S. M. T. Chan

C. P. Neu1

Center for Tissue Regeneration and Repair and

Department of Orthopaedic Surgery,

School of Medicine,

University of California, Davis,

Sacramento, CA 95817

K. Komvopoulos2

Fellow ASME,

Professor,

Department of Mechanical Engineering,

University of California,

Berkeley, CA 94720

e-mail: kyriakos@me.berkeley.edu

A. H. Reddi

P. E. Di Cesare

Center for Tissue Regeneration and Repair and

Department of Orthopaedic Surgery,

School of Medicine,

University of California, Davis,

Sacramento, CA 95817

Friction and Wear ofHemiarthroplasty Biomaterialsin Reciprocating Sliding ContactWith Articular CartilageFriction and wear of four common orthopaedic biomaterials, alumina (Al2O3), cobalt-chromium (CoCr), stainless steel (SS), and crosslinked ultra-high-molecular-weightpolyethylene (UHMWPE), sliding against bovine articular cartilage explants were inves-tigated by reciprocating sliding, nanoscale friction and roughness measurements, proteinwear assays, and histology. Under the experimental conditions of the present study, CoCryielded the largest increase in cartilage friction coefficient, largest amount of proteinloss, and greatest change in nanoscale friction after sliding against cartilage. UHMWPEshowed the lowest cartilage friction coefficient, least amount of protein loss, and insignif-icant changes in nanoscale friction after sliding. Although the results are specific to thetesting protocol and surface roughness of the examined biomaterials, they indicate thatCoCr tends to accelerate wear of cartilage, whereas the UHMWPE shows the bestperformance against cartilage. This study also shows that the surface characteristics ofall biomaterials must be further improved to achieve the low friction coefficient of thecartilage/cartilage interface. [DOI: 10.1115/1.4004760]

Keywords: articular cartilage, biomaterials, friction, hemiarthroplasty, wear

1 Introduction

Nonbiological materials are commonly used in contact withnative articular cartilage to restore load-bearing function to jointsfor a variety of orthopaedic conditions. Hemiarthroplasty is com-monly performed to treat femoral neck fractures and osteonecrosisof the hip and to manage degenerative joint disease or fracture ofthe shoulder. In knee arthroplasty, many surgeons elect not toresurface the patella, leaving the native articular cartilage in con-tact with the femoral component. Advantages of hemiarthroplastyover total joint replacement include a less invasive surgical proce-dure, reduced occurrence of dislocation, better range of motion,faster recovery for the patient, and preservation of normal carti-lage and bonestock [1–4]. However, because the artificial surfacearticulates against native articular cartilage, there is much concernfor wear of the tissue, which could eventually lead to conversionto total joint replacement [2,4–7]. Given that over 46� 106 Amer-icans suffer some form of arthritis [8], while 44� 106 have osteo-porosis, with 1.5� 106 of them sustaining bone fractures eachyear from this condition [9], there is an imperative need to mini-mize complications associated with wear to improve the longevityof both the cartilage and the implant.

A key factor in determining the longevity of the implant is thefriction and wear properties of the structural material [4,7]. Earlytribological studies of hemiarthroplasty biomaterials primarilyfocused on acetabular erosion, based on radiological and histologi-cal evaluations [5,6], which are only qualitative and possiblysubjective. The friction coefficient of articular cartilage againstcobalt-chromium alloy [7], pyrolytic carbon [7], steel [10],ceramics [10,11], glass [4,12], and polymers [4] has been measuredmore recently. However, direct comparisons between published

data are not possible due to widely varying techniques and materi-als, ranging from in vivo canine animal models to in vitro frictionmeasurements using different tissue sources, lubrication baths, testconfigurations, and counterface materials. Although histologicalevaluations of wear for hip hemiarthroplasties have been reported[5,11], only one study has quantified wear in terms of changes incartilage explant thickness after friction testing against zirconia andcobalt-chromium [11]. There remains a lack of comprehensive andquantitative data regarding wear of articular cartilage in hemiar-throplasties. The material properties (e.g., hardness) and surfacecharacteristics (e.g., roughness and wettability) also play importantroles in the wear of articular cartilage.

Wear of native articular cartilage after hemiarthroplasty is notwell understood despite the development of better biomaterials andsurgical techniques. Wear may be defined as the removal of mate-rial from solid surfaces due to mechanical action induced by surfacesliding, and includes various mechanisms such as adhesion, abra-sion, fracture, fatigue, delamination, and corrosion. Thus, despitethe advantages of hemiarthroplasty over total joint replacement,many studies recommend a total joint replacement, even when thecartilage appears normal, due to the high potential for wear andneed for later revision [1,13–15]. A better understanding of theunderlying reasons and the extent that native cartilage wears againsta foreign material would greatly aid the development of newimplants that might eliminate or decrease these complications.

The purpose of this study was to investigate the friction and wearcharacteristics of cartilage sliding against four common hemiarthro-plasty biomaterials, namely alumina (Al2O3), cobalt-chromium(CoCr) alloy, stainless steel (SS), and, to a limited extent, ultra-high-molecular-weight polyethylene (UHMWPE). The frictioncoefficient of cartilage slid against each biomaterial was measuredwith a pin-on-disk tribometer in reciprocating sliding mode. Wearwas assayed for total protein and amount of superficial zone protein(SZP), a marker of the articular cartilage surface zone believed tobe a key boundary lubricant [16–18], removed from the surfaceduring sliding. Friction coefficients and surface roughness of thebiomaterials were measured before and after sliding against

1Present address: Weldon School of Biomedical Engineering, Purdue University,West Lafayette, IN, 47907.

2Corresponding author.Contributed by Tribology Division of ASME for publication in the JOURNAL of

TRIBOLOGY. Manuscript received December 29, 2010; final manuscript received June21, 2011; published online October 10, 2011. Assoc. Editor: Thierry Blanchet.

