ABSTRACT:
The lower wear rate obtained with bulk alumina femoral hip balls compared to cobalt-chromium alloy balls on ultra-high molecular weight polyethylene is well documented in clinical studies. An alumina coating on a cobalt-chrome component would produce the superior surface properties of the ceramic combined with the desirable bulk properties of the metal. A new technique for forming a blended interface between the alumina coating and cobalt-chrome has been developed. The process uses a high energy ion beam of aluminum impinging on a cobalt-chrome substrate in an oxygen ambient gas to grow the alumina from beneath the metal's surface. Interface thicknesses up to 1200 Angstroms thick have been observed. To date, flat coupons and hip balls, have been coated with coatings 1 micron thick. The coatings were evaluated for polyethylene wear using a laboratory screening tribometer which has been designed to simulate non-linear tracking of the femoral ball on the UHMW polyethylene. The adhesion was measured using a diamond point scratch tester. Surface roughness is equal to or slightly improved over the cobalt-chrome substrate.

KEYWORDS:
ion implantation, alumina, ceramic, coating, polyethylene wear, pin-on-disk, cobalt-chrome

Polyethylene wear debris is one of the primary causes of loosening resulting in the need for subsequent revision surgery in total joint replacement. Numerous clinical studies [1-6] have shown a 38% to 73% reduction in wear when articulating against monolithic/alumina femoral heads. One clinical study, Bragdon et al [7] showed no improvement and one study by Livingston et al [8] showed increased wear of Hylamer UHMWPE which the authors did not investigate further. Figure 1 shows clinical data for both polyethylene wear and total hip replacement (THR) revision rates using CoCr balls vs alumina balls.

Wear rates measured from clinical radiographs of Weber THR designs [1] show a 3.7 times reduction in polyethylene wear on average when alumina balls were used instead of CoCr (left side of Figure 1A). Revision rates reported for Muller THR designs with CoCr balls were ten times higher than when ceramic balls were used [8]. (Right side of Figure 1.) These Ceramic advantages have not yet been realized on a widespread basis for total knee replacement components.

FIG. 1 - Clinical data for UHMWPE wear, CoCr vs monolithic alumina for Weber THR designs (left side) from Ref. 1. Revision rates for CoCr vs alumina for Muller THR design (right side) from Ref 8.

Materials and Methods

Alumina Coating Process
In order to specifically address the femoral knee application, an adherent alumina coating is being developed. However, aluminum oxide is not readily compatible with cobalt chrome alloy with respect to chemical affinity. Samples of CoCr alloy coated by the authors using magnetron sputtering, ion beam assisted deposition, and chemical vapor deposition (CVD) all debonded using a diamond scratch test method [9] at loads between 5 and 10 Newtons, while the goal was at least 30 Newtons. Failure occurred at the metal-ceramic interface.

Recently, we have ion implanted aluminum at high energy into CoCr while in an oxygen ambient in order to grow an alumina coating starting from beneath the alloy surface. Through a series of trials at various temperatures and ion beam current densities, it was found that alumina can be grown at a temperature as low as 500^oC at an ion beam current density of 25µA/cm².

The growth proceeds in three steps as shown in figure 2. In step 1 the aluminum ions are implanted below the surface of the CoCr by virtue of their 90keV kinetic energy. Subsequently in step 2, oxygen diffuses into the metal and reacts with aluminum atoms diffusing outward toward the surface. In step 3, residual surface CoCr is sputtered away by the ion beam and the alumina coating thickens as additional aluminum and oxygen are reacted. The graded interface between the alumina coating and the CoCr substrate determined by Auger (AES) depth profiling was 660 Angstrom thick, as shown in figure 3.

