Current status and future potential of wear-resistant coatings and articulating surfaces for hip and knee implants

Hip and knee joint replacements are common and largely successful procedures that utilise implants to restore mobility and relieve pain for patients suffering from e.g. osteoarthritis. However, metallic ions and particles released from both the bearing surfaces and non-articulating interfaces, as in modular components, can cause hypersensitivity and local tissue necrosis, while particles originating from a polymer component have been associated with aseptic loosening and osteolysis. Implant coatings have the potential to improve properties compared to both bulk metal and ceramic alternatives. Ceramic coatings have the potential to increase scratch resistance, enhance wettability and reduce wear of the articulating surfaces compared to the metallic substrate, whilst maintaining overall toughness of the implant ensuring a lower risk of catastrophic failure of the device compared to use of a bulk ceramic. Coatings can also act as barriers to inhibit ion release from the underlying material caused by corrosion. This review aims to provide a comprehensive overview of wear-resistant coatings for joint replacements – both those that are in current clinical use as well as those under investigation for future use. While the majority of coatings belong predominantly in the latter group, a few coated implants have been successfully marketed and are available for clinical use in specific applications. Commercially available coatings for implants include titanium nitride (TiN), titanium niobium nitride (TiNbN), oxidized zirconium (OxZr) and zirconium nitride (ZrN) based coatings, whereas current research is focused not only on these, but also on diamond-like-carbon (DLC), silicon nitride (SiN), chromium nitride (CrN) and tantalum-based coatings (TaN and TaO). The coating materials referred to above that are still at the research stage have been shown to be non-cytotoxic and to reduce wear in a laboratory setting. However, the adhesion of implant coatings remains a main area of concern, as poor adhesion can cause delamination and excessive wear. In clinical applications zirconium implant surfaces treated to achieve a zirconium oxide film and TiNbN coated implants have however been proven comparable to traditional cobalt chromium implants with regards to revision numbers. In addition, the chromium ion levels measured in the plasma of patients were lower and allergy symptoms were relieved. Therefore, coated implants could be considered an alternative to uncoated metal implants, in particular for patients with metal hypersensitivity. There have also been unsuccessful introductions to the market, such as DLC coated implants, and therefore this review also attempts to summarize the lessons learnt.


Introduction
A damaged or diseased joint can cause severe pain and limited mobility, which impairs the quality of life of the afflicted individual. To treat this condition, it might be necessary to replace the joint with an implant in a surgical procedure. Two of these types of implants, namely hip and knee implants, will be the focus of this review. The number of primary, or first time, total hip surgeries performed every year has steadily increased and is predicted to continue to rise [1] as the ageing population increases, and indications are that younger patients can also benefit from these procedures. While hip joint implants have a generally high survival rate of approximately 95% or more at 10 years for all currently used material combinations, metal-on-polyethylene (MoP) being the most common, [2][3][4][5]. However, the increasingly active, ageing and younger populations require longer lasting implants and a focus of current research is on prolonging the lifespan of implants to spare patients from further pain and revision surgeries. Similarly, the revision rates for knee implants are reported to range from 4.3% [5] to 5.3% [4] at 10 years.
The primary cause of revision surgery for metal or ceramic hip replacement components paired with polyethylene (PE) is aseptic loosening of the implant (28-51% [2][3][4][5][6][7][8][9][10]). In these, one of the main causes for this is believed to be the presence of wear debris from the articulating surfaces, mainly polymeric particles. However, it should be noted that PE wear has decreased dramatically since the introduction of highly crosslinked polyethylene (XLPE) 20 years ago [11]. Necrosis, pseudo tumours and pain have also been found to be the cause of revision surgeries particularly in alternative metal on metal bearing systems. These complications are believed to be caused by metallic particulate and soluble (ionic) debris originating from the articulating and non-articulating surfaces of the implant [12][13][14]. While the particulate debris may originate from any surface of the implant, it is primarily formed at the articulating interface.
