Fretting-corrosion in Hip Taper Modular Junctions: The Influence of Topography and pH levels – an in-vitro study

Abstract

Contemporary hip implants feature a modular design. Increased reported failure rates associated with the utilization of modular junctions have raised many clinical concerns. Typically, these modular interfaces contain circumferential machining marks (threads or microgrooves), but the effect of the machining marks on the fretting-corrosion behavior of total hip implant materials is unknown. This study reports the effects of microgrooves on the fretting-corrosion behavior of hip implant materials. The flat portions of two cylindrical, polished, CrCrMo alloy pins were loaded horizontally against one rectangular Ti alloy rod. Two surface preparation groups were used for the Ti6Al4V rod (polished and machined). The polished group was prepared using the same methods as the CoCrMo pins. The machined samples were prepared by creating parallel lines on the rod surfaces to represent microgrooves present on the stem tapers of head-neck modular junctions. Newborn calf serum (30 g/L protein content; 37°C) at pH of levels of 7.6 and 3.0 were used to simulate the normal joint fluid and a lowered pH within a crevice, respectively. The samples were tested in a fretting corrosion apparatus under a 200N normal force and a 1Hz sinusoidal fretting motion with a displacement amplitude of 25μm. All electrochemical measurements were performed with a potentiostat in a three-electrode configuration. The results show significant differences between machined samples and polished samples in both electrochemical and mechanical responses. In all cases, the magnitude of the drop in potential was greater in the machined group compared to the polished group. The machined group showed a lower total dissipated friction energy for the entire test compared to the polished group. Additionally, the potentiostatic test measurements revealed a higher evolved charge in the machined group compared to the polished group at both pH conditions (pH 7.6 and 3.0). The machined surfaces lowered the overall dissipated friction energy at pH 7.6 compared to pH 3.0, but also compromised electrochemical performance in the tested conditions. Therefore, the role of synergistic interaction of wear and corrosion with surface topographical changes is evident from the outcome of the study. Despite the shift towards higher electrochemical destabilization in the machined group, both polished and machined groups still exhibited a mechanically dominated degradation.

1.0. INTRODUCTION

Total hip replacement (THR) is one of the most common surgical procedures performed on more than 330,000 patients in the US annually^1–3. Recent investigations demonstrated that synergistic wear and corrosion processes (tribocorrosion) occurring at modular junctions of these implants are one of the factors limiting implant survivorship^4–6. Metal ions and debris released in the body due to tribocorrosion processes may result in local and systemic consequences^7–12. The release of tribocorrosion debris can lead to implant failure due to pain, inflammation, swelling, pseudo-tumors, cobaltism, and potentially genotoxicity^13,14. In addition, it was reported that the metal debris, in turn, can induce macrophage polarization and cell accelerated corrosion at the taper interface^15,16. Hence, it is imperative that the release of metal ions and tribocorrosion debris is held at a minimum.

Contemporary THRs feature a modular design, whereby the individual components of the implant are assembled intraoperatively¹⁷. Modular implants allow for a reduction of inventory and flexibility during surgery to accommodate patient anatomy^17–19. Typically, the surfaces of the femoral neck, or the stem taper, within the head-neck modular interface contain circumferential machining marks, which are designed to lock to the smoother femoral head surface, or head taper, and endure large compressive, tensile, and rotational loads²⁰. Despite the many benefits, the modular implant design has recently become a major concern in orthopedics as clinical reports have linked the use of these implants with adverse local tissue reactions (ALTRs) and increased failure rates^6,21–24. In addition, body fluids from the surrounding tissue penetrate into the modular junction, facilitating mechanically assisted crevice and fretting-corrosion^25–27. Fretting-corrosion is produced by relatively micro-scale (≪ 100 μm) motions between components/surfaces that are induced by cyclic loading²⁸. The fretting process increases the corrosion rate because the protective oxide layer or passive film is continuously fractured and reformed (repassivation) on the metallic surfaces. Relative motion within the crevice also accelerates corrosion reactions in concert with local oxygen depletion²⁹. Mechanically assisted crevice corrosion fundamentally alters the electrolyte constituents and electrochemistry within the crevice, a process driven by the reduction in free energy associated with a metal-oxide formation that results in the sequestration of oxygen from any available source (primarily H₂O)²⁶. Repassivation of metallic surfaces also results in a buildup of H⁺ and C1⁻ and a lowering of the pH within the crevice³⁰. Some studies suggest that the presence of cells within the confined space of modular junctions can further alter the chemical environment^31,32. Collectively, the above-mentioned factors are expected to cause a transition to acidic pH levels and thus enable severe corrosion processes, including pitting and dissolution, in addition to fretting wear and corrosion. However, the actual pH occurring in vivo within THR modular junctions are currently unknown.