FIG. 2 - Ion beam synthesis of alumina coating on cobalt-chrome

FIG. 3 - Auger election spectroscopy (AES) of alumina-cobalt-chrome interface

The alumina coatings thus grown were found to have the same average surface roughness as the uncoated Cobalt-Chrome. Figure 4 shows sample surface roughness profiles for bare a cobalt-chrome control (left) and an alumnia coated cobalt-chrome taken with a Sloan Dektak II profilometer located at Implant Sciences Corporation. The alumina coatings also had a straw color typical of oxidized chromium.

FIG. 4 - Profilometry traces of a Co-Cr control surface (left), Ra=0.015 micron and an alumina coated CoCr (right), Ra=0.0013 micron.

The smoothness of the coating and the low temperature deposition process used indicate that the alumnia is probably amorphous, but x-ray diffraction measurements to determine the crystallinity have not yet been done. Constant load scratch tests using a standard Rockwell-C diamond indentor were preformed to measure the adhesion of the alumina coating. Figure 5 shows magnified scratches for the ion beam synthesized alumina (left) and ion beam assisted deposition (IBAD) alumina (right). The left scratch made at 30 Newtons has no delamination at the edges of the scratch whereas the IBAD alumina shows severe delamination at only 10 Newtons.

FIG. 5 - Thirty Newton scratch test on ion beam synthesized alumina (left) and a 10 Newton scratch test on IBAD alumina (right).

Wear Testing
Laboratory screening tests based on the geometries of pin-on-disk, block-on-journal, and flat-on-disk have been used in the past to evaluate UHMWPE wear. While these devices do not simulate physiological loads or geometries, they do attempt to provide equivalent contact stresses, frequency, stroke amplitude and speed. Typical results from the CoCr/UHMWE wear couple for clinical data, joint simulators, and laboratory screening devices are shown in table 1.

Table 1 - Typical average wear for the CoCr/UHMWPE couple

As can be seen from the table a good joint simulator can replicate the low end of the clinical wear data which tend to have a wide variation. Laboratory screening tests typically show wear a factor of 10 or more lower than joint simulators. Recently Bragdon[14] and Wang[15] have proposed that laboratory screening devices have linear tracking which tends to actually strengthen the UHMWPE and reduces the observed wear. It appears that non-linear tracking, where the track of the metal on the UHMWPE traces out a complex path over an area of the polyethylene, is the key parameter which is responsible for the much higher wear values observed in joint simulators.

Thus the simple, low cost screening test may actually be counter-productive if it does not simulate the complex wear mechanism which actually occurs in-vivo. In fact, Wang[15] has shown that a simple wear machine without non-linear tracking (line contact linear reciprocating) actually reversed the rankings of 2 UHMWPE materials obtained on a subsequent joint simulator test. However, the pin-on-disk test is inexpensive and offers the additional advantage that a dimensional measurement of UHMWPE wear can be used which is much less complicated than a weight loss measurement. Although dimensional measurements of UHMWPE wear and pin-on-disk screening tests have fallen into disuse recently, the technology may be useful to screen coatings if the wear mode can be modified to eliminate linear tracking.

In order to properly simulate the in-vivo wear mechanism, we have modified a conventional pin-on-disk machine to provide a non-linear tracking wear test. This was done by placing the disk and its motor assembly on a translation stage which then allows the radius of the wear track to change continuously during the test. The amount of the radius change was selected using a random number generator in a personal computer to randomly change the radius. The direction of the rotation was also changed at 6 second intervals. Figure 6 shows a photograph of the disk and the linear translation stage. Figure 7 shows a computer simulation of the first 15 rotations of disk. After a few hundred cycles the entire region between the minimum radius and the maximum radius is completely filled with cris-crossing tracks. The ball was also continuously rotated at 1 rpm to eliminate the single wear spot obtained in conventional pin-on-disk tests.

FIG. 6 - Mechanical arrangement of modified pin-on-disk with randomly varied track radius

FIG. 7 - Locus of wear track for first 15 turns of disk

In order to evaluate this new wear tester concept it was necessary to verify two basic concerns:

1. That any creep of the UHMWPE is adequately accounted for
2. That the total wear volume is much higher than in the linear tracking pin-on-disk test.

