Tribological Behaviour of Ti6Al4V Alloy: An Application in Small Joint Implants

Vol.:(0123456789) 1 3 Tribology Letters (2023) 71:125 https://doi.org/10.1007/s11249-023-01795-4 ORIGINAL PAPER Tribological Behaviour ofTi6Al4V Alloy: AnApplication inSmall Joint Implants LukášOdehnal1 · MatúšRanuša1 · MartinVrbka1 · IvanKřupka1· MartinHartl1 Received: 15 June 2023 / Accepted: 4 October 2023 / Published online: 8 November 2023 © The Author(s) 2023 Abstract The presented study deals with the analysis of the tribological behaviour of the Ti6Al4V alloy manufactured conventionally. The study aimed to verify whether the titanium alloy is suitable for use as a contact material in small joint implants, as additive manufacturing of this alloy can in the future provide certain benefits, such as individualization and simplification of the implant construction, or controlled porosity. Thetested pair consisted of a pin and a glass plate lubricated with model synovial fluid. The contact area was observed with colorimetric interferometry. Alongside film thickness, friction, and wear scars were measured. From the designed experimental conditions, the titanium alloy was not able to create a sufficiently thick lubrication film to overcome its surface roughness and damage to contact surfaces occurred. Friction was comparable for all the tested configurations. The application of conventionally manufactured titanium alloy as a contact surface in small joint implants seems to not be suitable since its performance fell short when compared to conventional cobaltous alloy. Nevertheless, there are various alternative methods available, such as unconventional manufacturing, polishing, surface texturing, and coating. * Lukáš Odehnal [email protected] 1 Department ofTribology, Faculty ofMechanical Engineering, Brno University ofTechnology, Technická 2896/2, 61669Brno, CzechRepublic Tribology Letters (2023) 71:125 1 3 125 Page 2 of 15 Graphical Abstract Keywords First MTP joint replacement· Biocompatible materials· Pin-on-plate configuration· Lubrication film thickness 1 Introduction Interventions, especially the replacements of human joints, have become a necessity for today's population. The number of all joint replacements applied to the human body is rising every year. In the human body, there are various kinds of joints, and their full functionality is necessary for everyday life, for example, small joints, such as the big toe's first metatarsophalangeal joint (first MTP joint). This joint is responsible for a person's stability and is the most stressed joint in the foot during normal movement activities. Based on the study by Korim etal. [1] which reviewed the conducted arthrodesis, hallux valgus and hallux rigidus are the most frequent diseases affecting the first MTP joint, accounting for 36.6% and 34% of the cases, respectively. Recently, mainly hallux valgus has been discussed, as many middle-aged women suffer from this deformity because it is closely associated with wearing tight shoes or high heels. There are two main approaches to treating these deformities. The first, which is mainly invasive, is arthrodesis (fusion). This is adefinitive and irreversible operation in which the affected joint is immobilized by fusing the bones together. The second option is joint arthroplasty. With this method, the joint motion can be preserved and this intervention does not affect the gait pattern [2]. Based on the statistics from Germany [3], between 2008 and 2017, only a few of the first MTP joint arthroplasties were made compared toarthrodesis, although it may have certain benefits. On the other hand, the first MTP joint replacements are not as reliable as arthrodesis these days. One of the most common types oftotal arthroplasty for the first MTP joint is ToeFit-Plus® [2]. Titchener etal. [4] reported a revision rate of 24% at an average of 33months post-operatively for this implant, while the majority of revisions were caused by frank loosening or progressive lucency, mainly on the phalangeal side. The first MTP joint replacements evolved in shape or used materials in the past [5]. During these days, the most common type is a metal-on-polyethylene total MTP replacement. These replacements consist oftwo stems from titanium alloy with porous structures, a CoCrMo metatarsal head, and Tribology Letters (2023) 71:125 1 3 Page 3 of 15 125 UHMWPE phalangeal