Journal of Tribology OCTOBER 2011, Vol. 133 / 041201-1Copyright VC 2011 by ASME

cartilage with an atomic force microscope (AFM), which was alsoused to examine material transfer and the development of weartracks on the sliding surfaces. The main objectives for these experi-ments were to measure the friction coefficient of each biomaterial/cartilage interface, quantify cartilage wear generated by slidingagainst each biomaterial, and determine if surface rubbing againstcartilage causes changes in the surface characteristics of thebiomaterials.

2 Materials and Methods

2.1 Tissue Acquisition. Articular cartilage was harvestedfrom 1–3 week old bovine stifle joints obtained from a local abat-toir within 16 h of sacrifice. Joints were opened using aseptic tech-nique, and one osteochondral explant was harvested from theanterior region of each distal femoral medial condyle [19]. Thislocation was chosen for consistent harvest of samples from load-bearing regions of the joint [19–21], enabling the study of articularcartilage at locations more susceptible to accelerated wear in hemi-arthroplasties. In addition, because this region of the stifle joint ischaracterized by high levels of SZP expression [19,22], it facilitatesthe quantification of protein loss due to mechanical wear. A 5-mm-diameter coring reamer was used to extract the explants, and anadjustable custom jig to trim the height of the explants to 4 mm.Before testing, the explants were immersed in Dulbecco’s modifiedEagle’s medium (DMEM)/F12 culture supplemented with 1%penicillin/streptomycin (both from Invitrogen, Carlsbad, CA), and0.2% bovine serum albumin and 50 lg/mL ascorbic acid 2-phos-phate (both from Sigma-Aldrich, St Louis, MO).

2.2 Macroscale Friction and Wear Experiments. To deter-mine the dependence of the coefficient of friction of each bioma-terial/cartilage system on sliding time, articular cartilage plugs(n¼ 7 for each biomaterial) were slid against each hemiarthro-plasty biomaterial on a pin-on-disk tribometer operated in recipro-cating sliding mode. The pin specimens consisted of explantsaffixed to acrylic pins by ethyl cyanoacrylate. The articular sur-face was brought into contact with one of four biomaterials:Al2O3, CoCr, SS, and crosslinked UHMWPE. Specimens wereprepared as disks with their surfaces finished by the supplier (Mer-ton Tech, Paterson, NJ) to an average roughness of less than 0.254lm. Chemical compositions and mechanical properties of eachbiomaterial are given in Table 1. Before testing, each disk wassonicated in phosphate buffer saline (PBS). Cartilage explantswere completely immersed in 10 mL of PBS throughout the dura-tion of testing to maintain tissue hydration. Prior to the initiationof testing, each explant was allowed to equilibrate [23] for 2 minunder the applied normal load [19,24]. Data were acquired in 0.1 sintervals for 60 min using Labview (National Instruments, Austin,TX). All the experiments were carried out under the followingconditions: 0.5 mm/s sliding speed, 5 mm wear track radius,�7.85 mm stroke length, and 1.8 N normal load, resulting in anaverage contact pressure of 0.1 MPa, which is in the low range ofphysiological contact pressure. Under these conditions, slidingoccurred in the boundary lubrication regime, as determined by theminimum film thickness calculated for highly loaded, rigid cylin-drical contacts immersed in an isoviscous fluid [25–27].

The friction coefficient of each biomaterial/cartilage interfacewas averaged in each minute of testing and plotted versus time ofsliding. The initial and final friction coefficients measured after 1and 60 min from the onset of sliding, respectively, were used todetermine changes in the friction properties of each biomaterial/cartilage interface due to sliding. Statistical analysis was performedusing a two-way ANOVA and Bonferroni/Dunn post hoc tests witha significance level a¼ 0.05 (corresponding to p¼ 0.0083) and astandard software package (SAS Institute, Cary, NC).

2.3 Total Protein, SZP, and Wear Quantification. Imme-diately following testing, the explant was removed from thetribometer and rinsed twice with 500 lL of PBS to remove anywear debris adsorbed to the tissue surface. The PBS bath was col-lected in 15 mL conical tubes (BD Falcon, San Jose, CA), the diskwas rinsed three more times with PBS, and the fluid was collectedto ensure the collection of all wear debris. Wear samples were fro-zen for further analysis of protein content. The PBS in which weartests were conducted was concentrated by centrifugation usingCentricon filter devices (Millipore, Billerica, MA), and assayedfor total protein using the Bicinchoninic acid (BCA) assay (Pierce,Rockford, IL) and for SZP using enzyme-linked immunosorbentassay (ELISA). Prior to each assay, the actual volume of eachconcentrated sample was measured and used to calculate the totalmass of protein loss. BCA assay was performed on a microplatewith 10 lL of the sample applied into each well. ELISA was per-formed using standard methods and SZP purified from bovine sy-novial fluid as a standard [28]. Samples were serially diluted withPBS and reacted with S6.79 (1:5000) as the primary antibody andantimouse horseradish peroxidase-conjugated goat (1:3000, Vec-tor Laboratories, Burlingame, CA) as the secondary antibody.

Wear was quantified by the mass removed per unit surface areanormalized to the mean contact pressure and total sliding distance(lg/N�m). Conventionally, wear is expressed in terms of the vol-ume change in worn material, as in the classical wear equation[29]. Wear calculations based on mass loss almost always involvemeasurements of the dry mass of the worn material [30,31].Because tissue hydration is essential for articular cartilage viabilityand function, measuring the dry weight of the tissue samples wasimpractical. Therefore, the mass of material removed from the sur-face was estimated from the wear debris contained in the collectedmedium (PBS). The wear rate, expressed as material loss per slid-ing time, was not calculated because the wear debris was onlyassayed after the end of testing. A one-way ANOVA and Fisher’sPLSD post hoc tests with p< 0.05 were used to determine weardifferences between different biomaterial/cartilage interfaces.

