the oxide film on the surface. This was shown by depth distribution measurements [13]. A systematic screening of the literature (Table I ) indicatestheinfluenceofthedifferentimplantation parameters used by various authors [7-15], As a general trend, nitrogen implantation leads to a significant wear reduction when the dose exceeds 4 × lO '? cm -2. Since only few energy-dose combinations were investigated in these studies, it is not yet clear which implantation parameters lead to an optimized tribological behaviour. In the present study, biocompatible elements (C, N, O, Y, Hf, Pt, Au) are implanted into Ti6AI4V under systematically varied conditions. The aim is to correlate the wear behaviour with the depth distributions of implanted elements and the microstructure in the near-surface region of the alloy in order to determine implantation parameters which enhance wear resistance of the Ti6AI4V surface and minimize wear of the UHMWPE counterbody.

50

H. Schmidt et aL / Wear 209 (1997) 49-56

Table I

Literature survey of wear measurements of Ti6AI4V/UHMWPE couples (pin-on-disc and annulus-on-discconfigurations) and conditions used during ion implantationof the alloy (n.a.: no data available) Reference

Tribological test system

Implantedelements

Sioshansiet al. [8]

pin: Ti6AI4V disc:UHMWPE pin: Ti6AI4V disc:UHMWPE pin: Ti6AI4V disc:UHMWPE disc: Ti6AI4V pin: UHMWPE disc: Ti6AI4V pin: UHMWPE disc: Ti6AI4V pin: UHMWPE disc: Ti6AI4V pin: UHMWPE

C, N

50

3

N

50

n.a.

N

40-100

5-7

Matthews et al. [9] Jianqiang et al. [ 10] Rieu et al. [ I I l Poggie ct al. [ 7 ] Alonso etal. [ 121 Schmidt et al. [ 13l

Martinella et aL [ 14]

Dose D ( tO 17cm -2 )

N

n.a.

N

60

3

C, N, O

50. 100, 180

4

C, N, O Y Hf Pd, b, Pt, Au N

disc: Ti6AI4V annulus: UHMWPE disc: TtOA]4V annulus: UHMWPI~

Cotchia eta]. 115]

Energy E (keV)

80 220 400 200 100

N

2. E x p e r i m e n t a l p r o c e d u r e Ti6AI4V discs ( diameter 38 m m ) and annuli (inner diameter 14 m m , outer diameter 20 m m ) with a globular (or + / 3 ) microstrucmre were used. The fiat surfaces to be ion implanted were mechanically ground and mirror polished to an average surface roughness less than 0.03 tim. Ion implantation w a s performed using the G S 1 3 0 0 kV accelerator [ 19]. The target temperature w a s kept below 100 °C by mounting the samples on water-cooled tmget holders. H o m o g e n e o u s implantation w a s achieved b y magnetic beam scanning. Discoloration o f the titanium samples due to surface contamination was avoided by keeping the residual gas pressure in the implantation c h a m b e r below 3 × 10 - 6 mbar. The ion species, energies and doses are summarized in Tables 2 and 3. Carbon monoxide ( C O + ions) and carbon dioxide (CO~" ions) were implanted at acceleratior, voltages o f U = 160 a n d 200 kV, respectively. The different energies

I--5

6 0.2-1.5 0.2-1.5 0.05-0.5 5

50. 90, 175

6--12

per atom o f C a n d O are due to the ratio o f the atomic masses. In addition to single implantations, nitrogen and carbon were also implanted at two a n d three different energies (double and triple implantations, Table 3 ) , in order to achieve a broad distribution o f implanted atoms. During triple implantation o f nitrogen, energies and doses as were used by Sioshansi [20] were applied. To gimulate non-normal incidence o f the ion beam during ion implantation o f joint prostheses, two aJmuli were implanted at an incidence angle o f ~ = 4 5 ° ( Table 3 ). Backscattering spectrometry analysis [ 21 ] was carried out for a carbon monoxide implanted Ti6AI4V disc (Table 2 ) . The spectrum w a s taken at the 7 M V Van de G r a a f f accelerator o f the Institute for Nuclear Physics, Frankfurt, using doubly charged helium ions (incident projectile energy: 7.6 MeV; scattering angle: 171°; specimen tilt angle: 70°). The depth distributions o f carbon a n d oxygen were calculated with the R U M P program [ 2 2 ] , using the scattering crosssec-

