Wear 259 (2005) 984991

The effect of counterface roughness on the wear of UHMWPE

for rectangular wear paths

M.E. Turell
a
,
b
, G.E. Friedlaender
b
, A. Wang
c
, T.S. Thornhill
a
, A. Bellare
a
,

a
Department of Orthopaedic Surgery, Brigham and Womens Hospital, Harvard Medical School, Boston, MA 02115, USA
b
Department of Orthopaedics and Rehabilitation, Yale University School of Medicine, New Haven, CT 06510, USA
c
Stryker Howmedica Osteonics, 300 Commerce Court, Mahwah, NJ 07430, USA
Received 1 August 2004; received in revised form 30 December 2004; accepted 31 January 2005
Available online 14 March 2005


Abstract

Ultra-high molecular weight polyethylene (UHMWPE) is used worldwide as a bearing material in total joint replacement prostheses.
Multi-directional motion has been identied as a major factor affecting the wear rate of UHMWPE in total hip replacement prostheses. The
trajectory of relative motion between a femoral head and an acetabular cup takes a general quasi-elliptical or rectangular shape during the
patients gait cycle. Differences in motion pattern can affect the in vivo wear rates of UHMWPE cups in patients when all other factors are
equal. In a previous study that utilized smooth CoCr counterfaces, we compared wear factors, k, for UHMWPE articulated in a series of
rectangular wear paths (width = A, length = B) with systematically increasing aspect ratios (B/A) and linear tracking (A = 0), all with identical
path lengths per cycle. The results showed that the wear factor signicantly decreased in the rectangular wear path with the highest aspect
ratio and in linear tracking. The goal of our current study was to quantify the effect of a roughened counterface on the cross-path wear of
UHMWPE. UHMWPE pins were articulated against both smooth and rough CoCr disks in diluted calf serum using a multi-directional
wear tester under physiological loading conditions. Five different rectangular wear path geometries and linear tracking, all with identical path
lengths per cycle, were employed for each wear test. Gravimetric weight loss was converted into volumetric wear rates and wear factors, k.
The results showed that roughened counterfaces produced a larger increase in the wear factor in rectangular wear paths with higher aspect
ratios. The ratio of krough/ksmooth decreased monotonically as a function of increasing width of rectangles, normalized by total path length, or
A/(A + B).
2005 Elsevier B.V. All rights reserved.

Keywords: Ultra-high molecular weight polyethylene; Multi-directional wear; Total joint replacement; Joint arthroplasty



1. Introduction

Ultra-high molecular weight polyethylene (UHMWPE)
has been used worldwide as a bearing surface of total joint
replacement prostheses since 1962. The high entanglement
density associated with high molecular weight imparts su-
perior fracture toughness [1,2] and high wear resistance to
UHMWPE, making it an attractive choice as a bearing surface

Corresponding author. Present address: Orthopaedic Research Labora-

tory, MRB 106, Brigham and Womens Hospital, 75 Francis Street, Boston,
MA 02115, USA. Tel.: +1 617 732 5864; fax: +1 617 732 6705.
E-mail address: anuj@alum.mit.edu (A. Bellare).

0043-1648/$ see front matter 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2005.01.050
in both total knee and total hip joint replacement prostheses.
Despite its superior mechanical and tribological properties,
wear damage of UHMWPE continues to limit the clinical
lifespan of implanted knee and hip joint prostheses. Volumet-
ric wear rate of an acetabular cup articulating against a 32 mm
femoral head is approximately 80 mm
3
/year, which translates
into an average linear wear of 0.1 mm annually [36]. The
thickness of UHMWPE components is usually 1 cm or more.
Therefore, it would take 100 years for UHMWPE compo-
nents to wear through. While this estimate eliminates the
likelihood that UHMWPE implant components are in any
danger of wearing through over the course of a patients life-
time, billions of sub-micron wear debris particles are pro-

P(2A + 2B)

P(2A)


P(2A + 2B)