2.4 Histology Analysis. To verify the distribution of SZP inthe tissue, two sets of explants for each biomaterial/cartilage inter-face were fixed in Bouin’s fixative overnight for histology immedi-ately following friction and wear testing. An unworn explantimmersed in PBS was used as the positive control. SZP localizationat the articular surface was determined by immunohistochemistry(IHC) using S6.79 (1:5000) as the primary antibody and an ABCkit with mouse IgG as the secondary antibody (Vector Laboratories,Burlingame, CA). Images were obtained with an optical micro-scope (LSM510, Carl Zeiss, Jena, Germany) at 40�magnification.

Table 1 Composition and mechanical properties of biomaterials used for friction and wear testing against articularcartilage

Biomaterial Composition (wt%) Elastic modulus (GPa) Hardness (GPa)

Al2O3 99.8 Al2O3 393.0 19.8CoCr 63 Co, 28 Cr, 5 Mo,< 1 Mn,< 1 Si,< 1 Ni,< 1 Fe,< 1 C 257.0 4.2SS (316L) 0.043 C, 0.72 Mn, 0.034 P, 0.023 S, 0.55 Si, 4.12 Ni,

15.24 Cr, 3.16 Cu, 0.27 Cb, 0.008 Ta, balance Fe206.2 4.3

UHMWPEa (GUR 1050) 61% crystallinity (crosslinked at 75 kGy) 0.99 —

aCrosslinked by a three-time exposure to c-radiation of 3 MRad; annealed below melt to annihilate free radicals; plasma gas sterilized.

041201-2 / Vol. 133, OCTOBER 2011 Transactions of the ASME

2.5 Nanoscale Friction Coefficient and Surface RoughnessMeasurements. To determine changes in the biomaterial surfacedue to sliding against cartilage, the nanoscale friction coefficientand surface roughness of the biomaterials were examined with anAFM (MFP-3D-CF, Asylum Research, Santa Barbara, CA). Trian-gular silicon nitride tips (MSCT-AUNM, Veeco, Santa Barbara,CA) with a nominal tip radius of 10 nm and spring constant of 0.02N/m were used in all scans. Contact-mode scanning of 60� 60lm2 areas was performed in air with a resolution of 128� 128 pix-els, scan rate of 1 Hz, and constant set point of 4 V. Force calibra-tion on glass using Igor Pro software (version 5, WaveMetrics,Oswego, OR) based on the thermal calibration method [32]revealed that this set point corresponded to an applied normal loadof 7.07 6 0.12 nN. Lateral force calibration, performed on a silicongrating (TGF11, Mikromasch, Wilsonville, OR) using the directforce balance method [33], showed that the friction voltage and thefriction force were related by a constant equal to 980.8 nN/V.

Height and lateral force signals were acquired at five differentlocations within the vicinity of the sliding path. An average nano-scale friction force was obtained as the half-width of the frictionloop, i.e., half of the difference between the mean lateral trace andretrace values. Friction force values were divided by the appliedload and then averaged over the five locations. The root-mean-square (RMS) surface roughness, RS, indicating the statisticaldeviation of surface height data from the average surface height,was calculated as the average of RS values determined from fivedifferent locations on each sample surface. Differences betweenfriction coefficients and surface roughness values obtained beforeand after testing of each biomaterial were analyzed for signifi-cance using a paired t-test with p< 0.05.

3 Results

3.1 Macroscale Friction Coefficient. For all biomaterial/cartilage interfaces, the friction coefficient increased nonlinearly,demonstrating a steeper rise at the beginning of sliding andincreasing at a lower rate toward the end of testing (Fig. 1(a)).Friction coefficient data showed that UHMWPE produced the low-est friction coefficient both at the start (1 min) and the end (60min) of testing, while Al2O3 resulted in the highest friction coeffi-cients at both time points (Fig. 1(b)). Both the 1-min friction coef-ficient (0.035 6 0.001) and the 60-min friction coefficient(0.391 6 0.012) of UHMWPE were significantly lower than thoseof all other biomaterials (p< 0.0001). The 1-min friction coeffi-cient of Al2O3 (0.134 6 0.003) was significantly higher than thoseof all other biomaterials (p< 0.0001), whereas the 60-min frictioncoefficient of Al2O3 (0.569 6 0.009) was similar to that of CoCrand significantly higher than those of all other biomaterials(p< 0.0001). CoCr produced the largest increase in friction coeffi-cient after 60 min of sliding, from 0.074 6 0.003 (1 min) to0.566 6 0.013 (60 min). Fitting the data shown in Fig. 1(a) with abiphasic exponential function of time indicated that the trendsshown in Figs. 1(a) and 1(b) hold for any value of time, with R2

values ranging from 0.9865 to 0.9944. From the fitted curves, theestimated steady-state friction coefficients of Al2O3, CoCr, SS, andUHMWPE were found equal to 0.61, 0.57, 0.51, and 0.46, respec-tively. These friction coefficients are close to the 60-min frictioncoefficient data shown in Fig. 1(b).

3.2 Protein Wear. The total amount of protein removed fromthe cartilage surface during sliding was not significantly differentbetween any of the biomaterials (Fig. 2). Total protein removedfrom the surface ranged from 0.589 6 0.167 lg/N�m for Al2O3 to0.685 6 0.257 lg/N�m for CoCr. The measured amount of SZPremoved from the surface was multiplied by a factor of 100 toenhance the comparison with the total protein removed. SZP in thewear debris ranged from 0.0004 6 0.0002 lg/N�m for Al2O3 to0.0010 6 0.0007 lg/N�m for UHMWPE.