Table 2 Ion implantationparameters and microhardness ( at a load of 2 mN ) of the Ti6AI4V discs and wear behaviour against PMMA pins Ion

Energy/atom E (keV)

Atomic dose D ( 10*7cm- 2)

Microh~rdness (GPa)

Pt ÷ Au ÷

200 200

0.1 0.1

3.7 3.5

C+

CO* CO/" N~" C~

N~"

Non-implanted

80

C: 68.6 O: 91.4 C: 54.6 O: 72.7 80 80 80

Revolutions prior to scratching

6

5.2

C: 3 O: 3 C: 3 O: 6 3

5.1

50 20 5 000 I 500

4.9

2 000

6.2

> 50 000

6.4

>

3

6

3

50 000 <5

H. Schmidt et al. / Wear 209 (I 997) 4 9 - 5 6

51

Table 3 Ion implantation parameters of the Ti6AI4V annuli and their tribologieal behaviour against U H M W P E d i ~ s (& incidence angle of the ion beam with respect to the surface normal; b,p.: black particles in U H M W P E wear, see Fig, 8; s.l,: formation of a surface layer, see Fig. 4). No hydroxyapatite was added Ion

Pt + CO + N~" N+

Energy/atom E

Atomic dose

(keV)

D ( l0 ~ cm- 2)

200 C: 68.6 O: 91.4

0. I C: 3 O: 3 6 3

80 80

(~

Specimen

Number of

Friction

Wear volume

designation (.see Fig. 7)

~ratches on annulus

torque (N m)

(UHMWPE) AV (mm3)

0° 0°

Pt single CO single

~5 ~ 20

I. I 1.5

1.6/1.5 / 1.7 2.7/2.4/1.7/2.5/1.5

0° 0°

N single N + C single

~ 20 ~5

1.9 1.6

4.4 (b.p.)/3.1 1.6/1.4/1.5/ 1.0



N double

~ 10

1.6

1.5/0.8/I.6/I.3

N 45 ° double

~ 15

c ~

80

3

N+

130

3

N,+

6O

5

N+ N+ C+ C+

130 6O 120 60

3 5 3 4

45 °

1.5

1.210.9/I.g/1.4



C double

~5

l

1,310,7/0,7

N+

160

7



N triple

~ 15

2

1.211.9/2.0/1.2

N.,+ N_,+ N~

80 35 160

3 2.5 7

N 45 ° triple

~ 20

2

2.611.610.7

N_.*

Ni C+

gO 35 200

3 2.5 4.2



C triple

~ 15

1.9

1.2/1.4/1.7/1.4

C+

100

1.6 0°

Hf+ N

~ 10

1.2

2.0/1.4/1.7



Y+C

~ 15

1.6

1.9/1.7/I.5

Non. impl.

> 40

2.2

14.2 I b.p.)/7.3

C+

50

2.2

400 160

0.5 7

N2+

80 35

3 2.5

Y+ C+

220 200

0.5 4

Ht"z* N+

N2+

C÷ I00 C÷ 50 Non-implanted

45 °

{s.!.)

2 2

tions measured by Feng et al. [ 23 ] and A lmeida et al. [ 2 4 ] . In otdzr to simulate the actual depth profile, the ion-implanted layer was divided into 15 sublayers, each o f uniform composition. The microhardness of the Ti6AI4V discs was measured with a Fischerscope H 100 device [25] using a Vickers indenter. Hardness values were determined by measuring the depths o f the indentation during increasing loads from 0.4 to l 0 mN. Eight measurements were performed with each disc, and the average hardness values were calculated. Deviations from the average values at the different loads were smaller than 20%. A pin-on-disc device (CSEM, Neuchatel) was used to investigate the tribological behaviour of the Ti6AI4V discs against P M M A pins (load F = 5 N; contact area 1.5 mm2; contact pressure 3.3 M P a ) , with Ringer solution as a lubricant (0.146 M NaCI; 0.013 M KCI; 0.007 M C a C I z . 2 H 2 0 ) . A constant sliding velocitiy o f 5 c m s - ~ and test durations up to 50 000 revolutions ( ~ 20 h) were chosen. To determine the friction coefficient (it=f/F), the friction force f was measured with a load cell. In the production o f the P M M A pins, prepolymerized powder and liquid m o n o m e r ( provided by Heraeus Kulzer, Wehrheim) were mixed and pressed into