The most common materials currently implanted in this context are cobalt chromium alloy (CoCr) and highly crosslinked polyethylene (XLPE) in MoP implants [3,4,9,10], together with a zirconia/alumina ceramic combined with PE (CoP) [4,10]. In these cases, the dominant particulate debris is that of the polymer, which can range between 10 nm and 1 mm in size. Particles in the size range 0.1-1.0 μm are believed to play an important role in the activation of macrophages [15], which might initiate a cascade of reactions that eventually cause wear-induced osteolysis (loss of bone), possibly resulting in implant loosening [15][16][17][18][19]. Metal-on-metal (MoM) hip re-surfacing implants were considered an alternative for young patients, however, the implants experienced high short-term failure rates and an increased rate of revision surgeries [20]. Due to these catastrophic outcomes MoM hip replacements are now rarely used [2,4,5,8]. Debris from MoM implants is generally in the nanoscale range (<100 nm) and there is evidence to suggest that metal particles in this size range and ions can cause a variety of biological effects such as hypersensitivity, pseudo-tumours [12,14,[21][22][23][24][25] and aseptic lymphocytic vasculitis-associated lesions [13]. In addition high concentrations of cobalt and chromium ions are believed to cause e.g. acute visual and auditory impairment, peripheral neuropathy and cardiomyopathy [26]. It should be noted that the latter are extreme cases, and the incidence of adverse reactions to metal debris has been found to be less than 1.2% for MoM hip implants [27]. The revision per 1000 prosthesis years caused by adverse reactions to particulate debris for MoM was found to be 9.90, which can be compared to 0.19 for MoP [5].
Several different strategies have been explored for reducing wear and the negative biological effects of metal ions and PE wear debris. Among the most noteworthy of these are the improvement of polymer wear properties [11,28] and the development of e.g. textured surfaces [29]. However, this review will focus on the use of ceramic coatings or surface modifications, such as nitriding, to reduce wear and ion release. These aim to combine the advantages of a wear resistant ceramic with the ductility and toughness of a metal. The ceramic coating can also act as a barrier to, and decrease, ion release from the underlying metal [30].

Methodology
The material in this review is based on published research articles, patents and product information in order to cover all phases from early stage research to products in clinical use. The research articles were found using the database SCOPUS with the primary search terms "wear" AND "coatings OR film" AND "joint implant" for coatings and the search terms "wear" AND "surface modification" AND "joint implant" for surface modifications. An additional search with the terms "biocompatibility" AND "coatings OR film" AND "joint implant" was conducted to include studies regarding the biocompatibility and finally, through references, an additional 6 papers were identified. The patents were found by using the same search terms on Espacenet, and the products were found using Google (with search terms "joint replacement" "products" "coating"). Research papers, patents and products concerning dental implants, osteoinductive/conductive coatings, protein and other biological coatings as well as bulk materials were excluded and the joint types were limited to hip and knee. This left a total number of 89 research papers (27 of them being case reports or revision studies) and 4 patents. Information such as wear rates and surface roughness was compiled in a table that may be found in the supplementary information. Therefore, some of the references are not included in the main paper but can be found in the supplementary information. Fig. 1 shows a flow diagram of the procedures followed in selecting the reports and associated data for inclusion in this review.

Substrate material and deposition methods
A coating used for joint implants should be hard, wear resistant, corrosion resistant, biocompatible, not release particulates and any particulates that are generated should be biocompatible. In addition, the coating should have a low surface roughness and a good adhesion to the substrate material. Some of these properties are specified in standards such as ISO 7206-2 while others, such as wear resistance, have no specified target value. The important properties and what to consider are specified in Table 1. It should be noted that the table distinguishes between wear resistance and corrosion resistance, however theses are processes that occur synergistically. Typically, these properties are however reported separately.
These properties are covered in more detail in the supplementary information section.