3.3 Histology. Although SZP was removed from the surfaceduring sliding, IHC analysis revealed that SZP was still present at thecartilage surface (Fig. 3). Explants worn against CoCr showed visiblylower SZP content compared to explants worn against Al2O3, SS,and UHMWPE. Since the SZP distribution patterns were similar forboth sets of samples assayed, only one set is presented here. Theseresults correspond to the macroscale friction coefficient patterns ofthe biomaterials worn against cartilage, where CoCr yielded the larg-est increase in friction coefficient (Fig. 1).

Fig. 1 (a) Friction coefficient of different biomaterials slidingagainst articular cartilage versus time of sliding showed a non-linear increase in friction coefficient for all biomaterials. Eachcurve is an average of seven experiments. (b) Comparison offriction coefficients of different biomaterial/cartilage interfacesobtained at the start (1 min) and the end (60 min) of sliding. Thehighest and lowest friction coefficients were obtained withAl2O3 and UHMWPE, respectively, while CoCr resulted in thelargest increase in friction coefficient during testing. All groupswere significantly different (p < 0.0083) from each other exceptfor Al2O3 and CoCr at the 60-min time point.

Fig. 2 Normalized wear of total protein and SZP from the articularcartilage surface due to sliding against different biomaterials. Dif-ferences among the four biomaterial/cartilage interfaces werestatistically insignificant. The amount of worn SZP was multipliedby a factor of 100 to enhance the comparison with the total proteinloss.

Journal of Tribology OCTOBER 2011, Vol. 133 / 041201-3

3.4 Nanoscale Friction Coefficient and Surface Roughness.Nanoscale friction measurements obtained with an AFM revealeda significant decrease in friction coefficient of Al2O3 (p¼ 0.003)and CoCr (p¼ 0.004) after sliding against cartilage (Fig. 4(a)).Although lower friction coefficients were also obtained forUHMWPE and SS, these changes were relatively less significant.Surface roughness of the biomaterials was not significantlyaffected by the sliding process (Fig. 4(b)).

4 Discussion

The objective of this investigation was to evaluate the tribologi-cal properties of biomaterials commonly used in hemiarthroplastyand determine the biomaterial/cartilage interface that demon-strates the lowest friction coefficient and least tissue wear. Amongthe tested biomaterials, CoCr exhibited the least desirable tribo-logical properties, producing both the largest increase in frictioncoefficient through the duration of sliding and the highest frictioncoefficient at the 60-min time point (Fig. 1). IHC analysis of SZP,a marker of the superficial zone and a well-known boundary lubri-cant of articular cartilage [18,19,34], showed that cartilage slidagainst CoCr contained the least amount of SZP near the surface,both in density and distribution through the tissue (Fig. 3). How-ever, the amount of total protein and SZP in the wear debris of theCoCr/cartilage interface was not statistically different from any ofthe other biomaterial/cartilage interfaces (Fig. 2).

Friction coefficient and surface roughness results are consistentwith those of previous studies. CoCr has been reported to produce ahigher friction coefficient (�0.021–0.023) and less staining for col-lagen and proteoglycan in bovine osteochondral explants comparedto ceramics (�0.015) [11]. Histology from in vivo experiments ofCoCr femoral heads in canines revealed more degradation, fibrilla-tion, eburnation, and bony changes in the acetabulum compared topyrolytic carbon [5], suggesting a progression of osteoarthritis fol-lowing implantation. In another similar study, Vitallium caused pre-mature and increased loss of proteoglycans, ingrowth of pannusinto the cartilage, and disruption of the lamina splendens [6]. Theseresults are consistent with an inferior biomaterial causing excessivewear of articular cartilage. Even custom-fit Vitallium patello-femoral groove implants were found to cause a loss of proteogly-cans and fibrillation [35]. Noncontact profilometry yielded asurface roughness of cast CoCr implants of 42 nm, which is approx-imately 4–5 times less than the roughness of CoCr measured in thepresent study with the AFM. This discrepancy may be attributed todifferences in alloy composition, manufacturing process, and mea-surement method. The roughness of CoCr dental implants measuredwith an AFM was found equal to �134 nm [36], which is similar tothe roughness measured in this study (Fig. 4(b)). The surface rough-ness of CoCr femoral condylar implants measured with a stylusprofilometer was reported to be between �40 and 120 nm, depend-ing on the manufacturer [37]. Clearly, differences in measurementtechniques and material sources contributed to the wide ranges offriction coefficient and surface roughness of CoCr alloys examined

Fig. 3 Immunohistochemistry showed a lower SZP content at the articular cartilage surfacesliding against CoCr and a higher surface content for that sliding against Al2O3, SS, andUHMWPE. The lower staining density of SZP at the articular surface slid against CoCr correlateswith the higher friction coefficient at the CoCr/cartilage interface. The presence of SZP at thecartilage surface after sliding against different biomaterials for 60 min also corresponds to thelow levels of SZP detected in wear debris, suggesting that other proteins were also removedfrom the surface during sliding.

Fig. 4 Friction coefficient and surface roughness of differentbiomaterials obtained before and after sliding against cartilagewith an AFM. (a) The mean friction coefficients of Al2O3 andCoCr decreased significantly after sliding against cartilage (asignificance level of p < 0.05 is indicated by an asterisk). Themean friction coefficients of UHMWPE and SS also decreasedafter testing, but the changes were not significant. (b) Thesurface roughness of all biomaterial surfaces did not changesignificantly as a result of sliding against cartilage.

041201-4 / Vol. 133, OCTOBER 2011 Transactions of the ASME

in previous studies. A key advantage in the present investigationwas that using identical assays in all of the tests enabled direct com-parisons between biomaterial behaviors.