a mould. The powder contained small bulky zirconiaparticles (size approximately 20 v.m; opacifying agent in the bone cement). The behaviour of the Ti6AI4V annuli was investigated with an annulus-on-disc device developed at the Biomechanics Laboratory of the University o f Munich [ 26,27 ]. Tests were performed with flat U H M W P E discs (diameter 25 ram) which were cut and machined from bars. The discs were mounted in a gimbal sample holder to achieve a homogeneous load distribution. The friction torque was measured with a strain g a u g e attached to an arm which prevented the rotation o f the disc holder. The annuli were pressed against the discs, performing an oscillatory movement ( s w i n g angle 38°), using Ringer solution as a lubricant. A contact pressure o f 9.4 MPa, a m a x i m u m sliding velocity o f 4.2 cm s - ) ( 143 cycles min - t ) and a test duration o f 42 h ( 3.6 × I 0 s cycles ) were chosen. Each annulus was tested with several discs successively (Table 3). Some additional tests were carded out with hydroxyapatite ( H A ) which is similar to the mineral phase in bone. H A was used in powder form consisting of fine particles (grain s i z e < 1 Izm, provided by BenckiserKnapsack, Ladenbnrg) which were applied to the discs prior to the tests.

52

14. Schmidt et al. / Wear 209 (1997) 49-56

The sliding surfaces were investigated with an optical microscope (Optiphot-2, Nikon, Tokyo) and a scanningelectron microscope (DSM 962, Carl Zeiss, Oberkochen). The wear volumes A V of the UHMWPE discs were determined with a Dektak 8000 profiiometer (Veeco Sloan Technology, Santa Barbara).

Energy (MeV)

1

3.OxlO* 2.5

2

3

4

5

i

O (imprint¢~)

2.0

O (oxk~e

[~m)

ol.5 1.0

3. Results and discussion 0.5

3.1. Depth distributions and phases

The backscattering spectrum of the alloy disc implanted with carbon monoxide is shown in Fig. 1. Owing to the large scattering crosssections the yields of carbon and oxygen are strongly enhanced with respect to the substrate. The formation of an oxide film in air after implantation leads to a splitting of the oxygen peak. Due to the oxidation the oxygen concenwation at the surface is higher than the enrichment of implanted atoms ( < 10at.%, Fig. 2). The maximum of implanted carbon (29 at.%) andoxygen (23 at.%) is reached at a depth of 170 nm. The thickness of the implanted layer is 350 nm. Nitrogen or carbon implanted with an energy of 80 keV and the integral dose used for carbon monoxide implantation (6X 10~ cm -2) show similar depth distributions of implanted atoms (Fig. 2 ). The different depth distributious after single, double and triple implantations of nitrogen or carbon (Table 3) are schematically illustrated in Fig. 3. During implantation of nitrogen or carbon with doses exceeding 2 X 10I~ cm -2, finely dispersed TiN or TiC precipitates are formed, as shown by transmission electron microscopy investigations reported in the literature [ 28-32 ]. X-ray diffraction measurements revealed that TiO is formed during carbon monoxide or carbon dioxide implantation [ 13]. The size of the titanium compound precipitates (530 n m [ 28,30,31 ] ) increases at higher doses [ 28,29,31 ]. Carbon, nitrogen and oxygen stabilize a-titanium, and consequently the matrix in the subsurface region consists of aTi. The successive implantation of hafnium and nitrogen (Table 3) leads to the formation of the ternary compound (Ti,Hf)N in the near-surface region [33]. Investigations by Bolster et al. [28] and Qiu et al. [29] reveal that carbon or nitrogen implantation can lead to the formation of continuous TiC or TiN layers when low ion energies ( E < 80 keV, normal incidence) and doses exceeding 7XI017cm -2 are used. During the nitrogen triple implantation at 8=45 ° ( E = 35-160 keV, integral dose 1.25 X 10Is cm -2, Table 3), a continuous surface layer was formed on the alloy, resulting in a distinct discoloration. Owing to the lower penetration depth of the ions at ~5=45°, the enrichment of implanted nitrogen in the near-surface region is higher when compared to normal incidence. The original silvery colour changes into a dark bluish grey (Fig. 4) which indica,es that the layer consists of oxygen-

0.0

=;o

,,~o

Channel

s;o

noo

Fig. i. Backscattering spectrum of CO-implanted Ti6AI4V ( C O +, U = 160 kV, integral dose 6 × 10 j7 c m - 2) and R U M P calculation ( smooth

curve). 60-

~40-50!