plateau. Lately, additive manufacturing (AM) has become more frequent in all production areas, and this method is also suitable for joint replacements as the AM might bring benefits intheproduction of personalized metal-on-metal (MoM) small joint implants [6]. The most common method for producing artificial joints is selective laser melting (SLM) [7]. In such manufacturing, Ti6Al4V is more suitable than common CoCrMo alloy which is the most common contact surface injoint replacements these days [8]. The titanium alloy has excellent biocompatibility, good corrosion resistance, and high strength ratio [9]. Moreover, this alloy fabricated with a porous structure has comparable mechanical properties to a natural bone and can improve ingrowth of the stems and increases bone-implant stability [10]. To ensure the applicability of Ti alloy, the material must be subjected to detailed testing and compared with conventional Co alloy based on their performances. To bring the experimental conditions close to reality, boundary conditions, such as kinematics andload, have to be defined. The kinematics of the first MTP joint was described by Durrant etal. [11]. The model provided knowledge about the joint movements and described their variance between the individual subjects. The loading of the joint, or the contact pressure, was examined by Flavin etal. [12] andAl-Munajjed etal. [13]. The main feature ensuring the proper behaviour of the replacement in the human body is the ability to form a sufficiently thick lubrication film that can separate interacting parts ofthereplacement in order to reduce the generation of wear particles. The formation and thickness of the lubrication film can be influenced by various parameters. Myant etal. [14] demonstrated the impact of contact pressure, where its increase resulted in a decrease of lubrication film thickness. Kinematics, respectively the relative speed and slide-to-roll ratio (SRR) was described by Nečas etal. [15], who showed that the relative speed works differently for various SRR, i.e., the increase in relative speed for SRR 0 led to an increase in film thickness; on the other hand, for SRR 1.5, it led to a decrease infilm thickness. Another factor potentially influencing the behaviour may be connected to contact surfaces, as human joints are tested with synovial fluid (SF) consisting ofproteins with the ability to adsorb on the contact surfaces and its presence can affect the overall behaviour [16] of the system. The ability to form a stable and sufficiently thick lubrication film is usually connected tothe number of proteins in the contact area. Nečas etal. [17] presented that the behaviour of lubrication film thickness is dependent on the composition of SF and tried to connect the behaviour of individual constituents, such as albumin and γ-globulin, to the lubrication film thickness. It was found that the trends of albumin and lubrication film thickness were comparable, while the γ-globulin development was observed only on a small scale [15]. Therefore, it appears that the main role of forming the lubrication film lies in albumin, while its presence and maintenance in the contact area might be affected by other constituents [18, 19]. The study made by Ranuša etal. [20] showed differences in the behaviour of samples with differences in surface topography. The tested Ti alloy had significantly worse surface roughness compared to Co alloy, resulting in the presence of a larger amount of proteins in the contact area. The well-formed and stable lubrication film is closely connected to other observed parameters, such as friction and wear, while these two parameters are closely related to good functionality and the lifetime of the replacement. In the case of load, the increase of normal force leads to a decrease in friction [21, 22]. On the other hand, while lower friction occurs in higher load conditions, findings by Gao etal. [23] showed that the higher load contact is more likely to produce a higher rate of wear particles; therefore, lower friction does not always mean a lower wear rate. The studies dealing with the lubricant concentration [24–26] similarly showed that friction increases with an increasing number of proteins in the lubricant. Whencomparing the Co and Ti alloys according to the coefficient of friction (CoF), the values are not so different [27] On the other hand, when comparing the generation