UHMWPE exhibited the best performance with respect to fric-tion coefficient, consistent with previous findings. Significantlylower friction coefficient was found for cartilage slid against poly-urethane than glass, which was attributed to the higher complianceof the polymer [4]. In the same study, fibrillation of the articularsurface and delamination of the collagen fibers occurred only forcartilage samples articulated against glass, while cartilage samplestested against polyurethane appeared normal even after 500 slid-ing cycles (60 m). It appears that the lower polymer stiffness maybe an important factor in reducing friction and wear of articularcartilage. Additionally, crosslinking of the UHMWPE used in thepresent study likely contributed to lower friction and wear, com-pared to uncrosslinked UHMWPE. Crosslinking of the polymerchains should produce lower and more stable friction coefficients(less plowing) and greater wear (shear) resistance due to preven-tion of molecular chain alignment during sliding that weakens thepolymer surface [38]. Profilometry measurements showed that theroughness of UHMWPE tibial inserts ranged from 2.5 to 1870nm, depending on the manufacturer [37]. Surface roughness val-ues of 69.8 and 2481.8 nm have been reported for molded andmachined polyethylene tibial inserts, respectively [39]. Theseroughness results are also consistent with the roughness measure-ments of the present study.

The nanoscale friction coefficient of UHMWPE demonstratedthe smallest change following wear testing against cartilage, asopposed to CoCr, which exhibited the largest change in frictioncoefficient after testing (Fig. 4(a)). These results reflect frictioncoefficient trends similar to those observed at the macroscale,where UHMWPE produced the smallest increase while CoCrresulted in the largest increase in cartilage friction coefficient(Fig. 1). For all biomaterials, the nanoscale friction coefficientsmeasured after testing were found to be close to that of healthyarticular cartilage (�0.20) [40,41], especially for CoCr and SS.This suggests that some material transfer from the cartilage to thebiomaterial surface may have occurred during sliding. Although theroughness did not change significantly for any biomaterial surface,the general trend was for the roughness of CoCr, UHMWPE, andSS to decrease as a result of sliding (Fig. 4(b)). This decrease mayhave contributed to the decrease in nanoscale friction coefficient af-ter testing (Fig. 4(a)). Changes in the surface composition may alsohave occurred during sliding contact. Material transfer from the car-tilage to the biomaterial surface would tend to lower the frictioncoefficient to levels comparable to that of the cartilage/cartilageinterface and smooth out the biomaterial surface as wear particlesaggregate and become embedded in the transferred material.

Quantification of total protein and SZP loss from the cartilagesurface revealed insignificant differences between biomaterials.SZP comprised only 0.1–0.2% of the total protein wear and IHCresults showed that worn cartilage still contained SZP at or near thesurface (Fig. 3). These data suggest that SZP was not entirely wornoff from the surface and that other surface molecules contributed tothe friction force generated at each biomaterial/cartilage interface.CoCr resulted in less SZP staining in both histology samples, whileAl2O3, SS, and UHMWPE showed similar SZP distributions. Theseresults are consistent with the macroscale friction coefficients ofarticular cartilage slid against each biomaterial, where the CoCr/cartilage interface produced the largest increase in friction coeffi-cient after sliding for 60 min (Fig. 1). One limitation of this study isthat under the relatively low contact pressure (0.1 MPa) applied inthe sliding tests, wear of SZP was not as pronounced as for otherproteins present at the articular surface. Since the 0.1 MPa contactpressure is at the low end of the physiological range [42,43], addi-tional studies of the wear dependence on normal load should pro-vide more insight into the wear performance of implants acrossvarious locations in the joint. Additionally, study of the wear de-pendence on sliding time (distance) would yield wear rate measure-ments and provide a means of identifying the best biomaterial for

in vivo use. Based on the macroscale friction coefficient data, it isexpected that CoCr would produce the highest wear rate due to thesteeper initial rise in friction coefficient. Conversely, becauseUHMWPE demonstrated the slowest friction coefficient rise, it isexpected to also result in less wear. However, these effects are notrepresented in the wear data because wear was only quantified atthe end of sliding. Wear measured by the decrease in thickness ofcartilage explants slid against CoCr has been reported to be higherthan that of cartilage slid against zirconia [11]. These results arealso consistent with the wear measurements of this study, whichshow a lower mean loss of both total protein and SZP for Al2O3

compared to CoCr (Fig. 2). The additional assays of nanoscale fric-tion coefficient and surface roughness of the biomaterials measuredbefore and after wear testing also suggest that more tissue wearaccumulated in the case of the CoCr/cartilage interface.

Under the experimental conditions of this study, sliding occurredin the boundary lubrication regime, which is relevant to the specificaim of determining the biomaterial/cartilage system with the besttribological properties. While other modes of lubrication, includingweeping [44] and elastohydrodynamic [45] lubrication, provideprotection by separating the cartilage surfaces by a fluid film,boundary lubrication protects cartilage during solid-solid contact,which is associated with excessive wear. Boundary lubrication wasconfirmed by the very small fluid film thickness (�0.14 nm), esti-mated from elastohydrodynamic fluid film theory [26,27], and phe-nomenological observations of the tested surfaces revealing directsurface interaction (wear) and friction coefficients typical of slidingin the boundary lubrication regime. The cartilage specimen wasallowed to equilibrate under the applied normal load for 2 min priorto the initiation of sliding. Others have found that for contact pres-sures in the range of 0.5–4 MPa, the fluid film thickness after 5 sfrom the application of the static load was in the range of 0.01–0.04lm, and decreased to values in the range of 0.002–0.02 lm after120 s under the applied static load [46]. Similarly the fluid filmthickness due to entraining fluid was estimated to be equal to 0.01lm [46]. Since the surface roughness of articular cartilage is in therange of 0.3–0.8 lm [24,41,47,48] and the average roughness of thetested biomaterials was found to be about 0.1–0.2 lm (i.e., both atleast an order of magnitude greater than the film thickness), bound-ary lubrication conditions were predicted based on theoretical cal-culations [29].