FIG. 8 - Static creep test measurements vs soak time of UHMWPE

A comparison of a single track wear test and a non-linear (cross-shear) wear test were then run on CoCr femoral heads to assess the differences in wear volume. The pin-on-disk wear test parameters are shown in table 2. The creep test results were also used to reduce the apparent wear volumes measured in both tests but the magnitude of these corrections was smaller than the data symbols on the graph. Note that the cross-shear tests yielded UHMWPE wear rates which are about 8 times the linear tracking values, which is close to the discrepancy seen between joint simulators and linear tracking screening devices as previously shown in table 1. Figure 9 shows the linear tracking wear vs the cross-shear wear of UHMWPE using CoCr balls in both cases. The reason for this large difference in wear rate can be explained by examining the photographs of the wear track on the two CoCr balls shown in figures 10 and 11. Both CoCr balls themselves also showed some wear. This manifested as a flat spot on the ring that contacted the UHMWPE. The linear tracking ball (figure 10) was very highly polished in the contact track and showed 3.6 times less wear than the cross-shear ball. The ball used in the cross-shear ball. The ball used in the cross-shear (figure 11) test was heavily worn and showed numerous scratches which may have been responsible for the heavy UHMWPE wear. Both of these pin-on-disk machines were in a common particle-free clean hood so these scratches could not have been due to particle contamination in only the cross-shear test apparatus. These data give us reasonable confidence to use this cross-shearing test methodology when comparing ceramic coated pin versus CoCr pins.

TABLE 2 - Laboratory wear test parameters

FIG. 9 - Linear tracking wear vs cross-shear wear for CoCr ball on UHMWPE disk

FIG. 10 - Micrograph of linear tracking groove on CoCr ball. Marker lines are 500 microns apart.

FIG. 11 - Micrograph of cross-shear tracking groove on CoCr ball. Marker lines are 500 microns apart.

Results
One CoCr control ball and one alumina coated ball were tested to 2 1/2 million-cycles using a 2 station wear test machine both in cross-shear mode in ordered to obtain a preliminary indication of reduction of wear.

The wear results of the alumina-coated CoCr ball and the CoCr control are shown in figure 12. The error bars shown are the standard deviations computed from measuring the groove area at four equally spaced points around the UHMWPE wear track. In both cases the data have been corrected for creep. The alumina coating produces about 1/3 the wear of UHMWPE than was produced by the CoCr after 2 1/2 million cycles in this single ball test.

FIG. 12 - Cobalt-chrome cross-shear wear vs alumina cross-shear wear of UHMWPE

Figure 13 shows a comparison of the depth profiles of the UHMWE wear grooves in (a) alumina coated ball and (b) a CoCr control ball at the end of the 2 1/2 million cycle test. The photograph of the wear track on the alumina coated ball in figure 13 shows a very polished contact region which may be responsible for the reduced UHMWPE wear observed.

FIG. 13 - Profilometer trace of wear grooves in UHMWPE from (a) alumina coated ball and (b) Cobalt-chrome control ball

Discussion
These preliminary tests show that the use of an alumina coating on a CoCr femoral head may reduce the wear of UHMWPE to about 1/3 of that caused by an uncoated CoCr head. This comparable to the wear observed from a monolithic alumina head.

It has also been shown that a modified pin-on-disc test where the radius of the track is changed randomly may be a useful device to screen coatings although the exact amount of cross-shear to simulate in-vivo conditions needs further validation. While the absolute wear differences between ceramic coated CoCr and a control would have to be determined using a good joint simulator and many more test specimens, we believe that this test may properly rank the wear performance in the proper order. In addition, the anticipated problems of dimensional UHMPWE creep effects can be surmounted.