of wear particles, the titanium alloy is worse, while the released particles were approximately twice as high as for CoCrMo [28]. Thesefindings could mean that the titanium alloy is incompatible with the use in the joint implants forcontact surfaces, as the excessive number of released particles is undesirable for the patient’s health [29]. Nevertheless, the benefits coming with additive manufacturing and simplifying the implant shape construction predetermine that the use of this alloy could find its application. Based on the performed research and the knowledge gained in the past in this area of interest, further and complementary research seems more than needed since a better understanding of the given phenomena can lead to an increase in the lifespan and functionality of the implant, which are currently a big problem, as the revision operations rate is still too high for this joint [4]. To obtain such research that can reveal the behaviour in the contact area, the problem has to be examined at its roots. The main question posed in this publication was “What are the main differences in the tribological behaviour of conventionally manufactured CoCrMo and Ti6Al4V alloys in the simulated small joint implant (first MTP joint)?” andthesub-question: “Is the Ti6Al4V alloy suitable for further investigation as a potential material for human joints?” To answer these questions, the research is supplemented with the development of lubrication film thickness, an analysis of friction, and an analysis of contact pair roughness, or its wear scars. Combining these observed aspects should provide the necessary insight into the issue and show whether it makes sense to conduct further research on Ti6Al4V alloys as possibilities for contact Tribology Letters (2023) 71:125 1 3 125 Page 4 of 15 surfaces in small human joints, where additive manufacturing might find its application. 2 Materials andMethods 2.1 Experimental Apparatus The experiments were performed on the universal tribometer [20, 30], providing reciprocal movement in a pin-on-plate configuration with a stable sample (pin) and a moving plate (glass). Theexperimental apparatus is shown in Fig.1. The tribometer construction enables the use of optical methods to observe the contact area and measurement of the friction simultaneously. As the main goal of the study was to describe thelubrication mechanism, colorimetric interferometry [31] was used to observe the contact area toobtain information on the film thickness and its development. 2.2 Contact Pair andLubricant The contact pair consisted of a pin made from alloys used in implantology: CoCrMo(ASTM-F75) or Ti6Al4V (ISO 5832-3). The implant manufacturer [32] produced the test pins using a certified process involving cold drawing, followed by machining to a 15mm radius, and polishing to a roughness ofRa = 0.01 ± 0.005μm forCoCrMo, and Ra = 0.04 ± 0.01μm for Ti6Al4V as the conventional machining for joint implants allowed for the used materials. The finishing process followed the certified method, which showed that it was not possible to bring the surface roughness of Ti6Al4V alloy closer to the one obtained for CoCrMo alloy. The second articulating part was a plate made from B270 glass, as one transparent part was necessary for observations using colorimetric interferometry. Toenhance the observation conditions, the plate was on the contact surface coated with the Chromium layer (reflectivity of 25%). The other side of the plate was covered with an antireflective layer. The use of counterpart made from glass shows that the observed simulated joint more likely corresponds to the metal-on-metal joint implant type, as the material characteristics are closer to each other than for the metal-on-polymer type. Thematerial characteristics of contact bodies are shown in Table1. The presented study used the model SF as the testing lubricant. The model SF was synthetically prepared based on the samples extracted from patients with arthroplasty [33]. Thefinal composition of the model SF was made by diluting the required number of constituents inphosphate-buffered saline (PBS). The concentration of diluted components, respecting the properties of SF for patients after arthroplasty, is as follows: albumin (26.3mg/ml), γ-globulin(8.2mg/ml), hyaluronic acid (0.82mg/ml), andphospholipids (0.35mg/ ml). 