The presented results revealed that CoCr, the biomaterial mostfrequently used against cartilage in hip, knee, and shoulder replace-ments, exhibited the least desirable friction and wear propertiesamong the four biomaterials examined. Al2O3 resulted in betterfriction characteristics and less cartilage wear; however materialtransfer was discovered in the AFM studies of the disk surfaces fol-lowing wear testing. Ceramics are also commonly used in hemiar-throplasty of the hip and smaller joints (hand and foot). UHMWPEdemonstrated a better overall performance with respect to cartilagefriction and wear behavior. However, the current use of UHMWPEin joint replacements is in the form of hip or meniscal inserts orpatellofemoral replacement, where contact with cartilage is mini-mal. Results of the present study suggest that expanding the use ofUHMWPE and other polymers in orthopaedic applications wherecartilage is replaced by an engineered biomaterial is worth explor-ing. While UHMWPE yielded the lowest friction coefficientthroughout the duration of sliding against cartilage, it produced fric-tion coefficients between two to ten times higher than that of a car-tilage/cartilage interface. Constant and low friction coefficients of�0.01 [49] and �0.05 [50] have been reported for cartilage slidingagainst cartilage under a similar contact pressure (0.5 MPa) for 120and 500 min, respectively. An increase in friction coefficient fromapproximately 0.01 to 0.02 was observed after sliding cartilageagainst cartilage for 7 h [51]. The friction coefficients of any bio-material/cartilage interface tested to date have yet to achieve suchlow values. Research should be directed toward the development ofceramics and polymers that mimic the tribological properties ofnormal articular cartilage, while maintaining the mechanicalstrength necessary to sustain joint loading.

Journal of Tribology OCTOBER 2011, Vol. 133 / 041201-5

5 Conclusions

An experimental investigation of the tribological properties ofcommon hemiarthroplasty biomaterials articulated against cartilageexplants was conducted using identical assays for each biomaterial/cartilage interface. Friction and wear experiments, nanoscale fric-tion and roughness measurements, protein wear assaying, and his-tology analysis were used to compare the tribological performanceof each biomaterial/cartilage interface under physiological condi-tions. The obtained results suggest that UHMWPE, the most com-pliant and softest material tested, should produce the least wear inhemiarthroplasty applications. SS, which has a lower stiffness andhardness than Al2O3 and a lower stiffness than CoCr, also shouldresult in good tribological properties. The worst tribological per-formance, highest wear of surface protein, and least amount of SZPstaining in cartilage corresponded to the CoCr/cartilage interface,while Al2O3 performed moderately better with respect to wear butnot with respect to friction. In the context of the obtained results, itappears that UHMWPE and SS have a greater potential as hemiar-throplasty biomaterials because they produce significantly better tri-bological properties, while preserving the healthy articular cartilagemuch more effectively. Therefore, further improvements to the sur-face properties of hemiarthroplasty biomaterials must be made toreduce friction and wear of cartilage to near those of the cartilage/cartilage interface.

Acknowledgment

The authors gratefully acknowledge the use of the atomic forcemicroscope at the NEAT-ORU, Spectral Imaging Facility of theUniversity of California, Davis, and Dr. T. Schmid for providingthe S6.79 monoclonal antibody.

References[1] Edwards, T. B., Kadakia, N. R., Boulahia, A., Kempf, J.-F., Boileau, P., Nemoz,

C., and Walch, G., 2003, “A Comparison of Hemiarthroplasty and TotalShoulder Arthroplasty in the Treatment of Primary Glenohumeral Osteoarthri-tis: Results of a Multicenter Study,” J. Shoulder Elbow Surg., 12, pp. 207–213.

[2] Sah, A. P., and Estok II, D. M., 2008, “Dislocation Rate After Conversion FromHip Hemiarthroplasty to Total Hip Arthroplasty,” J. Bone Joint Surg. Am., 90,pp. 506–516.

[3] Springer, B. D., Scott, R. D., Sah, A. P., and Carrington, R., 2006, “McKeeverHemiarthroplasty of the Knee in Patients Less Than Sixty Years Old,” J. BoneJoint Surg. Am., 88, pp. 366–371.

[4] Foy, J. R., Williams III, P. F., Powell, G. L., Ishihara, K., Nakabayashi, N., andLaBerge, M., 1999, “Effect of Phospholipidic Boundary Lubrication in Rigidand Compliant Hemiarthroplasty Models,” Proc. Inst. Mech. Eng., Part H: J.Eng. Med., 213, pp. 5–18.

[5] Cook, S. D., Thomas, K. A., and Kester, M. A., 1989, “Wear Characteristics ofthe Canine Acetabulum Against Different Femoral Prostheses,” J. Bone JointSurg. Br., 71-B, pp. 189–197.

[6] Cruess, R. L., Kwok, D. C., Duc, P. N., Lecavalier, M. A., and Dang, G.-T.,1984, “The Response of Articular Cartilage to Weight-Bearing Against Metal.A Study of Hemiarthroplasty of the Hip in the Dog,” J. Bone Joint Surg. Br.,66-B, pp. 592–597.

[7] Jung, M., Wieloch, P., Lorenz, H., Gotterbarm, T., Veyel, K., Daniels, M., Mar-tini, A. K., and Daecke, W., 2008, “Comparison of Cobalt Chromium, Ceramicand Pyrocarbon Hemiprostheses in a Rabbit Model: Ceramic Leads to MoreCartilage Damage Than Cobalt Chromium,” J. Biomed. Mater. Res., Part B:Appl. Biomater., 85B, pp. 427–434.