Our laboratory has previously reported that the geometry of the machining lines may have an influence on the ultimate damage seen on retrieved implants, and that machining lines/surface topography change the contact conditions present at the head-neck modular junction^33–36. In addition, fretting-corrosion processes at metal interfaces are influenced by the number and magnitude of loading cycles, pH level, and displacement amplitude. Hence parametric studies are necessary to determine the effect of microgrooves/surface topography on the mechanical and corrosion behavior and subsequent transitions in the failure mechanisms at hip implant modular junctions. Therefore, the objective of this study is to compare the fretting-corrosion behavior of surfaces representative of current hip implant systems with and without the presence of microgrooves and under normal (7.6) and acidic (3.0) pH conditions. For this study, we utilize a flat-on-flat fretting corrosion apparatus that allows us to control both the microgroove driven contact mechanics, electrochemical potential and solution chemistry.

2.0. METHODS

2.1. Study design

Two different factors were examined in this study (Fig. 1): surface topography and solution pH level. Surface topography consisted of either polished or surfaces machined with microgrooves. The solution pH level consisted of a low (pH 3.0) and neutral (pH 7.6) group. Three tests were conducted for each group (N=3) to ensure repeatability for a total of 12 tests.

1.

Schematic diagram of the experimental design. The diagram shows the material couple/interfaces, conditions, factors, measurements, and outputs.

2.2. Sample preparation

CoCrMo pins (representing the head taper) and Ti6Al4V rods (representing the stem taper) were used in this study. Table 1 lists the alloy chemical composition. The CoCrMo pins were polished with silicon carbide grinding paper and diamond suspensions to attain an average roughness of less than 25nm. The Ti6Al4V rods were prepared with two different surface topographies, (i) polished (ii) machined. The “polished” Ti rod samples were prepared using the same methods as the CoCrMo pins. The polished group served as control samples and represented the surface topography typically used in fretting-corrosion tests. The “machined” samples were produced using a diamond blade fly cutter to create microgrooves. The height of and spacing between microgrooves were established to be similar to microgrooves measured on retrieved stem tapers^33,34 (median microgroove height of 11μm and spacing between microgrooves of 200μm). Before testing, all pins and rods were cleaned using 10-minute agitation cycles in an ultrasonic bath in 70% isopropanol solution, followed by a 10-minute agitation in deionized water and, finally, air drying using a nitrogen gas stream.

Table 1:

Chemical composition (in weight percent) of Ti6Al4V and CoCrMo alloys.

	Composition (in wt%)
Ti alloy	Ti	Al	V	C	Fe	O2	N2	H2
	89.62	6.1	4.0	0.004	0.16	0.106	0.008	0.0022
CoCrMo alloy	Co	Cr	Mo	C	Si	Mn	Al
	64.60	27.63	6.04	0.241	0.66	0.70	0.02

2.3. Solution

Newborn calf serum (Gibco, CA, USA) was used to simulate the proteinaceous joint fluid environment. The original newborn calf serum, which was reported by the manufacturer to contain a protein content of 67g/L, was diluted to a final protein content of 30 g/L using a buffer solution composed of: deionized water, sodium chloride (9 g/L), tris (hydroxymethyl) aminomethane (THAM, 27 g/L), ethylenediamine tetraacetic acid (EDTA, 200 mg/L), and enough hydrochloric acid to bring the pH of the solution down to the neutral level of pH 7.6. Once the initial pH of 7.6 was reached, lactic acid was added to attain a pH of 3.0 for the acidic group.