A

k =
V




M.E. Turell et al. / Wear 259 (2005) 984991 985

duced annually [7,8], which continues to be the primary con-
cern in total hip replacement prostheses. The accumulation
of particulate debris in the peri-prosthetic tissue can elicit a
biological response leading to bone loss (osteolysis) and im-
plant loosening, necessitating complicated revision surgery
to replace the implant [911].
Osteolysis associated with particulate wear debris has
prompted numerous investigations, which study the size,
shape and morphology of particulate debris and the biolog-
ical pathway that leads to particle-induced osteolysis. Wear
debris collected from peri-prosthetic tissue retrieval studies
showed that the majority of debris particles associated with
hip replacement prostheses were micron or sub-micron in size
and that particles could generally be categorized as brous or
particulate in shape [1216]. Particulate wear also occurs in
total knee replacement (TKR) components, although the par-
ticle size distribution differs from that of total hip replacement
(THR) wear particles, and TKR particles have been shown to
be generally larger than THR wear particles [17].
Computational and experimental studies have shown that
multi-directional motion or cross-shear motion can affect
wear rate of UHMWPE in total hip and total knee replace-
ments [18,19]. Linear tracking motion, whether unidirec-
tional or reciprocating produces an extremely low wear rate,
and in fact produces two to three orders of magnitude less
wear than is observed clinically in total knee replacement
prostheses [1924]. The higher clinical wear rates of total
knee replacement prostheses can be attributed to both coun-
terface scratching [25,26] as well as the effects of cross-shear
motion. Wang et al. have proposed that in a wear environ-
ment, UHMWPE macromolecules orient preferentially along
the principal direction of sliding [19,2224]. Unlike in lin-
ear tracking, where orientation results in strain hardening of
surface material and ultimately increases wear resistance as
sliding progresses, in multi-directional motion, the wear sur-
face experiences both compressive and shear forces in mul-
tiple directions. As sliding proceeds, the UHMWPE wear
surface may strengthen along the direction of sliding, while
it weakens in the perpendicular direction. It has been demon-
strated that the loci or trajectory of motion at the contact
point between a femoral head of an orthopaedic implant and
an acetabular cup is a quasi-ellipse or an approximate rectan-
gle during a gait cycle [18,19,27]. Some patients have either
more elongated (approximately rectangular) or more closed
(approximately more square-like) motion patterns than oth-
ers [27]. Examples of elongated and closed motion patterns
traced by the contact point between femoral and acetabular
components are illustrated in the examples shown in Fig. 1.
Bennett et al. have postulated that the differences in motion
patterns affect the in vivo wear rates of UHMWPE acetabular
cups in patients where factors such as age, weight and body
proportion were similar, but gait patterns varied widely [27].
The multi-directional sliding, or orientation-softening
wear model, proposed by Wang has been termed the unied
theory of wear for UHMWPE [28]. This theory proposed
that when a femoral head slides against an acetabular cup
Fig. 1. Elongated motion pattern (left) and closed motion pattern (right)
traced by the contact point between the formal and acetabular components
of orthopaedic implants.

along the wear path dened by the rectangular loop, fric-
tional energy is dissipated in both the A and B directions (see
Fig. 1). Since A < B, B denes the principal direction of slid-
ing motion while A is secondary. Previous studies indicated
that motion in the principal sliding direction, B, leads to plas-
tic deformation or macromolecular orientation whereas mo-
tion in the secondary direction, A, leads to material removal
by fracture [19,2224,28]. Therefore, only energy released
in the A direction is directly responsible for wear [28]. The
expression for wear rate can then be expressed as:
V mP(2A) (1)
where V is volumetric wear rate, is the coefcient of fric-
tion, P is the applied normal load and 2A is the sliding dis-
tance in the secondary direction per cycle. Since the wear
coefcient or wear factor, k, is dened as (volumetric wear
rate)/(load total sliding distance), in the case of a rectan-
gular wear path illustrated above, k can be written as:

A + B
(2)

Eq. (2) theoretically quanties the effect of cross-path mo-
tion on the wear factor of UHMWPE. In a previous set of
experiments, we compared wear factors for UHMWPE artic-
ulated in a series of graded wear paths in which the aspect
ratio of the wear path was systematically increased. These
wear tests were conducted using a cobaltchromium articula-
tion surface with an implant-grade smooth nish. The results
showed that the wear factors from these experiments were in
agreement with the proposed model only at high aspect ra-
tios, i.e. wear factors were found to signicantly decrease in
the rectangular wear path with the highest aspect ratio and in
linear tracking [29]. A question that remains to be answered
is whether the proposed model remains valid under condi-
tions of abrasive wear. Studies have shown that the addition
of bone cement (with zirconium and barium sulfate additives)
and bone particles to test serum lubricant produces signi-
cantly greater surface damage to stainless steel articulation
counterfaces which in turn results in surface roughening and
increased wear rates [25]. Similarly, other authors have found
that roughening of the femoral head, to a degree seen typically
in retrieval specimens, can increase the observed variability
of volumetric wear rates approximately seven-fold. This fact
may explain why random femoral head scratching in vivo
accounts for otherwise difcult to explain variations in wear
rates as abrasive wear may be a key factor causing excessive
wear in the most problematic subset of the patients with to-