~

L

_O.3o_~ ~

~

(implanted)

o

°

0

.

.

.

.

!

.

~ m lnml Fig. 2. Depth distributions o f C and O (conesponding to the calculated curve in Fig. I ) as well as that o f implanted atoms ( C + O ) , which was estimated at depths ofO-15 nm.

.y~ngletrn~. oouuetm~. Tri~tm~.

'V\

<<

Fig.3. Schematicdepthdistributionsaftersingleand multipleimplantations at normalincidence. containing TiN. This is due to oxygen absorption during implantation. After platinum or gold implantation (E=200keV, D = 1 0 ' 6 c m -2) the thickness of the implanted layers is approximately 100 n m [ 13]. The enrichment of implanted atoms starts at the titanium surface, and the maximum Pt or Au concentration (~5.5 at.%) is reached at a depth of 35 nm. At high titanium concentrations noble metals (e.g. Pd, Ir, Pt, Au) form intermetallic compounds (Ti4Pd, Ti2Pd, Ti3Ir, Ti3Pt, Ti3Au) [34]. During the noble metal implantations precipitates of intermetallic compounds are formed. The implanted dose used in the present study ( 10 '6 cm -2) is the same as that used by Tomashov et aL [35], who found that the implanted layer formed during palladium implantation into titanium ( E = 4 0 keV) cousists of Ti2Pd crystallites (size 3-6 rim).

53

1"1. Sc'hmidt et al. / W e a r 2 0 9 ( 1 9 9 7 ) 4 9 - 5 6

10000(

~1000{ t0~ "6 10(] 7~

tc

~

x

Hardness [GPal

Icm

Fig. 6. Correlation between microhardness (load 2 mN, Table 2) and number of revolutions prior to scratching.

Fig. 4. Ti6AI4V specimen after nitrogen triple implantation at 8 ffi45°: ion implanted (dark grey ) and non-implanted (light grey) surface regions.

3.2. Microhardness and tribological behaviour of the Ti6AI4V discs Hardness diagrams of ion-implanted and non-implanted samples are shown in Fig. 5. The values observed at a load of 2 mN are listed in Table 2. Non-implanted discs show a microhardness of about 3 GPa. Low-dose implantation of platinum or gold ( D = 1.016cm-2) slightly t'ncrease~ the microhardness in the subsurface region. Implantations of carbon, nitrogen oroxygen ( D > 6 × i0 '7 c m - 2 ) produce highex hardness values than Pt or Au implantations do. For loads above 5 mN the penetration depth of the indenter exceeds the thickness of the implanted layer, and the hardness values drop to those of the bulk. During wear tests of non-implanted discs against PMMA pins scratching of the titanium surface occurs, and the friction coefficients exceed 0.5. Black particles which consist mainly of titanium oxide are removed from the alloy surface. The depth of the scratches and the ameunt of wear particles both increase with the test duration. Discs subjected to low-dose implantation of platinum or gold show scratching of the surface after several revolutions. The depth of the wear marks on the discs exceeds the thickness of the ion-implanted layers, and the layers are completely removed. With carbon, nitrogen or oxygen implanted discs the number of revolutions after which scratching occurs is higher ( > 1500). After nitrogen implantation or the successive implantation of nitrogen and carbon no scratching is observed within 50 000 revolutions. On the alloy surface a PMMA film is formed in which zirconia

-~ / ~

~ ~

\

N

implanted ...CO implanted

implanted '1,

I 2

,

I 4

,

i 6

Load [mNl

,

I 8

, 10

Fig. 5. Microhardncss of non-implanted Ti6AI4V and of Ti6AI4V implanted with nitrogen, carbon monoxide and platinum (Table 2 ).

particles are embedded. A smoothening of the sliding surface of the pins occurs, and the friction coefficients drop below 0.35. After exposure to air, non-implanted as well as ionimplanted Ti6AI4V surfaces are covered with a thin oxide film. Due to the adhesion between the polymer and the oxide film during sliding, oxide particles break away from the titanium surface and promote scratching. Owing to the high oxygen affinity of titanium and aluminium the oxide film reforms rapidly after removal. Fig. 6 shows that a linear relationshtp exts~,sbetween the microhardness and the logarithm of the number of revolutions prior to scratching. This demonstrates that the increase of the wear resistance of the Ti6AI4V discs against PMMA can be predicted by microhardness measurements.