Ceramic coatings are typically deposited onto a metal implant using a wide range of techniques [74][75][76]. Possible substrate materials include the metals commonly used as bulk materials for joint implants i.e. CoCr and stainless steel but also titanium. Titanium is not used in articulating surfaces due to its poor wear properties [77][78][79], but this becomes possible with a coating or surface treatment that improves such properties [75]. The use of titanium as a substrate material could even be advantageous in terms of coating adhesion [80]. It is however important to keep in mind that the coating should be compared with the material it aims to substitute, e.g. in the case of articulating surfaces typically CoCr.
The choice of deposition method or surface treatment will have a strong influence on the above-mentioned parameters. The most common deposition techniques used for coatings for articulating surfaces can be divided into either physical vapor deposition (PVD) methods [81] or chemical vapor deposition (CVD) methods [82], the former group is more commonly used than the latter in the studies found in this review. These deposition techniques, as well as their advantages and challenges, have been covered in detail in previously published articles [82][83][84][85][86]. PVD techniques include, amongst others, vacuum evaporation, magnetron sputtering (MS), reactive magnetron sputtering (rMS), pulsed laser deposition (PLD) and high-power impulse magnetron sputtering (HiPIMS). They utilise metallic sources that are either evaporated, typically using an electron beam, or sputtered with the help of a plasma to produce a flux of metal atoms/ions which is deposited onto the surface to be coated in the presence of one or more reactive gases to form a layer of ceramic material. Coatings deposited using PVD techniques typically have a low surface roughness, are hard and wear resistant. There are however limitations due to the technique being line-of-sight but these are usually mitigated by manipulation of the substrate during deposition. The second group of methods, CVD, entail exposing the substrate to a mixture of gases that react at high temperature (typically >600 C) to form a ceramic compound, though it is possible to use a plasma to enhance the reactivity of the gas precursors (plasma-enhanced CVD (PECVD)) to reduce the deposition temperature. Using CVD techniques it is possible to grow uniform, well adhering coatings on complicated substrates. However, there are limitations related to e.g. heat resistance of the substrate. Another approach is to treat the surface by e.g. heat or laser in the presence of gases such as nitrogen [87]. This often yields an increased hardness and wear resistance, however the surface roughness is often increased and the techniques are limited to materials that can form suitable ceramic surface layers such as zirconium oxide and titanium nitride.
Another family of surface deposition techniques is thermal spraying in which heated or melted material is sprayed onto a surface [88]. The coating material is either in the form of powder or a filament that is typically melted by flame or plasma [75].
A harder surface can also be achieved by surface treatment processes such as plasma electrolytic oxidation where metals are oxidized through a process similar to anodization but with higher potentials [89].
A technique with future potential is additive manufacturing where it is theoretically possible to manufacture implants with a surface layer with different properties compared to the underlying material. However, the manufactured materials currently require extensive post processing to achieve smooth surfaces suitable for wear-resistant applications [90].
The most important properties relevant to coatings for joint implants, as well as the methods used in their evaluation, are discussed in the supplementary information section. It should be noted however that somewhat different requirements exist depending on joint application, e.g. hip joints may experience more complex movement patterns, as well as more severe edge loading [91] than knee joints [92], and different standards are available to evaluate their performance.
TiN coated implants are available with either a Ti6Al4V [100] or a CoCr [101] substrate and the coated surface is usually paired with PE in the articulating surfaces. These implants are aimed for young, active patients, in particular patients with metal sensitivity [102] and studies have shown them to be stable over time [103].
As TiN, ZrN and OxZr are commercially available it is possible to study patient outcomes ( Table 2). Since their introduction, long-term follow up studies have found that OxZr implants perform well, with low revision rates [55], comparable to uncoated CoCr implants [104,105] (Table 2). The implant has performed better, with lower rates of corrosion and fretting wear, in comparison to CoCr, as seen in retrieved implants [106]. Taking all these results together Oxinium implants show great promise as a viable option for joint implants, not just for patients suffering from or at risk of developing metal hypersensitivity but also for otherwise healthy patients.