986 M.E. Turell et al. / Wear 259 (2005) 984991

tal joint replacements [26]. The purpose of the current study
is to determine whether the unied theory of wear is appli-
cable to conditions of abrasive wear, such as the case of a
roughened counterface articulating against UHMWPE along
rectangles of various aspect ratios. It was hypothesized that
under conditions of abrasive wear, the model would in gen-
eral describe the cross-path wear of UHMWPE as accurately
as it had under conditions in which a smooth counterface was
employed.


2. Materials and methods

2.1. UHMWPE starting material

Commercially available, ram-extruded GUR 1050
(Hoechst-Ticona, Bayport, TX) rod stock (PolyHi Solidur,
Ft. Wayne, IN) was used as the starting material for all wear
tests. Rod stock with a diameter of 7.6 cm was machined into
cylindrical pins with overall dimensions of 20 mm in length
and 9 mm in diameter for all wear tests. All pins were tested
in the as-machined condition (no sterilization).

2.2. Wear testing protocol and apparatus

A series of six different articulation patterns including
a 5 mm 5 mm square, 4 mm 6 mm, 3 mm 7 mm,
2 mm 8 mm, and 1 mm 9 mm rectangles, and a
0 mm 10 mm linear tracking pattern were digitized
into an OrthoPOD
TM
(Advanced Mechanical Technology
Inc.) multi-directional wear tester. For each wear test, the
OrthoPOD
TM
multi-directional wear tester was loaded
with six UHMWPE pins as shown in the experimental
set-up in Fig. 2. All wear tests were conducted at a cycle
frequency of 1 Hz (constant sliding speed of 20 mm/s along
the wear track) and with a constant applied load of 192 N
or an applied stress of 3 MPa, well within the physiological
range of 25 MPa for the hip joint. A bovine serum (JRH
Biosciences) lubricant was used for all wear tests, and
was diluted with distilled water to contain 23 g/L protein,
20 mM EDTA, and 0.2% sodium azide. The lubricating
bovine serum temperature was maintained at 37

C using
a re-circulating water bath. Cobaltchromium disks with
dimensions of 25 mm diameter and 3 mm thickness were
used as the articulation counterface. Two series of wear tests
using each of the six experimental articulation patterns were
performed. In one series of tests, cobaltchromium disks
were polished to implant-grade surface smoothness with
a centerline roughness of 0.015 m (Ra). In a second set
of experiments, the cobaltchromium disks were scratched
along random directions using 320 grit emery paper in
accordance with the previously established method of Wang
et al., resulting in an average roughness, Ra, of 0.45 m
[30]. A total of six pins plus an additional pin serving as a
soak control were used for all wear tests.

2.3. Data analysis

Wear rates for each of the six articulation patterns were
determined in experiments in which both smooth and rough
counterfaces were employed. Gravimetric weight loss per pin
was determined approximately every 200,000 cycles. The
soak control consistently revealed that the amount of absorp-
tion of bovine serum by the UHMWPE specimen was unde-
tectable and therefore any corrections to compensate for uid
absorption were unnecessary. For each group, wear tests were





















Fig. 2. AMTI OrthoPOD
TM
multi-directional wear tester.



M.E. Turell et al. / Wear 259 (2005) 984991 987

continued until at least one million cycles had been reached.
The gravimetric weight loss per pin was determined by taking
the average weight loss of the six wear tested pins. Prior to
being weighed, UHMWPE pins were rst washed twice with
distilled water followed by an alcohol and acetone rinse, re-
spectively. The pins were dried in air at 25

C for 15 min prior

to determination of weight loss. Gravimetric weight loss was
further converted into volumetric wear data by using a den-
sity value of 0.943 g/cm
3
for UHMWPE. Wear factor values,
k, were also calculated for each sample by dividing the vol-
umetric wear rate by the product of the load and the sliding
distance per cycle.