3.3. Tribological behaviour of the Ti6AI4V/UHMWPE couple The results of the tests on the Ti6AI4V annuli and the friction torques are listed in Table 3. The wear behaviour is illustrated in Fig. 7. Although the friction remains fairly constant for each annulus the wear of the UHMWPE counterbodies alters significantly. The creep deformation of a UHMWPE disc induced by a static load test of 42 h and a relaxation time of 6 days (A V= 0.5 mm3) is smaller than the lowest UHMWPE wear volume (A V= 0.7 mm3). During the tests ot non-implanted annuli the friction torque (2.2 N m) and the UHMWPE wear are high, and black particles are abraded (Fig. 8). Ion-implanted Ti6AI4V annuli show lower friction torques ( < 2 N m ) and a strong reduction of UHMWPE wear. During the first wear test of each annulus a polyethylene film adhering to the alloy surface is formed (Fig. 9). When the film is carefully removed, narrow grooves are observed (Fig. I0 ). During the sliding movement small oxide particles or particles from the ion-implanted region are removed from the titanium surface due to adhesion between the polyethylene and the alloy. When the particles are embedded in the UHMWPE counterface they induce scratching of the titanium surface. The polyethylene film is initially formed in those regions of the alloy surface where no particles are removed. With increasing test duration the film grows, and the grooves

54

H. Schmidt et al. / Wear 209 (1997) 4 9 - 5 6

i! ~o impl.

4 4Ii -3

J

I

o

I

Fig. 7. Wearl~havioucof ti~ Ti6PJ4V annuli. The implantationp~umcterscon'espondingto ll~ specimendesignalionsare listed in Table 3.

Fig. 8. UHMWPEdiscs after the wear tests: high polyethylenewear with metallic blackparticles (left); low polyethylenewear (right).

Fig. 9. Sliding surface (polyethylene film) of the C-implantedTi6AI4V annulus(triple implantation). are covered with polyethylene. Oxide particles are embedded into the film. The final thickness of the film drops with the number of scratches and with decreasing wear of the UHMWPE counterface. The addition of hydroxyapatite

Fig. 10. Slidingsurfaceof the CO-implantedTi6AI4Vannulusafterremoval of the polyethylenefilm with a polishingpaste.

(HA) to the lubricant initially increases the polyethylene wear, and the film formed on the Ti6AI4V annuli is thicker when compared to tests without HA. During the movement the particles which are entrapped between the sliding surfaces are embedded into the polyethylene film and no longer act as an abrasive medium. The number of scratches is not increased by the presence of HA particles. The length of the grooves on the aanuli often exceeds the length corresponding to the swing angle, and some wear marks are too closely spaced to be clearly distinguished in the microscope. Approximate values for the number of scratches on the annuli are given in Table 3 and in Fig. 7. The number observed with non-implanted specimens exceeds 40. Ion implantation leads to a strong reduction of the number of scratches ( < 2 0 ) . A high friction torque (_> 1.9 N m) is accompanied by a large number of scratches ( >_ 15). At lower torques no systematic correlation between friction and wear is observed.

H. Schmidt et al. / Wear 209 (1997149-56

The following correlations between the wear behaviour (Fig. 7) and the depth distributions and phases induced by ion implantation can be discerned: (i) Precipitates of intermetallic compounds (ThPt, Ti31r, T h A u ) formed below the oxide film lead to a reduction of oxide particles removed from the allot surface. The number of scratches on the annulus and the UHMWPE wear are strongly reduced after implantation of platinum, although the hardness increase is not pronounced (Table 2). Similar observations were made with iridium or gold implanted Ti6AI4V discs sliding against UHMWPE pins [ 131. (ii) After nitrogen or carbon implantation a strong wear reduction occurs when the enrichment of implanted atoms starts directly at the surface ( Fig. 3 ), i.e. when a large fraction of the dose is implanted at energies below 70 keV. After single implantation of nitrogen or carbon monoxide the wear rates are higher than after multiple implantations. When the concentration of implanted atoms at low depths is very high and a continuous surface layer is formed ( Fig. 4), the reduction of the number of scratches is less pronounced. Successive implantations of hafnium and nitrogen or of yttrium and carbon do not reduce wear when compared to nitrogen or carbon triple implantations. (iii) The f o ~ a t i o n of TiN or TiC precip;'~te~ in the nearsurface region strongly reduces the number of particles removed from the Ti6Al4V surface. After double implantations the wear rates are lower than after triple implantations. This indicates that more particle~ are removed from the surface when the size of the precipitates increases, i.e. when higher total doses are used during the implantation. The most pronounced reduction of scratches and UHMWPE wear is achieved by double implantation, i.e. when the enrichment of implanted carbon or nitrogen (Fig. 3) and the formation of precipitates start directly below the oxide film on the titanium surface, followed by a continuous decrease of the concentration of implanted atoms with increasing depth.