Further tests are planned on variations of the alumina processing to minimize the UHMWPE wear. Zirconia ceramic, made with a similar ion beam technique, will also be evaluated for wear. However, these wear tests however did not test for third body wear from bone cement and other hard debris which may become trapped in the joint in-vivo. Further tests are planned with thicker alumina coatings to evaluate performance under these conditions.

Progress has also been made in coating complex shapes, such as femoral knee prostheses. These knees coated with 1 micron of alumina show some non-uniformities in the straw color but these are due to optical interference effects rather than large variations in coating thickness. Thus a thin alumina coating on CoCr components could reduce UHMWPE wear to less than 1/3 of present values and therefore, prove to be an important advance in TJR prosthesis design.

FIG. 14 - Micrograph of wear track on alumina coated CoCr ball

Contact us for additional information on our products and how they can benefit your products.

References
[1] Schuller H. M., Marti R. K., "Ten-year socket wear in 66 hip arthroplasties - ceramic versus metal heads", Acta Orthop Scand, 61 (3): 240,1990

[2] Zichner L. P., Willert H. G., "Comparison of alumina - polyethylene and metal- polyethylene in clinical trials", Clin Orthop, 1992; 282: 86-94

[3] Semlitsch, M. and Willert, H. B., "Clinical Wear Behavior of Ultra-High Molecular Weight Polyethylene Cups Paired with Metal and Ceramic Ball Heads in Comparison to Metal-on-Metal Pairings of Hip Joint Replacements," Proc. Instn. Mech. Engrs./IMechE, 211, Part H (1997): 73-88.

[4] Oonishi, H., Igaki, H., and Takayama, Y. "Comparisons of Wear of UHMWPE Sliding Against Metal and Alumina in Total Hip Prosthesis," Bioceramics, 1(1989):272

[5] Semlitsch, M., Streicher, R. M., and Weber, H., Orthopaede, 18 (1989): 337-381

[6] Weber, B. G., "Total Hip Replacement: Rotating Versus Fixed and Metal Versus Ceramic Heads," The Hip, 1981: 264-275

[7] Bragdon, C. R., et al., "Wear at the Ceramic on Polyethylene Articulation," Trans. 21st Society for Biomaterials Meeting, March 1995: 50.

[8] Livingston, B. J., Chmell, M. J., Reilly, D. T., Spector, M., and Poss, R., "The Wear Rate of Hylamer Cups is Higher Than Conventional PE and Differs with Heads from Different Manufacturers," Trans. 43rd ORS, February 1997: 141.

[9] Paschen U., "In-vivo-verschleib der gleitflachpaarung Al2O3 -keramic/polyethylen bei huftendoprosthesen," Taf Der Mundlichen Parfung 7: 1-15, 1986

[10] Selker J., Steinmann P. A., and Hintermann H.E., "The Scratch Test: Different Critical Load Determination Techniques," Surface Coatings and Technology, 36 (1988) 519-529

[11] Griffith M. et al., "Socket Wear in Charnley Low Friction Arthoplasty of the Hip," Clin Orthop, 137, 37, (1978)

[12] McKellop H. A. and Clarke I. C., "Degradation and Wear of Ultra-High-Molecular-Weight Polyethylene," Corrosion and Degradation of Implant Materials, 2nd Symp, ASTM STP 859, A. C. Franker and C. D. Griffin Eds., ASTM, Philadelphia 1985 pp 351

[13] Dumbleton J. H., and Shen C., "The Wear Behavior of Ultra High Molecular Weight Polyethylene," Wear, 37, 279 (1976)

[14] Bragdon C. R., et al, "The Importance of Multidirectional Motion for the Wear of Polyethylene in the Hip," 5th World Biomat. Cong., May 28, Toronto, Canada, I-582 (1996)

[15] Wang A., et al, "Wear Testing Based on Unidirectional Motion: Fact or Artefact?," 5th World Biomat. Cong., May 28 Toronto, Canada p I-583 (1996)