2.4. Statistics

Statistical analysis was conducted using ANOVA with
Fishers protected least signicant difference (PLSD) post-
hoc test in which a p-value of less than 0.05 was used to dene
statistical signicance. In addition, single-factor ANOVA us-
ing a 95% condence interval ( = 0.05) was also used to
more rigorously assess the statistical signicance between
groups. Wear factor values from each wear path geometry
using both smooth and rough counterfaces were compared
against one another in order to establish the level of signi-
cance between different groups.


3. Results

3.1. Wear testing

Wear tests showed that the wear rates increased incre-
mentally as the wear path decreased in aspect ratio and ap-
proached the 3 mm 7 mm conguration. Interestingly, for
both experiments in which smooth counterfaces and rough
counterfaces were used, wear reached a maximum when a
3 mm 7 mm wear path was employed (Fig. 3). In the smooth
counterface test series, the wear factor for the 5 mm 5 mm
square path was found to be 2.5 times greater in comparison
to the 2 mm 8 mm path (p < 0.05, ANOVA with Fishers
PLSD post-hoc test), which agrees remarkably well with the













Fig. 3. Comparison of wear factors for both smooth and rough
cobaltchromium counterfaces.
prediction of Eq. (2). For the same comparison in the rough
counterface test series, the wear factor for the 5 mm 5 mm
square path was found to be only 1.1 times greater than
the wear factor for the 2 mm 8 mm path. Although this
trend was observed, in the case of the rough counterface
test, the differences between the 5 mm 5 mm path and the
2 mm 8 mm path was not statistically signicant.

3.2. Effect of counterface roughness

As predicted, counterface roughness had a substantial ef-
fect on wear rate. Although standard deviations between
groups were large, the differences in wear rates between
rough and smooth tests for each wear path geometry were
statistically signicant (p < 0.05) according to ANOVA (us-
ing a 95% condence interval). This is due to the large
number of data points included in each group. An ex-
ception is that the differences in the wear factor for the
4 mm 6 mm rough versus smooth tests were not statisti-
cally signicant. Using ANOVA (with the 95% condence
interval criteria) within the smooth counterface series, the
wear factor of the 1 mm 9 mm wear path was signicantly
greater than that of the 0 mm 10 mm path as was the
wear factor of the 3 mm 7 mm versus the 2 mm 8 mm.
However, the wear factors of 2 mm 8 mm versus the
1 mm 9 mm, the 4 mm 6 mm versus the 3 mm 7 mm,
and the 5 mm 5 mm versus the 4 mm 6 mm were not sig-
nicantly different from one another. Using this same anal-
ysis, in the rough counterface set of wear tests, the wear
factor associated with the 1 mm 9 mm path was signi-
cantly greater than that of the 0 mm 10 mm path as was
the 3 mm 7 mm path compared to the 4 mm 6 mm path.
The wear factor of the 2 mm 8 mm versus the 1 mm 9 mm
path, the 3 mm 7 mm versus the 2 mm 8 mm path, and the
5 mm 5 mm versus the 4 mm 6 mm path were not statis-
tically different. An important nding of this study was that
the counterface roughness affected wear rates to a greater ex-
tent in experiments in which more linear rectangular paths
were used as compared to the more square-like wear paths.
In comparison to the smooth counterface series, the rough
counterface tests demonstrated wear rates that were signif-
icantly greater than predicted when more linear wear paths
were tested. There was a general decreasing trend in the ratio
of krough/ksmooth as the aspect ratio of the rectangles decreased
(approached the square conguration), as shown in Fig. 4.

3.3. Unied theory of wear

In order to assess the accuracy of the unied theory of
wear in predicting the wear rate of UHMWPE based on a
particular wear path geometry, the experimentally observed
wear factor values for both the smooth and rough counterface
tests were plotted as a function of the ratio A/(A + B) (Fig. 5).
These curves were then compared to predicted wear factor
trends, in which the maximum predicted wear factor was
taken to be the value of the wear factor obtained from the



988 M.E. Turell et al. / Wear 259 (2005) 984991


















Fig. 4. Comparison of wear factor ratios (rough/smooth).