2.4. Fretting-corrosion testing apparatus

A previously developed fretting-corrosion testing apparatus was used to measure mechanical and electrochemical responses due to fretting and corrosion³⁷. The novelty of this set-up is that it uses a flat-on-flat contact geometry. This contact geometry contrasts with the typically used ball-on-flat or cylinder-on-flat contact geometries and is believed to be a better representation of the in vivo contact conditions of hip implant modular junction. Details of the construction, set-up configurations, and compliance evaluation of this apparatus have been previously published and will only be briefly described^37–39. In order to simulate the head-neck modular junction, the flat portions of two CoCrMo pins were axially loaded under a 200N force against the flat portion of a single, rectangular, Ti6Al4V rod. The Ti6Al4V rod underwent displacement in a direction perpendicular to the axial loading of the pins and parallel microgrooves (if present) (Fig. 2A). It should be noted that the microgrooves were made only on the Ti rod, oriented perpendicular to the axis of the rod and pins. Displacement consisted of a sinusoidal fretting motion with a displacement amplitude of 25μm applied to the rod at a frequency of 1Hz for 3600 cycles (3600 s). The samples were mounted into an electrochemical cell and the temperature of the test solution was maintained at 37°C. Care was taken to ensure that the rod was aligned properly with respect to the pins. After placing the pins in contact with the rod, acrylic cement was injected through peripheral holes in the pin holder to fix the pins in place. The cement was allowed to dry before filling the chamber with electrolyte/lubricant. The position of the Ti6Al4V rod and the resulting shear load were monitored and recorded (frequency of recording = 100 Hz at 10-second intervals). Fretting regime (i.e. partial slip vs. gross slip) was assessed by evaluating the shear load-displacement hysteresis loop behavior to obtain the ratio of dissipated energy to total energy as defined by Mohrbacher et al.⁴⁰ It was determined that all testing groups were within the gross slip regime.

2.

The details of the material couples. (A) Diagram of the contact conditions featuring two polished CoCrMo pins that are loaded axially against a Ti6AL4V rod which is either polished or machined with parallel microgrooves. (B) Example of the wear pattern observed after the completion of the experiment (load of 200N, displacement amplitude of ± 25um, applied as a sinusoid at a frequency of 1 Hz for 3600 cycles) for a pin against a machined rod.

All electrochemical measurements were performed with a potentiostat (Interface 1000, Gamry, Warminster, PA, USA) in a three-electrode configuration: 1) contacting metals as the working electrode, 2) a graphite rod as the counter electrode, and 3) a saturated calomel electrode (SCE) as the reference electrode. Fretting-corrosion tests were conducted under free-potential mode (no applied potential) and potentiostatic mode (applied potential).

2.5. Electrochemical tests

Free potential tests were performed to evaluate the effect of mechanical stimulation (fretting) on the electrochemical stability of the system. For these tests, changes in potential were monitored before, during, and after the applied fretting motion. To evaluate the electrochemical stability of the system we measured (i) ‘V_Drop’, the drop in potential due to the onset of fretting motion, and (ii) ‘${\overline{V}}_{\mathit{\text{Fretting}}}$ ‘, the average potential of the system during the fretting phase.

Potentiostatic tests at an applied potential of −250mV vs. SCE were used to determine the contributions of wear and corrosion to the overall degradation of the metal samples, and to evaluate the regime transitions, as described by Stack et. al.^41,42. The applied potential of −250mV corresponds to the anodic domain according to the previously conducted potentiodynamic polarization tests⁴³.

2.6. Surface area calculations

The total exposed surface area between the rod and pins was used to calculate the current density and charge density. The surface area calculations were estimated using different methods for the polished and machined groups. For the polished group, surfaces were assumed to be flat and the rectangular surface area of the immersed portions of the sample was used to calculate surface area. For the machined groups, the increased exposed surface area due to the microgrooves was estimated from images collected from the machined rod surfaces. The surface topography of machined surfaces was measured with a white light interferometer (New View 6300, Zygo Corporation, Middlefield, CT, USA). The obtained topographical data was used to assess the actual surface-exposed surface area during testing.