2.3 Loading andKinematic Conditions ofExperiments As there is no available ISO standard for testing of small joint implants, especially the first MTP joint investigated in this study, the loading and kinematic conditions had to be determined based on the analytic models or ISO standards used for testing of total knee replacement (TKR) [34] or total hip replacements (THR) [35]. Durrant etal. [11] presented a model where the initial metatarsal declination angle is 15°, and the terminal declination angle is approximately 80°; these angles can vary for different subjects. Using a Fig. 1 Scheme of the experimental apparatus Table 1 Material characteristics, loading, and the resulting contact area based on the Hertz theory Material Young’s modulus Poisson’s ratio Applied load Contact pressure Calculated contact area CoCrMo 241 GPa 0.29 0.5 N 105.5MPa 0.095mm Ti6Al4V 114 GPa 0.34 0.73 N 105.5MPa 0.115mm Glass B270 71 GPa 0.22 Tribology Letters (2023) 71:125 1 3 Page 5 of 15 125 typical radius of total replacements for this joint, the contact path was calculated for ~ 12mm. Nevertheless, the 20mm stroke was used in the experiments due to dead ends occurring in the testing cycle as the measuring of friction and observing the contact area requires a sufficiently long stroke with stable normal load and speed stabilization. The tested speeds were chosen according to the speeds that may appear in the 1. MTP joint during the gait cycle based on the sagittal plane kinematics presented by Allan etal. [36]. Based on the development of the declination angle, the recalculation of peripheral speed was made. Considering that the duration of the cycle is one second [34, 35] and the typical radius of rotation for the 1. MTP joint is about 13mm [37, 38]; the recalculation showed that for the most of the cycle, the speed is very low (approx. 1–6mm/s). The considerable speeds occur between 40 and 65% of the cycle, varying from 7 to 65mm/s. Two speeds, 20mm/s and 40mm/s, were selected from this range for the experiments. The expected contact pressure is in the 1. MTP joint highly influenced by high conformity of the implant´s parts, where both curvature radiuses are 10mm. This results in contact pressure at about 7–8MPa [12]. When transformed to the experimental conditions (pin-on-plate), this contact pressure (~ 7.9MPa) is obtained for the material combination of CoCrMo and UHMWPE by applying 0.5N. In order to maintain the recalculated contact force of 0.5N on the material combination used in experiments (CoCrMo on Glass), the resulting contact pressure was 105.5MPa. To compare the behaviour of two tested materials (CoCrMo and Ti6Al4V), the contact pressure was unified in this study. The resulting contact forces applied on two tested configurations after recalculation based on the contact Hertz theory are as follows: 0.5N (CoCrMo/Glass) and 0.73 (Ti6Al4V/Glass). The load was monitored and controlled during the experiments to ensure no more than ± 5% deviation from the desired value. 2.4 Experimental Setup andEvaluation oftheResults The experiments were designed to observe the behaviour of lubrication film thickness in short terms. The experiment was divided into three parts. After each part, the contact pair was unloaded todepict the relief of the joint. Each partial experiment consisted of 20 cycles, which gives the number of 60 cycles for the whole experiment, composed of two unloading stages. Each configuration was tested on three different contact pairs to achieve sufficient repeatability. The presented results were evaluated from the observation of contact in only one direction (marked as “Evaluated area” in Fig.2A). This area is also cropped off the first and last 2mm of the deflection; as in these parts (dead ends) of the cycle, therelative speed was not constant and it was not marked as relevant to the results. Continuous development of both CoF and lubrication film thickness was recorded during the experiment (seeFig.2B, C). For CoF, the whole marked area – green oval (see Fig.2B) is represented in the results by the average value for each cycle. While observing the lubrication film thickness and its development over time, it was found out that the thickness was dependent on the observed point of the cycle. In light ofthis finding, three points were selected and observed to describe possible inconsistencies during the cycle. These points are marked green in Fig.2C. 3 Results 3.1 Analysis ofSurface