However, there have been cases with both TiN and OxZr in which retrieved implants were reported to exhibit surface damage, including an exposed substrate [100][101][102][103]108]. While the femoral heads in these studies have been paired with a PE liner the oxidized surface has shown damage and metallic transfer likely caused by contact with the metallic shell following dislocations. The authors concluded that Oxinium femoral heads should not be used in patients at risk of joint instability.

Potential candidate wear-resistant coatings
Several coatings are being investigated for their potential to reduce wear in joint implants. The following section will examine the published research, categorized into six groups according to coating composition: diamond-like carbon, silicon nitride, chromium nitride, zirconium based, titanium based and tantalum based. A summary of the properties can be found in Table 3 (hardness, Young's modulus and adhesion) and Table 1 in supplementary information (surface roughness and wear properties). The properties vary, with hardness ranging from 8 to 44 GPa and Young's modulus from 100 to 466 GPa. The tribological properties such as wear rates were obtained using different set-ups, with differences in contact pressure and counter surface, leading to large inherent variation between samples, making results difficult to compare.

Diamond-like carbon and nano-crystalline diamond
The term diamond-like carbon (DLC) covers a range of hard carbonbased materials with a wide range of properties such as hardness and wear resistance. The variety of structures and consequently properties for carbon is explained by its ability to exist in three hybridizations (sp 1 , sp 2 and sp 3 ). DLC coatings have a significant fraction of sp 3 bonds [137]. However, it is important to keep in mind that DLC coatings do not consist solely of amorphous carbon (a-C) but also hydrogenated amorphous carbon (a-C:H) alloys and the structure of the coating will largely depend on the deposition method. For example sputtered coatings typically can extend from sp 2 to sp 3 and plasma-enhanced CVD produced coatings have a higher fraction of hydrogenated carbons [137].
DLC coatings have been considered for joint implants because of the promise of a chemically inert, hard, wear resistant surface and significant research has been conducted on DLC-coated implants industrially; e.g. in 2001 the company Implant Design AG put forward a DLC-coated knee implant but had to withdraw it the same year after it was banned by the Swiss Federal Office of Public Health (SFOPH). The implants failed due to excessive wear caused by partial delamination of the coating, which led to early revisions [128,138]. Since then research has focused on understanding the mechanism behind the failures as well as improving the coating adhesion. Falub et al. observed an interlayer measuring approximately 5 nm in thickness that occurred gradually between the CoCr implant and DLC coating. This interlayer consisted mainly of carbides with an overall stoichiometry close to Me 2 C and delamination is believed to be caused by the instability of Co carbides in this interlayer ( Fig. 1) [139]. The mechanism is believed to be stress-corrosion cracking, i.e. delayed failure due to environmentally induced crack propagation [140].
Attempts to improve the adhesion by using interlayers was made by Thorwarth et al. [37] in a study whereby tantalum (Ta) was used as an interlayer for a DLC coating. The results were promising, with little noticeable wear for DLC coatings with Ta interlayers in the case of low concentrations of oxygen impurities. However, it was hypothesised that oxygen contamination would lead to an increased occurrence of the β-phase in the Ta interlayer, which could lead to mechanical failure due to its brittleness. Wang et al. on the other hand proposed the introduction of a fullerene-like structure and incorporating fluoride (F-FLC) to obtain long-term stability in in vitro environments [132]. The results revealed coatings with lower coefficient of friction and wear compared with DLC as well as promising in vitro results with e.g. cell adhesion of rabbit bone marrow mesenchymal stem cells. In addition Cr has been used to in an attempt to improve adhesion and these coatings exhibited improved corrosion resistance compared with CoCr [141].
The closely related nanocrystalline diamond coatings (NCD) consist of nano-sized diamond crystals in the range of 3-15 nm with a large fraction of amorphous carbon at the grain boundaries [142]. These coatings have similar wear properties to DLC coatings and have been shown to reduce the wear compared to uncoated Ti6Al4V components [143]. However, as previously mentioned CoCr is the common choice for articulating surfaces due to its superior wear properties, and a comparison to this material is lacking. Cell studies using primary human bone marrow cells have shown cell attachment, spreading and proliferation on the coatings, i.e. non-cytotoxicity. The available biocompatibility studies however seem to focus on osteoblasts, which may be more suitable for coatings aimed at bone ingrowth [144][145][146].