5 mm 5 mm square wear path. Analysis revealed that the
orientation-softening theory of wear was unable to accurately
predict wear over the entire range of rectangles for both the
smooth and rough counterface series. In smooth counterface
tests, the parameter A/(A + B) was particularly accurate in the
prediction of wear rates in cases where the aspect ratio of the
rectangular path was highest, such as in the 0 mm 10 mm
to 3 mm 7 mm wear paths. The monotonic increase in wear
factors in the A/(A + B) range of 0.0 and 0.3 for UHMWPE
in both rough and smooth counterfaces suggested a linear
correlation between wear factor, k, and A/(A + B). Therefore,
a linear curve t was conducted in a separate plot (Fig. 6)
for the wear factors in the A/(A + B) range of 0.00.3. Fig. 6
shows that the wear factor values from both the smooth and
rough counterface tests show a strong linear relationship in
this range.
In both the smooth and rough counterface experiments,
the wear factor increased with increasing A/(A + B) ratio, but
only up to a value of 0.3. There were signicant deviations in
the experimental data from the theoretical linear correlation
predicted by the orientational softening theory (Fig. 5). These
differences are particularly apparent in the set of experiments













Fig. 5. Wear factor, k, vs. A/(A + B) ratio.
Fig. 6. Liner t (with maximum A/(A + B) = 0.3) for wear factor, k, vs.
A/(A + B) ratio.

employing rough counterfaces. According to the theoretical
model of Wang [28], the maximum wear factor should occur
at A = B or A/(A + B) = 0.5. Instead the experimentally ob-
served wear factor peaked at A/(A + B) = 0.3 for both smooth
and rough counterface experiments. In fact, the 3 mm 7 mm
showed a 16% increase over the 5 mm 5 mm square path
in the series where smooth counterfaces were used (p < 0.05,
ANOVA with Fishers PLSD post-hoc test).


4. Discussion

4.1. Wear tests

This study was conducted as an extension of a previous
study that sought to quantify the effect of the cross-path mo-
tion on wear rate of UHMWPE. The aim of this study was
to address the effects of counterface roughness on wear rates
for different cross-path motions. For a rectangular wear path
with a width A and a length B (A < B), the numerical param-
eter A/(A + B) is a convenient measure to relate wear factor,
k, to cross-path motion for rectangles with various aspect ra-
tios, B/A. This numerical parameter, dened by the unied
wear model, was relatively accurate in predicting the gen-
eral trend that wear rates would increase as a function of
decreasing aspect ratio (i.e. as a function of the wear path
approaching a square conguration) but only in the range
of the 1 mm 9 mm wear path to the 3 mm 7 mm con-
guration. For both smooth and rough counterface tests, a
linear trend was particularly evident in cases where the as-
pect ratio of the rectangular path was highest, such as in
the 0 mm 10 mm, 1 mm 9 mm, and 2 mm 8 mm wear
paths. In both counterface groups, wear factors reached a
maximum when a 3 mm 7 mm wear path was employed,
and the trend was no longer observed as the aspect ratio of
the wear path continued to decrease beyond this point. The
failure of the unied wear model to predict wear rates in some
cases for both rough and smooth counterface tests raises ques-
A + B

6E08

(
R

=

0
.
8821)

A + B
+

4E07

(
R

=

0
.
9922)




M.E. Turell et al. / Wear 259 (2005) 984991 989

tions about some of the underlying assumptions upon which
the model is based. For example, for any rectangular wear
path with sides A and B, where A < B, the model assumes
that orientation on the molecular level occurs in the B di-
rection, while fracture, or the actual wearing of UHMWPE,
occurs in the A direction. Based on the results of this study, in
which wear reached a maximum for the 3 mm 7 mm wear
path, this assumption may not be a valid description of the
orientation and wear processes. It is possible that complete
orientation and strain hardening of UHMWPE in the B direc-
tion is required for wear to occur solely in the A direction,
and that this is only achieved at 7 mm of sliding. It is there-
fore possible that for the 5 mm 5 mm and 4 mm 6 mm
wear paths, there is signicant reorientation in the A direc-
tion as well rather than fracture of the fully oriented, strain
hardened UHMWPE brils. Reorientation would imply that
a different wear mechanism is operative since biaxial ori-
entation prior to wear would suggest that a more sheet-like
material wears rather than splitting of brillar UHMWPE.
These wear and orientation mechanisms appeared to occur in
two discrete A/(A + B) ranges, irrespective of the roughness
of counterface.
Fig. 6 shows that there was a linear correlation between
wear factor, k, and the numerical parameter A/(A + B), regard-
less of counterface roughness. The linear equations for these
correlations as well as the R-values (to measure the degree of
t) were obtained for both the smooth and rough counterface
tests. In the equation for the line, k = m[A/(A + B)] + c, the
constant c (y-intercept) represents the fraction of wear factor
that is due to linear, abrasive wear and the constant m (slope)
is a measure of the dependence of cross-path wear on the
wear path geometry. The following equations were obtained
for the linear correlations:
k = 7E06
A