Roughness Before theExperiment A roughness analysis of the samples was carried out to define the boundaries of lubrication film thickness, i.e., a separation of the contact surfaces with a certain probability. A sufficiently close zone adjacent to the contact area was selected for this observation. The zone was based on the expected size of the calculated contact area, which was doubled in cases where the contact area did not appear on the expected canopy. The surface roughness distribution Fig. 2 Scheme for the evaluation of the results: A relative speeds with a marked area of evaluation; B evaluation of CoF; C evaluation of lubrication film thickness Tribology Letters (2023) 71:125 1 3 125 Page 6 of 15 (its irregularities) of both examined alloys was Gaussian. Two borders were determined, the first representing a 95% probability of separation of contact pairs and the second a 99% probability. Thevalues were established based on the measurements (see Table2) to 35nm (CoCrMo) and 110nm (Ti6Al4V) for a probability of 95%, or to 45nm (CoCrMo), and 160nm (Ti6Al4V) foraprobability of 99%. These borders are represented in the results graphs as red dashed, or solid lines. 3.2 Lubrication Film Thickness For the lubrication film thickness, the experiments with an entrainment speed of 20mm/s (seeFig.3) showed similar results in case of reaching a sufficiently thick film layer for both of the materials, and the separation of contact pairs did not occur (except for the first few cycles for 2. PIN made fromTialloy). The measured thickness was generally higher for the Ti alloy (e.g. point at 10mm deflection at 30th cycle: ~ 15nm for CoCrMo and ~ 55nm for Ti6Al4V). Nevertheless, due to its worse surface roughness, it was not sufficiently high as the thickness did not reach either of the shown boundaries (red dashed or solid line). On the other hand, the experiments with an entrainment speed of 40mm/s showed different results for each material (see Fig.4). While using the CoCrMo alloy, the contact pairs were separated for almost the whole experiment (except for the first few cycles for 1. PIN). On the contrary, while the Ti6Al4V alloy was used, the film thickness showed quite similar behaviour as at a lower speed and it did not reach sufficient values to separate the contact pair completely (except for the first few cycles for 2. and 3. PIN). 3.3 Wear Scars afterExperiments The findings regarding the lubrication film thickness of tested samples correspond with the roughness measurements (wear scar analysis) after experiments, where a combination of CoCrMo and glass at 40mm/s showed no signs of wear. Overall, the wear scars on the tested samples differ for each material and relative speed applied (see Fig.5 and Table3). The CoCrMo samples are most likely to create a regular oval wear area oriented withalonger axis in the direction of the movement (Fig.5A). On the other hand, the Ti6Al4V pins did not create such an oval wear area, but these samples were characterized by the formation of separated grooves that were considerably deeper and also longer in the direction of the movement (Fig.5B). Table3 shows the characteristics of the wear scars for all tested samples. Significant differences were observed for Ti alloy pins at 20mm/s where 1. PIN showed similar wear scars asCopins (oval area) and 2. PIN had no scars at all. The typical wear scars for each material are illustrated more precisely in Fig.6. The wear mechanism of Ti6Al4V and CoCrMo alloys differs, as observed in the images. The images of Ti6Al4V show a non-uniform wear pattern accompanied by the creation of deeper separate grooves, while the wear area of the CoCrMo sample is characterized by the creation of regular oval wear areas without irregularities. The images depict the abrasion for both materials, but the release of Ti6Al4V wear particles caused the three-body abrasion. 