[8] Helmick, C. G., Felson, D. T., Lawrence, R. C., Gabriel, S., Hirsch, R., Kwoh,C. K., Liang, M. H., Kremers, H. M., Mayes, M. D., Merkel, P. A., Pillemer,S. R., Reveille, J. D., and Stone, J. H., 2008, “Estimates of the Prevalence ofArthritis and Other Rheumatic Conditions in the United States. Part I,” ArthritisRheum., 58, pp. 15–25.

[9] Office of the Surgeon General, 2005, “The 2004 Surgeon General’s Report onBone Health and Osteoporosis: What It Means To You,” U.S. Department ofHealth and Human Services, Washington, DC.

[10] Muller, L. P., Degreif, J., Rudig, L., Mehler, D., Hely, H., and Rommens, P. M.,2004, “Friction of Ceramic and Metal Hip Hemi-Endoprostheses AgainstCadaveric Acetabula,” Arch. Orthop. Trauma Surg., 124, pp. 681–687.

[11] Patel, A. M., and Spector, M., 1997, “Tribological Evaluation of Oxidized Zir-conium Using an Articular Cartilage Counterface: A Novel Material for Poten-tial Use in Hemiarthroplasty,” Biomaterials, 18, pp. 441–447.

[12] McCutchen, C. W., 1962, “The Frictional Properties of Animal Joints,” Wear,5, pp. 1–17.

[13] Blomfeldt, R., Tornkvist, H., Eriksson, K., Soderqvist, A., Ponzer, S., andTidermark, J., 2007, “A Randomised Controlled Trial Comparing BipolarHemiarthroplasty With Total Hip Replacement for Displaced Intracapsular

Fractures of the Femoral Neck in Elderly Patients,” J. Bone Joint Surg. Br.,89-B, pp. 160–165.

[14] Ravikumar, K. J., and Marsh, G., 2000, “Internal Fixation versus Hemiarthro-plasty Versus Total Hip Arthroplasty for Displaced Subcapital Fractures ofFemur�13 Year Results of a Prospective Randomised Study,” Injury, 31, pp.793–797.

[15] Waters, T. S., and Bentley, G., 2003, “Patellar Resurfacing in Total KneeArthroplasty. A Prospective, Randomized Study,” J. Bone Joint Surg. Am.,85A, pp. 212–217.

[16] Jay, G. D., Harris, D. A., and Cha, C.-J., 2001, “Boundary Lubrication byLubricin is Mediated by O-linked b(1-3)Gal-GalNAc Oligosaccharides,” Gly-coconj. J., 18, pp. 807–815.

[17] Schumacher, B. L., Block, J. A., Schmid, T. M., Aydelotte, M. B., and Kuett-ner, K. E., 1994, “A Novel Proteoglycan Synthesized and Secreted by Chondro-cytes of the Superficial Zone of Articular Cartilage,” Arch. Biochem. Biophys.,311, pp. 144–152.

[18] Swann, D. A., Slayter, H. S., and Silver, F. H., 1981, “The Molecular Structureof Lubricating Glycoprotein-I, The Boundary Lubricant for ArticularCartilage,” J. Biol. Chem., 256, pp. 5921–5925.

[19] Neu, C. P., Khalafi, A., Komvopoulos, K., Schmid, T. M., and Reddi, A. H.,2007, “Mechanotransduction of Bovine Articular Cartilage Superficial ZoneProtein by Transforming Growth Factor b Signaling,” Arthritis Rheum., 56, pp.3706–3714.

[20] Brown, T. D., and Shaw, D. T., 1984, “In Vitro Contact Stress Distribution onthe Femoral Condyles,” J. Orthop. Res., 2, pp. 190–199.

[21] Bruns, J., Volkmer, M., and Luessenhop, S., 1994, “Pressure Distribution in theKnee Joint. Influence of Flexion With and Without Ligament Dissection,”Arch. Orthop. Trauma Surg., 113, pp. 204–209.

[22] Nugent-Derfus, G. E., Takara, T., O’Neill, J. K., Cahill, S. B., Gortz, S.,Pong, T., Inoue, H., Aneloski, N. M., Wang, W. W., Vega, K. I., Klein, T.J., Hsieh-Bonassera, N. D., Bae, W. C., Burke, J. D., Bugbee, W. D., andSah, R. L., 2007, “Continuous Passive Motion Applied to Whole JointsStimulates Chondrocyte Biosynthesis of PRG4,” Osteoarthritis Cartilage,15, pp. 566–574.

[23] McCutchen, C. W., 1982, “Cartilage is Poroelastic, not Viscoelastic(Including an Exact Theorem About Strain Energy and Viscous Loss, andan Order of Magnitude Relation for Equilibration Time),” J. Biomech., 15,pp. 325-327.

[24] DuRaine, G., Neu, C. P., Chan, S. M. T., Komvopoulos, K., June, R. K., andReddi, A. H., 2009, “Regulation of the Friction Coefficient of Articular Carti-lage by TGF-b1 and IL-1b,” J. Orthop. Res., 27, pp. 249–256.

[25] Bhushan, B., 1999, Principles and Applications of Tribology, Wiley, NewYork.

[26] Dowson, D., and Jin, Z. M., 1986, “Micro-Elastohydrodynamic Lubrication ofSynovial Joints,” Eng. Med., 15, pp. 63–65.

[27] Dowson, D., and Higginson, G. R., 1977, Elastohydrodynamic Lubrication,Pergamon Press, New York.

[28] Su, J. L,. Schumacher, B. L., Lindley, K. M., Soloveychik, V., Burkhart,W., Triantafillou, J. A., Kuettner, K., and Schmid, T., 2001, “Detection of Su-perficial Zone Protein in Human and Animal Body Fluids by Cross-SpeciesMonoclonal Antibodies Specific to Superficial Zone Protein,” Hybridoma, 20,pp. 149–157.