(smooth counterface) (3)

k = 1E05
A


(rough counterface) (4)
It is evident that the constant c cannot assume a negative
value, and the low, negative wear factor value of 6E08
for linear tracking for smooth heads obtained from the lin-
ear t is merely a consequence of experimental error, which
can be taken to be zero or replaced by the experimentally
measured positive value of 5.06E08. The value of m was
higher in the rough counterface series of wear tests. The
ratio of mrough/msmooth was 1.4, revealing that there was a
steeper dependence of wear factor, k, on the numerical pa-
rameter, A/(A + B), in the case of rough counterface. For both
the smooth and rough counterface, the high R-values indi-
cated a strong linear correlation between the wear factor and
A/(A + B). The good t of the wear factors in the series of
wear paths in the range of 0 mm 10 mm to 3 mm 7 mm
paths suggests that complete orientation followed by frac-
ture of brils was the primary mechanism of wear in both
of these cases. Independent of the accuracy with which the
model investigated in this paper predicted wear rate, the over-
all experimental results support the hypothesis that wear rate
is in fact dependent upon the wear path geometry, counterface
roughness and the interplay of these two variables.

4.2. Clinical relevance

The nding that wear rate is dependent upon both wear
path pattern and counterface roughness has a number of clin-
ically relevant implications. The observation that differences
in wear rates, between tests employing rough versus smooth
counterfaces, are greater in more linear motion path patterns
and that these differences systematically decrease as the wear
approaches a square pattern is an important nding. In total
joint replacement applications where linear wear is known to
be operative and to have a signicant impact upon the life-
time of the joint replacement, for example, as is the case in
the knee joint, the effects of abrasive wear (simulated by a
roughened counterface in this study) are of greater concern.
It should be noted that the results of this study reveal a some-
what oversimplication of the wear mechanisms as they oc-
cur in clinical application since the effects of third body wear
and fatigue-related wear mechanisms were not investigated.
A question which remains to be answered is whether it may
be possible to predict motion path patterns by conducting
gait analysis and if so, whether motion path pattern can be
therapeutically manipulated using methods of gait training or
by making improvements to implant design. A challenge in
addressing these questions lies in the difculty in assessing
gait in patients who require total joint replacement surgery
as gait in these individuals may be altered from their normal
baseline gait due to the orthopaedic complications, which ne-
cessitate surgery in the rst place. Similarly, gait patterns in
these individuals may be dramatically altered following total
joint replacement surgery, making it difcult to predict the
effects of implant design in advance of surgery.

4.3. Limitations

The results of this study represent an attempt to quantify
the effect of the motion path pattern and counterface rough-
ness on wear rate of UHMWPE. It should be noted that the
data, in particular the wear factor values, that have been ana-
lyzed in this paper represent the results of preliminary wear
tests conducted for each wear path geometry to a period of at
least one million cycles. A more comprehensive study, and
one that would employ more rigorous wear testing would
generate a larger number of samples for tests encompassing
a broader range of motion path patterns. From such experi-
ments, trends in the values of UHMWPE wear rates would
be more reliably generated and would be of a greater level
of clinical signicance. A study of the morphology of wear



990 M.E. Turell et al. / Wear 259 (2005) 984991

particles generated from the various motion path patterns for
both smooth and rough counterface tests is also necessary
(and forthcoming) to obtain a more accurate understanding
of the wear mechanisms that are operative during the various
wear paths. In addition, the wear tests in this paper employed
a constant applied load and it is important for future studies
to include wear test loading parameters which more closely
model the loading conditions found in knee and hip joints
in vivo. The relationship between wear path and wear rates
established in this study applies only to uncrosslinked med-
ical grade UHMWPE. Further wear testing on the various
crosslinked UHMWPEs currently in clinical use is required
for such relationships to be established.