2.7. Weight loss estimation and wear-corrosion synergy

The synergistic interactions between wear and corrosion were estimated using Eq. 1, as proposed by Stack et al. for micro-abrasion-corrosion systems,^41,42

$$K_{WC}=K_C+K_W$$ Eq. 1

where K_WC is a total mass loss, K_C is mass loss due to corrosion, and K_W is mass loss due to wear. K_WC was measured by analyzing the total metal content of the solution using inductive coupled mass-spectroscopy (ICP-MS)⁴⁴ and adding together the total individual mass losses of Ti, Co, Cr, and Mo (Table 1).

K_C can be calculated using Faraday’s equation:

$$K_C=\frac{M\times Q}{n\times f}$$ Eq. 2

where M is the atomic mass of the metal alloy, Q is the charge passed through the working electrode in coulombs (equal to current times time), n is the number of valence electrons (we assumed n=2), and f is Faraday’s constant (96500C/mol⁻¹). It should be noted that this system contains two different alloys coupled together, and therefore some consideration has to be made as to how to account for the current levels generated from each individual component. From previous tests performed in our lab³⁷, mass loss due to corrosion of Ti6Al4V (under mechanical or tribological exposure) against an electrochemically inert counterbody is more than two orders of magnitude higher than CoCrMo⁴⁵. Therefore, we assumed that the K_C contribution of CoCrMo was negligible, and K_C was calculated using only the properties of Ti6Al4V. Once K_WC and K_C have been determined, the relationship in Eq. (1) can be rearranged to find K_W.

2.8. Surface characterization

Three-dimensional images of the polished and machined surfaces were obtained using a white-light interferometry microscope (Zygo New View 6300, Zygo Corporation, Middlefield, CT, USA). After free potential experiments, characterization of the worn surface was performed using scanning electron microscopy (SEM) (Joel JSM-6490 LV, Oxford Instruments, Oxford, UK) and energy-dispersive x-ray spectroscopy (EDS). The surfaces from the in vitro fretting-corrosion tests were also compared to images of modular junction retrieval samples in order to analyze the ability of the fretting-corrosion apparatus to reproduce the observed clinical wear features (i.e. provide face validity for the tests).

2.9. Contact pressure estimation

It is important to note the difference in contact conditions between the polished and machined rod surfaces. As previously described, most tribometers contain a ball-on-flat or cylinder-on-flat design due to the ability to accurately estimate the contact pressure based on the Hertzian contact model.⁴⁶ The unique contact conditions of a flat-on-flat-on-flat design did not allow for traditional Hertzian estimations. Therefore, the contact pressure had to be estimated after the fretting corrosion test by examination of the wear scar area. An example of the wear scar is shown in Fig. 2B, which shows the transfer of parallel marks onto the CoCrMo pins after the fretting corrosion test. The images of the tested surfaces were taken with white light microscopy (Zygo New View 6300, Zygo Corporation, Middlefield, CT, USA) and then imported into ImageJ software⁴⁷ and analyzed. The lengths of visible wear marks were measured and used to calculate the contact pressure based on a modified Hertzian contact model between a cylinder on a flat surface for each line contact. For this model, the pressure was believed to be equally distributed between the contact points/lines and along the contact path of each observed line. A study by Dunders et al. showed that contact between elastic bodies with wavy surfaces can be approximated by a cylinder on cylinder Hertz solution for small contact lengths such as those expected in this study⁴⁸. The following model was used to estimate the contact pressure. The normal force, F_n, acting on each cylinder/microgroove, was calculated as the applied force, F_a, divided by the number of contacts, N.