4. Conclusions Ion implantation into Ti6AI4V with the appropriate parameters leads to a reduction of wear during the sliding movement of the alloy against polymers. The friction decreases with the number of oxide particles between the sliding surfaces. The wear resistance of Ti6AI4V against PMMA increases with microhardness. Wear of Ti6AI4V/UHMWPE sliding couples is reduced after noble metal implantation. which can be attributed to the formation of intermetallic compounds. The formation of TiN or TiC precipitates during nitrogen or carbon implantation leads to a strong reduction of the number of scratches on the alloy surface and of UHMWPE wear. Scratching has also been observed on the sliding surfaces of Ti6AI4V hip and knee joint prostheses which had been retrieved from patients [4]. Although the annulus-on-disc device does not replicate all tribological conditions in a joint

55

( with respect to lubrication and dynamic loading), the reduction of the number of scratches and of UHMWPE wear are supposed to be suitable criteria for the comparison of different ion implantation conditions. With respect to the improvement of the tribological behavionr of Ti6AI4V/UHMWPE joint prostheses the following conclusions can be drawn: ion implantation of nitrogen and carbon with doses exceeding 6 x 1017 cm- "- is suitable for the modification of the sliding surfaces of the titanium alloy. The major part of the dose should be implanted at energies below 70 keV to ensure a high enrichment of iraplanted atoms directly below the oxide film. The integral dose should not exceed 10 ~s c m - 2 to avoid the formation of continuous surface layers.

Acknowledgements The authors thank Dr U. Fink (Aesculap AG, Tuttlingen) for providing the specimens and Dr H. Baumann (University of Frankfurt) for supplying the backseattering spectrum.Further thanks go to Prof. H.E. Exner (Technical University of Darmstadt) for critical discussions and improvements oftbe manuscript. Th:,s wGrk has been financially supported by GSI (Contract No. DAEXM) and by Deutsche Forschungsgemeinschaft (Contract No. Ex 8/3-1 ).

References [ I I R. Van Noon, Titanium:the implantmaterialof today,J. Mater. Sei.. 22 (1987) 3801-3811. [ 21 M. Semlit~h. Titaniumalloys for hlpjoint replacements.Clin. Mater.. 2 (1987) 1-13. 13} H.-G. Willert. G.H. Buehhoro and P. Eyerer. Ultra-High Molecular Weight Polyethylene as Biomaterial in Orthopedic Surge~', Hofgrebe and Huher. Toronto. 1991. [41 S. Nasser.P.A. Campbell,D. Kilgus.N. Kossovskyand H.C. Amstatz. Cementless total joint arthroplasty prostheses with titanium-alloy articularsurfaces.Clin. Orthop., 261 (1990) 171-185. [5 ] H.J. Agins. N.W. Alcock,M. Bansal,E.A. SalvatLP.D. Wilson.P.M. Pellicciand P.G. Bullough,Metallicwear in failedtitanium-alloytotal hip replacements.J. Bone Joint Surg.. 70.A (1988) 34%356. [61 A.V. Lombardi,T.H. Mallory,B.K. Vanghnand P. Drouillard.Aseptic looseningin total hip arthroplasty.secondaryto osteolysisinducedby wear debris from titanium-alloymodularfemoralheads.J. Bone Joint Surg., 71-A (1989) 133%1342. [71 R.A. Poggie, A.K. Mishra and J.A. Davidson. Three-body abrasive wear hehaviourof orthopaedicimplanthearingsurfacesfrom titanium debris.J. Mater. Sci. Mater. Med.. 5 (1994) 387-392. 18] P. Sioshansi,R.W. Oliver and F.D. MaUhews.Wear improvementof surgicaltitaniumalloys by ion implantation.J. Vac. Sci. Technol., A3 (1985) 2670-2674. [91 F.D.Manhews, K.W.GmerandD.L. Armstrong.The effectof nitrogen ion implantationon the abrasive wear resistanceof the Ti--6AI-4V/ UHMWPE couple, in J.M. Williams. M.F. Nichols and W. Zingg (Eds.). Mat. Res. Soc. Symp. Proc. Vol. 55, Materials Research Society,Pittsburgh, 1986.pp. 243-252. 1101 Z. Jianqiang.Z. Xiaozhong.G. Zintang and L. Hengde, Comparison of wear resistance of ion implanted surgical Ti-6AI--4V with other alloys, in J.M. Williams,M.F. Nicholsand W. Zingg (Eds.L Mat. Res.