5. Conclusions

Wear of UHMWPE as it articulates against a metallic
counterface in a particular wear path and under conditions
that mimic wear in a total hip replacement prosthesis, likely
occurs via two discrete steps for rectangles with an aspect
ratio greater than 2.33 (3 mm 7 mm rectangular path). For
such rectangles, the wear tests of this study support the hy-
pothesis that there is orientation or texturing of UHMWPE
on parallel edges of the rectangle followed by wear of the
textured UHMWPE on the other two parallel edges. How-
ever, for rectangles in the aspect ratio range of 1.02.33
(5 mm 5 mm to 4 mm 6 mm paths), the decline in wear
rates for both smooth and rough counterface experiments
provide evidence that wear and orientation processes may
not occur in discrete stages. In addition, the unied wear
model predicts zero wear for linear tracking, which is not
the case, especially when more abrasive conditions of wear
occur such as the case of a roughened counterface. A more
robust model is required to predict wear of UHMWPE during
articulation against a metallic counterface along a rectangu-
lar path covering the entire range of aspect ratios of rectan-
gles.


Acknowledgments

This project was funded through a biomedical engineering
grant provided by the Whitaker Foundation and by a fellow-
ship provided by the Orthopaedic Research and Education
Foundation.


References

[1] A. Gomoll, T. Wanich, A. Bellare, J-integral fracture toughness and
tearing modulus measurement of radiation cross-linked UHMWPE,
J. Orthop. Res. 20 (6) (2002) 11521156.
[2] L.C. Duus, H.A. Walsh, A.M. Gillis, E. Noisiez, S. Li, The effect of
resin grade, manufacturing method and cross linking on the fracture
toughness of commercially available UHMWPE, Trans. Orthop. Res.
Soc. 25 (2000) 544.
[3] M. Grifth, M. Seidenstein, D. Williams, J. Charnley, Socket wear
in Charnley low friction arthroplasty of the hip, Clin. Orthop. 137
(1978) 3747.
[4] J. Livermore, D. Ilstrupand, B. Morrey, Effect of femoral head size
on wear of the polyethylene acetabular component, J. Bone Joint
Surg. 72A (4) (1990) 518528.
[5] R. Hall, A. Unsworth, P. Siney, B. Wroblewski, Wear in retrieved
Charnley acetabular sockets, Proc. Inst. Mech. Eng. H: J. Eng. Med.
210 (1996) 197207.
[6] H. Oonishi, E. Tsuji, Y. Kim, Retrieved total hip prostheses. Part I.
The effects of cup thickness, head size and fusion defects on wear,
J. Mater. Sci.: Mater. Med. 9 (1998) 394401.
[7] I.C. Clarke, J.M. Kabo, Wear in total hip replacement, in: H.C.
Amstutz (Ed.), Hip Arthroplasty, Churchill Livingstone, New York,
1991.
[8] G. Lewis, Polyethylene wear in total hip and knee arthroplasties, J.
Biomed. Mater. Res. (Appl. Biomater.) 38 (1997) 5575.
[9] H. Amstutz, P. Campbell, N. Kossovsky, I. Clarke, Mechanisms and
clinical signicance of wear debris-induced osteolysis, Clin. Orthop.
Relat. Res. 276 (1992) 718.
[10] W. Harris, The problem is osteolysis, Clin. Orthop. Relat. Res. 311
(1995) 4653.
[11] T. Schmalzried, L. Kwong, M. Jasty, R. Sedlacek, T. Haire, D.
OConnor, C. Bragdon, J. Kabo, A. Malcolm, W. Harris, The mech-
anism of loosening of cemented acetabular components in total hip
arthroplasty: analysis of specimens retrieved at autopsy, Clin. Orthop.
Relat. Res. 274 (1992) 6078.
[12] P. Campbell, S. Ma, B. Yeom, H. McKellop, T. Schmalzried, H.
Amstrutz, Isolation of predominantly submicron-sized UHMWPE
wear particles from periprosthetic tissues, J. Biomed. Mater. Res.
29 (1995) 127131.
[13] A. Shanbhag, J. Jacobs, T. Glant, J. Gilbert, J. Black, J. Galante,
Composition and morphology of wear debris in failed uncemented
total hip replacement, J. Bone Joint Surg. 76B (1994) 6067.
[14] J. Savio, L. Overcamp, J. Black, Size and shape of biomaterial wear
debris, Clin. Mater. 15 (1994) 101119.
[15] K. Margevicius, T. Bauer, J. McMahon, S. Brown, K. Merritt, Isola-
tion and characterization of debris in membranes around total joint
prostheses, J. Bone Joint Surg. 76A (1994) 16641676.
[16] P. Campbell, P. Doorn, F. Dorey, H. Amstutz, Wear and morphology
of ultra-high molecular weight polyethylene wear particles from total
hip replacements, Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996)
167175.
[17] A. Shanbhag, H. Bailey, D. Hwang, C. Cha, N. Eror, H. Rubash,
Quantitative analysis of ultrahigh molecular weight polyethylene
(UHMWPE) wear debris associated with total knee replacements,
J. Biomed. Mater. Res. 53 (1) (2000) 100110.
[18] B. Ramamurti, C. Bragdon, D. OConnor, J. Lowenstein, M. Jasty, D.
Estok, W. Harris, Loci of movement of selected points on the femoral
head during normal gait, J. Arthroplasty 11 (7) (1996) 852855.
[19] A. Wang, D. Sun, S.-S. Yau, B. Edwards, M. Sokol, A. Essner, V.
Polineni, C. Stark, J. Dumbelton, Orientation softening in the defor-
mation and wear of ultra-high polyethylene, Wear 203204 (1997)
230241.
[20] V. Saikko, T. Ahlroos, Type of motion and lubricant in wear simu-
lation of polyethylene acetabular cup, Proc. Inst. Mech. Eng. H: J.
Eng. Med. 213 (1999) 301310.
[21] C. Bragdon, D. OConnor, J. Lowenstein, M. Jasty, W. Syniuta, The
importance of multidirectional motion on the wear of polyethylene,
Proc. Inst. Mech. Eng. H: J. Eng. Med. 210 (1996) 157166.
[22] A. Wang, C. Stark, J. Dumbleton, Mechanistic and morphological
origins of ultra-high molecular weight polyethylene wear debris in
total joint replacement prostheses, Proc. Inst. Mech. Eng. H: J. Eng.
Med. 210 (1996) 141155.
[23] A. Wang, V. Polineni, A. Essner, M. Sokol, D. Sun, C. Stark, J. Dum-
bleton, The signicance of non-linear motion in the wear screening
of orthopaedic implant materials, Test. Eval. 25 (2) (1997) 239245.