$$F_n=\frac{F_a}{N}$$ Eq. 3

The contact half-width for any given contact, b_i, can be expressed as:

$$b_i=\sqrt{\frac{2F_n}{\pi l_i}\frac{\frac{\left(1-v_1^2\right)}{E_1}+\frac{\left(1-v_2^2\right)}{E_2}}{\frac{1}{d_1}+\frac{1}{d_2}}}$$ Eq. 4

where l_i is the length of contact for a given cylinder, ν₁ and ν₂ are the Poisson’s ratio for the two contacting surfaces, E₁ and E₂ are the respective elastic moduli of the contacting surfaces, and d₁ and d₂ are the respective diameters of the contacting surfaces. The diameter of each cylinder/microgroove was 205 μm, as measured from the surface profile of the white light interferometry measurements. For flat surfaces, the corresponding d is infinitely large (d = ∞), and therefore, the expression drops out. Finally, contact pressure in each cylinder, i, was quantified as⁴⁶:

$${\mathit{\text{pressure}}}_i=\frac{2F_n}{\pi l_i{b}_i}$$ Eq. 5

For the polished group, the worn area was also estimated using ImageJ software; however, the relationship of the applied force over the apparent area (F_a/A) was used to quantify the pressure.

2.10. Statistical analysis

Each test was repeated 3 times. The data were statistically analyzed using student t-tests with p<0.05 significance level. The data was compared between samples tested at the same pH level (comparison between polished and machined groups) and between samples tested at different pH levels but having the same topography.

3. RESULTS

3.2. Surface characterization of original, untested surfaces

Fig. 3 shows the difference in the topography and surface profile between the polished and machined surfaces before electrochemical testing. The machined group had an average height of microgrooves of 11.32±0.5μm and spacing between microgrooves of 314±6.0μm. The polished and machined surfaces had an average surface roughness of < 0.025μm and 2.5μm, respectively. A free surface area of 15.4 cm² was estimated for the polished group surfaces. Machined surfaces had an estimated free surface area of 16.3 cm²

3.

Original, unworn topography and surface profile plots of the polished and machined Ti6AL4V rod surfaces.

3.3. Contact pressure estimate

The average worn area of pins contacting polished rod surfaces was 3.8 ± 0.07 mm². The worn area and applied force of 200N corresponded to a contact pressure of 52±10 MPa at each pin/rod interface. Based on the model described in Section 2.9, the average contact pressure for the machined rod surfaces was 1200±700MPa at each pin/rod interface. Therefore, given the same applied force, the machined surface topography increased contact pressure by a factor of 25.

3.4. Electrochemical measurements

The electrochemical stability of each group was evaluated from the evolution of potential during free potential tests (Fig. 4A). The graph shows that the fretting motion was initiated at 600 seconds and lasted for a period of 3600 seconds. Two parameters in the evolution of potential were evaluated, V_Drop and ${\overline{V}}_{\mathit{\text{Fretting}}}$. V_Drop, the initial cathodic drop in potential due to the onset of fretting, was lower in the polished group than the machined group, and lowest for pH 3.0 samples within each group (Fig. 4B). The polished pH 3.0 group had the lowest V_Drop (0.005±0.003 V vs. SCE), followed by the polished pH 7.6 group (0.026±0.010 V vs. SCE), the machined pH 3.0 group (0.078±0.022 V vs. SCE), and the machined pH 7.6 group (0.102±0.022 V vs. SCE). ${\overline{V}}_{\mathit{\text{Fretting}}}$, the average corrosion potential during the fretting phase, was higher in the polished group than the machined group, and highest for pH 3.0 samples within each group (Fig. 4C). The polished pH 3.0 group had the highest ${\overline{V}}_{\mathit{\text{Fretting}}}$ (−0.083±0.057 V vs. SCE), followed by the machined pH 3.0 group (−0.227±0.025 V vs. SCE), the polished pH 7.6 group (−0.299±0.004 V vs. SCE), and the machined pH 7.6 group (−0.333±0.024 V vs. SCE).

4.

Evolution of the potential during free potential tests. (A) Representative evolution of potential curves before, during, and after fretting motion. (B) The drop in potential (V_Drop) at the onset of the fretting motion. Polished surfaces showed significantly higher potential drop at both pH 3.0 and pH 7.6 levels (* P<0.05 and ** P<0.001). (C) The average potential throughout the applied fretting motion $\left({\overline{V}}_{\mathit{\text{Fretting}}}\right)$ (* P<0.05).