After investigating and addressing the risk of delamination in DLC coatings the use of interlayers could make them a viable option for joint implants as observed from the low wear rates ( Table 1 in supplementary information section) [38,129,147].

Silicon nitride
Bulk silicon nitride, Si 3 N 4 , and related materials have been used in applications such as combustion engines due to high temperature and Table 1 The important properties of coatings for joint implants, their evaluation methods and their target profile.

Property
Target profile Typical method for evaluation and standards Hardness A high hardness will help mitigate wear. However, the stiffness should also be considered as the ratio of hardness to Young's modulus gives a measure of the elastic limit in the contact and hence provides an indicator of wear performance.
Nano-or microindentation. Values are given in Pa or Vickers hardness number (HV). ASTM E2546-07 [58] ISO 14577-4 [59] Wear resistance A low wear rate is desirable, but actual values will depend on the specific tribological situation and are therefore not specified here. Attention should be paid to the generated wear debris, the size, shape and volume will likely influence the immune response in the final application.
Tribological set-up ranging for pin-on-disc to joint simulators. The resulting wear is measured as specific wear rate (mm 3 Table 2 Revised implants and follow-up studies of coated implants. In the case of revision-retrieved samples one must be aware of the fact that they are failures and may depart from the general performance of the cohort. wear resistance [148], and components made of Si 3 N 4 have been used in several applications subjected to wear, e.g. bearings [149][150][151][152]. It is also used in biomedical applications such as spinal implants [153]. The main advantage of silicon nitride, SiN x , as a coating for joint implants has been shown to be, not only its ability to reduce wear, but also to minimize the adverse immune response to released ions and particles. SiN x dissolves in aqueous solutions into only biocompatible elements [154], which could mean that the generated wear particles would dissolve in the body without triggering the immune system response that would eventually lead to bone resorption. It should be noted however that ammonia is formed during the dissolution of SiN x , which may result in an elevated pH. However, this has been found to be beneficial in terms of antibacterial properties [155,156]. Further, there needs to be a compromise between the dissolution of the wear particles and the requirement of a coating that provides sufficient performance for the intended period of use i.e. significantly greater than 20-25 years. SiN x coatings have been manufactured both through PVD and CVD methods with a wide range of hardness (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24) [47,49,50,53] and Young's moduli   [47,49,50]. Specific wear rate measurements through pin-on-disc setups have found reduced wear compared to CoCr [49,53]. The generated wear debris was investigated in a study where a ball-on-disc set-up was used to produce the same, which was found to lie in the range of 0.01-0.05 μm [157]. It was noticed that the debris agglomerated and that these agglomerates were in the range of 0.15-1.96 μm, while the individual particles measured between 0.01 and 0.05 μm in size. It was also found that the pH increased from 7.45 to around 8 after 20 days, likely due to the formation of ammonia [158,159], as mentioned above.
Biocompatibility studies have mainly been focused on simulated wear particles, often commercially available alternatives. The particles, both micron-scale and nanoscale, did not give rise to any significant release of proinflammatory cytokines in a study with primary human peripheral blood mononuclear cells [160]. Bulk Si 3 N 4 has also been found to show no cytotoxic effects [161,162] but instead antiviral properties [163]. The aforementioned dissolution behaviour of the coatings was investigated in simulated body fluid (25 vol% foetal bovine serum  reveal the presence of a metal carbide, which could lead to delamination when exposed to the environment of the body [139].
Reprinted by permission from Acta Materialia, Elsevier.
diluted in phosphate buffer saline solution) for up to 60 days (Fig. 2). The coatings successfully reduced the metal ion release by two orders of magnitude, as compared to uncoated CoCr references. The dissolution rates of the coatings were lower or comparable to the CoCr (0.2-1.2 nm/ day for SiN x coatings compared to 0.7-1.2 nm/day for CoCr) [48]. It should be noted however, that the dissolution rate will depend on factors such as composition and density of the coating -an increased nitrogen content has e.g. found to yield lower dissolution rates [48]. Recent studies have also found that alloying the SiN coating with Fe and C or Cr and Nb could give reduced dissolution rates [164].