3.4 Coefficient ofFriction Figure7 shows the evolution of CoF for all tested configurations. The yellow dots, representing the average value of CoF for three samples are in the graphs supplemented with standard deviation (SD). All configurations showed a similar behaviour, with CoF settled down at values around 0.4 after an initial increase. The rehydration stages after 20 or 40 cycles (marked with vertical black dashed lines) did not significantly affect the behaviour of CoF at the tested speeds of 20mm/s. On the other hand, when the relative speed of 40mm/s was applied, a slight decrease of CoF can be seen after the rehydration stages for Ti6Al4V and a much greater decrease for CoCrMo. Nevertheless, the decrease lasted only one cycle and then the friction returned to the expected value. The values of CoF forCoCrMo at 40mm/s (the only configuration that reached the selected boundaries for separated contact pairs) were the most inconsistent with the highest SD (see Fig.7D). The development of friction was supplemented with stacked graphs of friction force evolution throughout the experiments for each configuration (see Fig.8). The friction force values were evaluated using the methodology outlined in Fig.2A as averages of three repeated experiments. Grey stripes indicate the beginning and end of each cycle, representing parts of the cycle, where the relative speed was unstable. Each material displayed a unique stacked pattern of the evolution due to the application of different loading forces. The development in each cycle Table 2 Probability of full film lubrication based on contact pair roughness Material Sample 95% Probability of separated contact pairs 99% Probability of separated contact pairs CoCrMo PIN 1 30nm 41nm PIN 2 35nm 47nm PIN 3 34nm 48nm Ti6Al4V PIN 1 102nm 166nm PIN 2 113nm 182nm PIN 3 100nm 128nm Tribology Letters (2023) 71:125 1 3 Page 7 of 15 125 Fig. 3 Development of film thickness at 20mm/s for A CoCrMo and B Ti6Al4V. Interferograms with marked contact areas correspond to 3.PIN for each configuration (the inlet zone is on the left side) Tribology Letters (2023) 71:125 1 3 125 Page 8 of 15 Fig. 4 Development of film thickness at 40mm/s for A CoCrMo and B Ti6Al4V. Interferograms with marked contact areas correspond to 3.PIN (CoCrMo) and 2.PIN (Ti6Al4V) (the inlet zone is on the left side) Tribology Letters (2023) 71:125 1 3 Page 9 of 15 125 remained quite stable during the entire experiment after the initial rise due to the increase in relative speed for all configurations except for the Ti6Al4V at 40mm/s, where an additional lower peak occurred at around 5mm, and the upper peak at around 7mm of displacement. 4 Discussion General Discussion The presented study focused on comparing and describing the tribological behaviour in the contact area of two materials suitable for human joint implants, a conventional CoCrMo alloy, and a 3D printing-friendly Ti6Al4V alloy. The experiments were designed to provide a general overview of these two conventionally manufactured materials. Three tribological parameters were measured or observed: lubrication film thickness, friction coefficient, and wear scars on the tested pins. The kinematics and load used in the experiments on the reciprocal tribometer with pin-onplate configuration [30] were designed based on the literature [11–13] or ISO standards [34, 35], while the deflection in one direction was set at 20mm. To obtain a frequency of 0.5Hz and 1Hz, the applied entrainment speed was 20mm/s and 40mm/s, respectively. The load was set to0.5N for CoCrMo alloy and 0.73 N for Ti6Al4V, as it was recalculated using the Hertz contact theory based on similar contact pressures. The experiments showed that it is difficult to create a sufficiently thick lubrication film layer underthe experimental conditions designed, as the height of the lubricant did not reach values sufficient toovercome the average roughness of the contact pair (see Figs.3 and 4). This is accompanied by the formation of wear scars that are undesirable in human joints (see Fig.5 and Table3), as Fig. 5 Wear scars on A CoCrMo samples, and B Ti6Al4V samples. Graphs represent the profile cuts of PIN 1 of each material with marked borders (red and green lines), corresponding to surface pictures ofwear areas Table 3 Wear scars on tested samples * The wear scar of the Pin was similar to the ones in CoCrMo/Glass configuration Tested conditions Sample x-axis width y-axis width Central depth Tested conditions Sample x-axis width y-axis width Central depth CoCrMo Glass 20mm/s PIN 1 157μm 124μm 86nm Ti6Al4V Glass 20mm/s PIN 1* 168μm 123μm 127nm PIN 2 139μm 120μm 92nm PIN 2 Non-visible signs of wear PIN 3 126μm 86μm 81nm PIN 3 139μm 87μm 203nm CoCrMo Glass 40mm/s PIN 1 Non-visible signs of wear Ti6Al4V Glass 40mm/s PIN 1 255μm 154μm 583nm PIN 2 Non-visible signs of wear PIN 2 196μm 65μm 437nm PIN 3 Non-visible signs of wear PIN 3 298μm 131μm 984nm