Fig. 5A shows the representative curves of the evolution of current density, measured from potentiostatic tests during the fretting motion. The charge is indicative of the electron transfer and the resultant metal ion dissolution rate. The curves were integrated as a function of time in order to attain the values of charge density. The accumulated charge was used to calculate the mass loss due to corrosion (K_C) from Eq. 2. The average charge density of all trials was used as a comparative tool to evaluate the differences in the interfacial dissolution rates between the treatment groups. Charge density was lower in the polished group than the machined group and was lowest for pH 7.6 samples within each group (Fig. 5B). The polished pH 7.6 group had the lowest charge density (0.57±0.12 mC/cm²), followed by the polished pH 3.0 group (1.34±0.05 mC/cm²), the machined pH 7.6 group (2.01±0.14 mC/cm²), and the machined pH 3.0 group (3.90±0.72 mC/cm²).

5.

Electrochemical data from the fretting-corrosion test. (A) Evolution of current measured from potentiostatic tests during the fretting motion. (B) The total evolved charge density throughout the applied fretting motion (* P<0.05 and ** P<0.001).

3.5. Evolution of hysteresis behavior under free potential mode

Fig. 6A shows representative hysteresis response curves of the shear load to displacement amplitude at each cycle. For any given cycle, the corresponding cross-sectional area of the hysteresis loop represents the dissipated friction energy, in Joules. Overall, the graphs indicate that the system had a run-in period of approximately 300–500 cycles or seconds (the frequency was 1Hz), after which, the system reached a relatively stable state. This hysteresis behavior can be alternatively expressed as the averaged curves of the dissipated friction energy throughout the entire imposed fretting motion (Fig. 6B and C). In the pH 3.0 group (Fig. 6B), there was no observable difference in dissipated friction energy between polished and machined samples, with approximately 4 to 5mJ/cycle of dissipated energy throughout the entire fretting motion. For the pH 7.6 group (Fig. 6C), the average dissipated friction energy was higher in the polished group than the machined group (approximately 6.1 mJ/sec vs. 2.1 mJ/sec, respectively).

6.

Hysteresis loops and evolution of dissipated friction energy. (A) Representative hysteresis response curves showing the time-dependent, shear load to displacement response at each cycle of throughout the applied fretting motion. (B) and (C) show the averaged curves of the dissipated friction energy throughout the entire imposed fretting motion (mean and standard deviations), for pH 3.0 and pH 7.6 groups, respectively.

3.6. Analysis of the total metal content in the solution

Average total mass loss (K_WC) was smallest for the polished group, and for samples with pH 7.6 within each group (Fig. 7, Table 2). The polished pH 7.6 group had the lowest K_WC (42.06±3.84μg), followed by the polished pH 3.0 group (53.17±2.30μg), the machined pH 7.6 group (60.81±42.06μg), and the machined pH 3.0 group (131.23±3.36μg). The individual contributions to K_WC from corrosion (K_C ) and wear (K_W) follow the same relative trends in mass loss for with the polished pH 7.6 group having the lowest relative values of mass loss (K_C = 2.13±0.46μg and K_W = 39.93±2.78μg), followed by polished pH 3.0 (K_C = 5.25±0.25μg and K_W = 47.92±1.42μg), machined pH 7.6 (K_C = 8.15±0.58μg and K_W = 52.65±2.86μg) and finally machined pH 3.0 (K_C = 16.50±3.04μg and K_W = 114.73±0.67μg ).

7.

Results of the individual contributions to total mass loss measurements (K_WC) due to wear (K_W) and corrosion (K_C) as a function of pH and surface topography. It was clearly observed that the machined surfaces bear the significantly higher total mass loss (K_wc, ** P<0.001, see also Table 2) compared to the polished surfaces at pH 3.0. Compared to pH 3.0, pH 7.6 showed significantly lower K_W (*p<0.01 & **p<0.001) and K_C (**p <0.001 & **p <0.001) both in polished and machined surface respectively.