Chromium nitride
Similar to other ceramic coatings CrN coatings have been shown to increase hardness and reduce wear compared to CoCr [44,165], as well as reducing the release of metal ions. The added advantage of CrN coatings is the possibility to achieve a CrN surface layer through e.g. plasma nitriding. Because of its potential for strong adhesion CrN is also sometimes used as an interlayer between the substrate and a top coating [46,52].
In a study comparing the performance of CrN with TiN, TiAlN and DLC, CrN was found to exhibit superior corrosion and wear resistance [131]. CrN coatings have been subject to additional investigation whereby coated CoCr femoral heads were paired with PE cups and tested in a hip simulator for 5 million cycles. The CrN coatings were shown to produce similar amounts of PE wear under standard conditions compared to adverse conditions (9.5 mm 3 /mc and 12 mm 3 /mc for standard and adverse conditionsfor the latter alumina particles were introduced to simulate third body abrasive wear), while the uncoated CoCr head showed a large increase in wear for the adverse conditions (9.2 mm 3 /mc and 469 mm 3 /mc under standard and adverse conditions, respectively) [166]. Another study comparing TiN, CrN, CrCN and DLC coatings in a hip simulator showed a 36-fold reduction in wear rate for self-mating CrN and CrCN coatings compared with uncoated CoCrMo MoM implants. In addition, the ion release was dramatically reduced [165].
Another method of creating a CrN rich surface is through incorporation of nitrogen into the surface of CoCr using reactive plasma. This was reported by Wang et al. [136], who exposed CoCr substrates to a plasma of NH 3 at a constant pressure of 500 Pa for 9 h. The formation of a CrN and Cr 2 N layer was confirmed by X-ray diffraction (XRD), and evaluation showed that the nitrided surface was harder and had lower wear rates compared to the untreated surface when in contact with a cemented carbide (WC/Co) ball during ball-on-disc wear tests. By using different plasma gases it is possible to obtain both nitrided and e.g. carbonitrided surface layers [135]. This was investigated by Liu et al. [135], who found both plasma nitriding and plasma carbonitriding of CoCr increased the hardness and wear resistance as compared to untreated CoCr. In addition, corrosion resistance was improved for both treatments, and the carbonitrided surface showed a better corrosion resistance compared to the nitrided surface.
Reported wear rate values of CrN based coatings have been consistently low (110 À6 to 8.810 À7 mm 3 /Nm) [44,167] and adhesion tests have given high critical load values (L C2 ) -up to 50 N for CrN/NbN deposited on CoCr substrates [44,45]. Overall the low wear rates and contingency of good adhesion make CrN-based coatings potentially suitable for joint implants.

Zirconia and yttria-stabilized zirconia
Zirconia and yttria-stabilized zirconia have been investigated as potential candidate coatings for orthopaedic implants because of their ability to reduce wear. The number of studies available on these materials for this application are still limited but the results are comparable to other investigated coatings [168,169]. These coatings were deposited onto a substrate such as titanium and should not be confused with the surface treated Oxinium implants.
The wear properties of yttria-stabilized zirconium dioxide (YSZ)coated titanium balls paired with UHMWPE have been evaluated in a ball-on-disc set-up using different lubricants. The coatings were found to decrease the wear of the PE under dry conditions and when lubricated with a NaCl solution. However, when lubricated with a serum solution i.e. under conditions more closely emulating those of a natural joint, the results were similar to those obtained for the uncoated Ti spheres [169] (3-1310 À4 mm 3 /Nm for dry conditions and 3-810 À4 mm 3 /Nm for serum solution as a lubricant; the root mean square surface roughness of the coatings ranged from 28 AE 1 nm to 60 AE 5 nm). The coatings did not exhibit any cytotoxicity when tested with mesenchymal stem cells and pre-osteoblast cell lines [168], indicating that they are biocompatible and could be useful as coatings for orthopaedic implants.