56

H. Schmidt et al. / Wear 209 (1997) 49-56

Soc. Syrup. Proc. Vol. 55, Materials Research Society, Pittsburgh, 1986, pp. 229-235. [ 111 J. Rieu, A. Pichat, L.M. Rabbe. A. Rambert, C. Chabrol and M. Robelel, Structural modifications induced by ion implantation in metals and polymers used for orthopaedic prostheses, Mater. Sci. Technol., 8 (1992) 589-593. [ 12] F. Aloe.so, A. Arizage, S. Quainton, JJ. Ugarte, J.L. Vivante and J.L Onate, Mechanical properties and structure of Ti-6AI-4V alloy implanted with different light ions, Surf.. Cont. Technol., 74-75 ( 1995 ) 986-992. [13] H. Schmidt, H.E. Exner, D.M. ROck, N. Anger and U. Fink, Wear bebaviour of ion implanted Ti6AI4V against UHMWPE, in R. Kossowsky and N. Kossovsky ( Eds. ), Advances in Materials Science and Implant Orthopedic Surge~', Kluwer, Dordrecht. 1995, pp. 207221. ] 14] R. Martinella, S. Giovanardi and G. Palomburini. Wear of ultrahigh molecular weight polyethylene sliding against surface-treated Ti6AI4V, A1S1316 stainless steel and Vitallium, Wear. 133 (1989) 267-279. [15]M. Corchia, R. Giorgi, S. Turtu, F. Caccavale, S. LoRusso. P. Mazaoldi. G. Abutangelo, R. Cortivo. S. Giovanardi, R. Martinella, A. McCabe. D. Chivers and P. Tweedale. Wear bebuvionr improvement of the couple Ti6AI4V-UHMWPE by ion implantation. Abstr. Ilth Fur. Conf. on Biomaterials, Pisa, 1994, pp. 368-371. [ 16] J. Lausman. T. R~stlund and H. MeKellop. Wear of ion implanted pore titanium and Ti6AI4V alloy against ultrahigh molecular weight polyethylene, Surf.. Eng., 7 ( 1991 ) 311-317. [~7l A.K, Mishra and J.A. Davidson, PMMA abrasion resistance of zirconia, TiN and ADLC surfaces, Trans. 4th World Biomat. Cong., Berlin, 1992, p. 116. [ 18l R.M. Streieber, H. Weber, R. Sch6n and M. Senditsch, New surface modification for Ti--6AI-7Nb alloy: oxygen diffusion hardening (ODH), Biomatenals, 12 ( 199i ) 12.5-129. [ 19] D.M. ROck,H. Angert, H. Emig, K.D. Leible, P. Spltdtke, D. Vogt and B.H. Wolf, Ion implantation facilities at GSI, Nuc£ Tracs Radiat. Meas.. 19 ( 1991 ) 951-954. [20] P. Siosbansi, Surface modification of industrial components by ion implantation. Mater. Sci. Eng.. 90 (1987) 373-383. ]21]J.A. Leavitt, L.C. Mclntyra and M.R. Weller. Backscattering spectrometry, in J.R. Tesmer and M. Nastasi (Eds.), Handbook of Modem ion Beam Materials Aanlysis, Materials Research Society, Pittsburgh. 1995, pp. 37-81. 122l L.R. Dooliale, High energy backscattering analysis using RUMP, in I.R. Tesmer, CJ. Maggiore, M. Nastasi, J.C. Barbour and J.W. Mayer (Eds.). High Energy and Heavy Ion Beams in Materials Analysis, Materials Research Society, Pittsburgh, 1990, pp. 175-182. [23] Y. Feng, Z. Zbou, Y. Zhou and G. Zhao, Cross-sections for 165° backscattering of 2.0-9.0 MeV 4He from carbon. Nucl. Instr. Meth.. B86 (1994) 225-230. 1241 FJ.D. Almeida, J.A. Davies and T.E. Jackman, Absolute calibration of the re*O(a,a) t°O elastic scattering resonance at 7.30-7.65 MeV and applications to oxygen depth profiling, Nacl. Instr. Meth.. B82 (1993) 393-398. [25] W. Weiler, Hardness testing - - a new method for economical and physically meaningful microhardness testing. Brit. J. Non Destr. Test.. 31 (1989) 252-258.