Table 2:

Summary of the experimental results including the total mass loss measurements (K_WC), the individual contributions of the mass loss due to wear (K_W), and mass loss due to corrosion (K_C). The K_C/K_W ratio was used to identify the dominant degradation mechanism.

	pH	Polished (P)	Machined (M)
Kwc (μg)	3.0	53.17 (2.30)	131.23 (3.36)
	7.6	42.06 (3.84)	60.81 (4.66)
Kc (μg)	3.0	5.25 (0.25)	16.50 (3.04)
	7.6	2.13 (0.46)	8.15 (0.58)
Kw (μg)	3.0	47.92 (1.42)	114.73 (0.67)
	7.6	39.93 (2.78)	52.65 (2.86)
Kc/Kw	3.0	0.11 (0.00)	0.14 (0.03)
	7.6	0.05 (0.01)	0.15 (0.01)
Dominant mechanism	3.0	Wear-corrosion	Wear-corrosion
	7.6	Wear	Wear-corrosion

3.7. Surface characterization of the worn surfaces

The worn surfaces were examined by SEM to identify the variations in the wear pattern due to the fretting-corrosion tests. Fig. 8 shows the representative SEM micrographs of the wear scar regions of rods and pins from the machined group at pH 3.0 and pH 7.6. At pH 3.0, the CoCrMo pins exhibited grooves within the contact area in the sliding direction. Mild etching could also be observed in and around the contact area. Also, accumulations of predominantly titanium oxide were found outside of the contact area and also to a lesser degree within the contact area. On the machined Ti6Al4V rod samples, microgrooves were partially destroyed indicating plastic deformation, material breakout due to adhesion, and surface fatigue. Also, the presence of carbon-rich organic material and chlorine was detected on the surface and within crevices by EDS. At pH 7.6, the CoCrMo pins exhibited mainly the material transfer of titanium-oxide from the Ti6Al4V rod. Larger flakes on the surfaces contained titanium, aluminum, and vanadium indicating actual material transfer due to adhesion. Grooves along the sliding direction occurred as well, but to a lesser degree than at pH 3.0. At pH 7.6 the machined Ti6Al4V rod samples displayed prominent deformation and surface fatigue. In many areas, it appeared that oxidation of the surface occurred prior to crack formation and subsequent detachment of large oxide particles. Carbon-rich material and traces of chlorine were detected on the surface of both pH 3.0 and pH 7.6 samples. It is important to note that other damage features associated with pitting, intergranular corrosion, imprinting, and partial slip fretting as found on tapers from retrieved implants were not reproduced by these tests^32,49.

8.

SEM micrographs of the wear scar regions of rods and pins from the machined group at pH 3.0 and pH 7.6.

4. DISCUSSION

4.1. Fretting-induced electrochemical response: Influence of topography and pH levels

In this study, free potential experiments were used to compare the effects of fretting motion on the corrosion potential of the simulated modular junction system of a Ti6AL4V stem coupled with a CoCrMo head as a function of surface topography (polished vs machined with microgrooves) and pH variation (pH 3.0 and pH 7.6). The evolution of corrosion potential could be highly affected by the crevice conditions at the implant interfaces. In all free potential tests, there was a cathodic drop in potential, corresponding to the onset of the fretting motion (Fig. 4).

This drop in potential has been extensively reported in the literature as electrochemical destabilization⁵⁰. This is due to the disruption of the passive layers on the implant surfaces (due to loading and motion) and at the asperity interfaces between contacting materials⁵¹. A drop in potential has been well-reported in the literature under both sliding and fretting conditions (mechanical stimuli) in both Ti6Al4V and CoCrMo alloys^{25,26,37,52,53}. An increase in charge (Fig. 5A) is indicative of the electrochemical destabilization that occurs due to the presence of microgrooves. The higher current density in the machined group indicates that topography plays an important role in the corrosion kinetics of the system. In addition, the higher current density in the pH 3.0 groups indicates that the surrounding media also influences the corrosion kinetics, particularly in the case of Ti alloy, the passive film formation is inhibited^54,55.

4.2. Topography effects on the mechanical response

In general, the machined surfaces had inferior corrosion behavior. On the other hand, the machined group had a lower total dissipated friction energy throughout the entire test compared to the polished group. The dissipated friction energy is indicative of energy transfer between the surfaces leading to surface damage. In contrast, previous studies have indicated that the machined surface at taper junctions enhances mechanical stability⁵⁶. The induced high strain at both internal and external machined surfaces potentially led to cold welds at the taper junction. Furthermore, the long term stability of the taper junction also depends on the different geometric parameters, such as angle, diameter and engagement length which results in different moment arms under walking cycles and loadings. Taper dimensions influence the stability and subsequent fretting-corrosion behavior of the implant⁵⁷.