M.E. Turell et al. / Wear 259 (2005) 984991 991

[24] A. Wang, A. Essner, V. Polineni, C. Stark, J. Dumbleton, Lubrication
and wear of ultra-high molecular weight polyethylene in total joint
replacements, Tribol. Int. 31 (13) (1998) 1733.
[25] L. Caravia, D. Dowson, J. Fisher, B. Jobbins, The inuence of
bone and bone cement debris on counterface roughness in slid-
ing wear tests of ultra-high molecular weight polyethylene on stain-
less steel, Proc. Inst. Mech. Eng. H: J. Eng. Med. 204 (1) (1990)
6570.
[26] T.D. Brown, K.J. Stewart, J.C. Nieman, D.R. Pedersen, J.J.
Callaghan, Local head roughening as a factor contributing to vari-
ability of total hip wear: a nite element analysis, J. Biomech. Eng.
124 (6) (2002) 691698.
[27] D. Bennett, J. Orr, R. Baker, Movement loci of selected points on
the femoral head for individual total hip arthroplasty patients using
three-dimensional computer simulation, J. Arthroplasty 15 (7) (2000)
909915.
[28] A. Wang, A unied theory of wear for ultra-high molecular weight
polyethylene in multidirectional sliding, Wear 248 (2001) 3847.
[29] M. Turell, A. Wang, A. Bellare, Quantication of the effect of
cross-path motion on the wear rate of ultra-high molecular weight
polyethylene, Wear 255 (2003) 10341039.
[30] A. Wang, V.K. Polineni, C. Stark, J.H. Dumbleton, Effect of femoral
head surface roughness on the wear of ultrahigh molecular weight
polyethylene acetabular cups, J. Arthroplasty 6 (1998) 615620.