[29] Bhushan, B., 2001, Modern Tribology Handbook, CRC Press, Boca Raton, FL.[30] Bayer, R. G., 2004, Mechanical Wear Fundamentals and Testing, CRC Press,

Boca Raton, FL.[31] Turenne, S., Lavallee, F., and Masounave, J., 1989, “Matrix Microstructure

Effect on the Abrasion Wear Resistance of High-Chromium White Cast Iron,”J. Mat. Sci., 24, pp. 3021–3028.

[32] Hutter, J. L., and Bechhoefer, J., 1993, “Calibration of Atomic-Force Micro-scope Tips,” Rev. Sci. Instrum., 64, pp. 1868–1873.

[33] Asay, D. B., and Kim, S. H., 2006, “Direct Force Balance Method for AtomicForce Microscopy Lateral Force Calibration,” Rev. Sci. Instrum., 77, pp.043903 (1–9).

[34] Zappone, B., Greene, G. W., Oroudjev, E., Jay, G. D., and Israelachvili,J. N., 2008, “Molecular Aspects of Boundary Lubrication by Human Lubri-cin: Effect of Disulfide Bonds and Enzymatic Digestion,” Langmuir, 24, pp.1495–1508.

[35] LaBerge, M., Bobyn, J. D., Drouin, G., and Rivard, C. H., 1992, “Evaluation ofMetallic Personalized Hemiarthroplasty: A Canine Patellofemoral Model,”J. Biomed. Mater. Res., 26, pp. 239–254.

[36] Jang, K.-S., Youn, S.-J., and Kim, Y.-S., 2001, “Comparison of Castability andSurface Roughness of Commercially Pure Titanium and Cobalt-ChromiumDenture Frameworks,” J. Prosthet. Dent., 86, pp. 93–98.

[37] Chapman-Sheath, P., Cain, S., Bruce, W. J., Chung, W. K., and Walsh, W. R.,

2002, “Surface Roughness of the Proximal and Distal Bearing Surface of Mo-

bile Bearing Total Knee Prostheses,” J. Arthroplasty, 17, pp. 713–717.[38] Zhou, J., Chakravartula, A., Pruitt, L., and Komvopoulos, K., 2004,

“Tribological and Nanomechanical Properties of Unmodified and CrosslinkedUltra-High Molecular Weight Polyethylene for Total Joint Replacements,”ASME J. Tribol., 126, pp. 386–394.

[39] Benson, L. C., DesJardins, J. D., and LaBerge, M., 2001, “Effects of In VitroWear of Machined and Molded UHMWPE Tibial Inserts on TKR Kinematics,”

J. Biomed. Mater. Res., 58, pp. 496–504.[40] Chan, S., Neu, C., DuRaine, G., Komvopoulos, K., and Reddi, A. H., 2008,

“Articular Cartilage Friction Following Enzymatic Digestion Measured byAtomic Force Microscopy,” 54th Ann. Meet., Trans. Orthop. Res. Soc., 33,Poster No. 641, San Francisco, CA.

041201-6 / Vol. 133, OCTOBER 2011 Transactions of the ASME

[41] Coles, J. M., Blum, J. J., Jay, G. D., Darling, E. M., Guilak, F., and Zauscher,S., 2008, “In Situ Friction Measurement on Murine Cartilage by Atomic ForceMicroscopy,” J. Biomech., 41, pp. 541–548.

[42] Brown, T. D., and Shaw, D. T., 1983, “In Vitro Contact Stress Distributions inthe Natural Human Hip,” J. Biomech., 16, pp. 373–384.

[43] Lee, S. J., Aadalen, K. J., Malaviya, P., Lorenz, E. P., Hayden, J. K., Farr,J., Kang, R. W., and Cole, B. J., 2006, “Tibiofemoral Contact Mechanics AfterSerial Medial Meniscectomies in the Human Cadaveric Knee,” Am. J. SportsMed., 34, pp. 1334–1344.

[44] McCutchen, C. W., 1959, “Mechanism of Animal Joints: Sponge-Hydrostaticand Weeping Bearings,” Nature, 184, pp. 1284–1285.

[45] Wright, V., and Dowson, D., 1976, “Lubrication and Cartilage,” J. Anat., 121,pp. 107–118.

[46] Forster, H., and Fisher, J., 1996, “The Influence of Loading Time and Lubricanton the Friction of Articular Cartilage,” Proc. Inst. Mech. Eng., Part H: J. Eng.Med., 210, pp. 109–119.

[47] Park, S., Costa, K. D., and Ateshian, G. A., 2004, “Microscale FrictionalResponse of Bovine Articular Cartilage from Atomic Force Microscopy,”J. Biomech., 37, pp. 1679–1687.

[48] Forster, H., and Fisher, J., 1999, “The Influence of Continuous Sliding and Sub-sequent Surface Wear on the Friction of Articular Cartilage,” Proc. Inst. Mech.Eng., Part H: J. Eng. Med., 213, pp. 329–345.

[49] Northwood, E., Fisher, J., and Kowalski, R., 2007, “Investigation of the Fric-tion and Surface Degradation of Innovative Chondroplasty Materials AgainstArticular Cartilage,” Proc. Inst. Mech. Eng., Part H: J. Eng. Med., 221,pp. 263–279.

[50] Northwood, E., and Fisher, J., 2007, “A Multi-Directional In Vitro InvestigationInto Friction, Damage and Wear of Innovative Chondroplasty Materials AgainstArticular Cartilage,” Clin. Biomech., 22, pp. 834–842.

[51] Katta, J., Jin, Z., Ingham, E., and Fisher, J., 2009, “Effect of Nominal Stress onthe Long Term Friction, Deformation and Wear of Native and Glycosaminogly-can Deficient Articular Cartilage,” Osteoarthritis Cartilage, 17, pp. 662–668.

Journal of Tribology OCTOBER 2011, Vol. 133 / 041201-7