4.3. Topography effects on the synergistic interactions between fretting wear and corrosion

According to Stack et al.^42,58, the mechanical and electrochemical (corrosion) contributions to tribocorrosion can be classified by taking the ratio of K_C/K_W, with the transition criteria defined as:

$$K_C/K_W<0.1=\mathit{\text{Wear dominated}}$$ Eq. 6

$$1>K_C/K_W\ge 0.1=\mathit{\text{Wear}}-\mathit{\text{corrosion}}$$ Eq. 7

$$10>K_C/K_W\ge 1=\mathit{\text{Corrosion}}-\mathit{\text{wear}}$$ Eq. 8

$$K_C/K_W\ge 10=\mathit{\text{Corrosion dominated}}$$ Eq. 9

Based on these criteria, it is evident that synergistic interaction of wear and corrosion dominates in the case of the machined group and pH 3.0 samples within the polished group. For the machined group, the large surface area created by the gaps between microgrooves may increase corrosion processes. For the polished group, in the acidic environment of pH 3.0 samples, subsequent changes in electrochemical reactions may lead to accelerated corrosion of the polished group, compared to polished samples of pH 7.6 where wear dominates. As per Stack et al., the synergistic interaction of wear and corrosion may play a role in mechanistic transitions in wear and/or corrosion processes that eventually trigger early failures at hip modular junctions^6,20,59,60.

4.4. Topography comparisons with retrievals

The fretting-corrosion tests only reproduced a subset of damage features that have been observed THRs with a CoCrMo head and Ti6Al4V stem^32,57. Specifically, we found material transfer from the softer Ti6Al4V rod to the harder CoCrMo pins. The grooves that could be seen predominantly on the CoCrMo pins had some resemblance with grooves found parallel to the taper axis on retrievals. However, it has been reported that such features on retrievals are caused by plastic deformation and micro-ploughing during the assembly and disassembly of the femoral head. In addition, the grooves generated in this study are characteristic of gross-slip fretting, while the fretting damage that is typically found on retrieved components is suggestive of a partial slip fretting regime. Therefore, it appears likely that the in vivo acting fretting motion lies below the here achieved amplitude of 25 μm. The CoCrMo pins also showed no evidence of pitting, intergranular corrosion, and surface fatigue—damage features observed on retrieved CoCrMo head taper surfaces. Nonetheless, the use of fretting test samples with microgrooves leads to a better representation of the acting contact conditions, corrosion kinetics and surface passive film (Ti-oxide) formation as they occur in vivo. The oxide subsequently appears to crack and break off from the surface. A similar process has previously been reported by Fischer et al.⁶¹.

This in vitro parametric study has several limitations. The fretting test system used a flat-on-flat contact, which does not reflect the actual taper junction. Yet, the introduction of microgrooves to the sample surface is novel and allows for local contact mechanics that actually reflect in vivo local contact mechanics better than other commonly used setups such pin-on-flat. Another limitation is that only one microgroove geometry was investigated. However, the chosen topography represents the median dimensions of microgrooves on stem tapers in our retrieval database. In addition, the study could not evaluate the rheological/lubricating effect of fluid entrapped in the microgrooves during the fretting process. Finally, it the damage found on retrieved implant surfaces represents dynamic behavior over time and therefore involves many factors, such as patient-specific loading conditions beyond those investigated in the controlled experiments in this study.

5. CONCLUSIONS

The study suggests that the microgrooves increase the susceptibility to corrosion as a larger surface area is exposed compared to polished surfaces. The role of synergistic interaction of wear and corrosion is evident from the outcome of the study. Despite the shift towards higher electrochemical destabilization in the machined group, both polished and machined groups still exhibited a mechanically dominated degradation. The study also suggests that surface topographical changes play a critical role in the mechanical and electrochemical processes. Further studies are necessary to determine the optimal microgroove geometry in order to minimize the fretting induced by micro-motion and subsequent fretting corrosion at implant interfaces.