[ 26] J. Hinterberger. M. UngethUmand W. Plitz. Tribological properties of aluminium oxide ceramics, in G.W. Hastings and D.F. Williams (Eds.). Mechanical Properties of Biomaterials. Wiley. Chiehester. 1980. pp. 73-82. [27] A. Walter and W. piitz, The ring-on-disc method - - clinical sigeificance of a wear-screening test of biomaterials for hip joint alloplasty, in S.M. Penen and E. Schneider (Eds.), Biomechanics: Current Interdisciplina~. Research. Martinus Nijoff. Dordrecht. 1985. pp. 129-134. [28] R.N. Bolster. I.L. Singer and R.G. Vardiman, Composition. structure and wear resistance of Ti-6AI-4V implanted with carbon or boron to higIzdoses. Su~ Cont. Techaol.. 33 (1987) 469-477. ]29] X. Qiu. R.A. Dodd. J.R. Conrad. A. Chon and F.J. Worzala. Microstroctural study of nitrogen-implanted Ti-6AI-4V alloy. Nucl. Instr. Meth.. B59/60 ( 1991 ) 951-956. [30] R. Martinella. S. Giovanardi. G. Chevallard. M. Villani. A. Molinari and C. Tosello. Wear behaviour of nitrogen-implanted and nitrided Ti-6AI4V alloy. Mater. Sei. Eng.. 69 (1985) 247-252. [ 31 ] R.G. Vardiman and R.A. Kant. The improvement of fatigue life in Ti6AI-4V by ion implantation. J. Appl. Phys. 53 (1982) 690-694. [32] R. Hutchings. The subsurface microstructure of nitrogen-implanted metals, Mater. Sci. Eng.. 69 (1985) 129-138. [33] D.M. ROck. H. Schraidt. N. Angert. O. Yoda. Y. Aoki and H. Naramoto. Characterization of a hafnium and nitrogen implanted titanium alloy, in J.S. Williams. R.G. EIliman and M.C. Ridgway (Eds.). Ion Beam Modification of Materials, Elsevier. Amsterdam, 1996. pp. 1016-1019. [34] T.B. Massalski. Bina~" Alloy Phase Diagrams. ASM. Metals Park. 1986. [ 35 ] N.D. Tc,m ~ o ; ' , M.I. Gu~'.'~,G.A. Fedos~va. A K. GorodetskiI.A.P. Zakharov, D.B. Bogomolov, G.P. Cberoova, B.G. Vladimirov and S.M. Ivanov, Improving the con'osion resistance of titanium by pall~ium-ion implantation, .toy. Phys. Dokl,. 26 ( 1981 ) 222-224.

Biographies Harald S c h m i d t is research associate at the D e p a r t m e n t o f Physical Metallurgy o f the T e c h n i c a l University o f D a r m stadt. H e studied Physics and r e c e i v e d a P h D in Materials Science. H e is w o r k i n g on w e a r and corrosion properties o f titanium with respect to biomedical applications. A o d r e a s S c h m i n k e studied Materials Science and is a P h D student at the D e p a r t m e n t o f Physical Metallurgy o f the T e c h nical University o f Darmstadt. H e is w o r k i n g on the analytical characterisation and tribological properties o f ion-implanted titanium. D o r o t h e e M. Riick is with the Materials R e s e a r c h Departm e n t o f the G S I Centre for H e a v y Ion Research, Darmstadt. She is coordinator o f the ion implantation group. H e r m a i n field o f research is modification o f polymers.