One way of achieving a surface oxide layer is through thermal oxidation, this was conducted by Luo et al. who oxidized a ZrNb alloy. An increased treatment temperature (of 700 C compared to 500 C) improved both the hardness and wear resistance [170].
The hydrolytic long-term stability of zirconia-based material remains a concern however [171,172], and would have to be thoroughly investigated before being an option for biomedical implants.

Alloyed and structured titanium nitride and carbide
While titanium nitride has successfully made it to the market (Implantcast, Cellumed, Link medical technology and Link orthopaedics) there is still ongoing research aimed to improving these coatings. One such strategy is to alloy the TiN coating with one or more elements and another is using a multi-layer structure. These strategies have the potential to further reduce the wear and/or improve adhesion.
Although alloying TiN with elements such as aluminium has been found to increase the hardness and Young's modulus, and decrease the wear rate compared to Ti6Al4V, however it should again be noted that CoCr is more commonly used in articulating surfaces and should be the material used for comparison [54]. Other investigated alloying elements include niobium and carbon, which have been shown to be non-cytotoxic [39]. Titanium carbide has also been shown to reduce wear compared to uncoated Ti6Al4V when tested in a reciprocating pin-on-disc set-up [173].
Another strategy for the improvement of wear properties is to deposit multilayers. The mechanical and tribological properties of such a coating (TiN/CrN) has been investigated and compared to a TiN monolayer. The multilayer structures reduced the friction coefficient, however, the wear rate was not reported [42].
Another option is to create a ceramic surface layer, by e.g. heataccelerated diffusion. There are several techniques available for alloying the surface of a material by exposure to elevated temperatures in a controlled environment. Another example is the powder immersion reaction assisted coating (PIRAC). Yet another process whereby nitrogen is incorporated into the surface is laser nitriding, where the substrate is irradiated with a laser in a chamber with nitrogen gas. The laserilluminated area of the surface will melt and create a plasma above it. Subsequently the nitrogen, which is now ionized, will be absorbed by the melted surface. All methods result in increased hardness as well as increased wear resistance compared with Ti6Al4V. The PIRAC method and laser melting did however yield an increased surface roughness compared to an untreated reference [43,133,[174][175][176].
While TiN based coatings offer a possibility to improve the wear properties of Ti6Al4V, they are in many of the published current studies compared only to uncoated Ti6Al4V, which makes it challenging to assess the efficacy of these proposed coatings.

Tantalum carbide, tantalum oxide and tantalum nitride
Porous tantalum is used in orthopaedic applications such as cranioplasty plates and hip implant fixation due to its osseintegrative properties [177]. Proven to be non-cytotoxic, tantalum-rich coatings, such as TaN and TaO, have been proposed for application on the articulating surfaces of joint implants due to their corrosion resistance, promising results in vitro [177] and favourable mechanical properties.
Tantalum carbide coatings deposited on CoCr by thermal treatment in molten salts have been investigated with regards to their adhesion, mechanical and tribological properties [36,178]. The coatings were found to have a TaC or Ta 2 C-TaC structure depending on the carbon content, manufacturing process of the CoCr substrate and temperature during coating growth. In addition, the coatings were deposited in a multilayer structure with layers comprising different carbon contents. The thickness of the coatings varied depending on structure and ranged from 300 to 1000 nm. The structure of the coating proved important for adhesion, where a multilayer structure (TaC-Ta 2 C-Ta) yielded higher critical loads during scratch tests (delamination at 30 N compared with delamination at 14 N for a single layer TaC) [36]. When evaluating hardness and Young's modulus by nanoindentation (with a maximum load of 10 mN), the hardness was significantly increased (27 GPa for a multi-layer structure and 23 GPa for a single layer coating compared to 12 GPa for the uncoated substrate) [178]. The wear volume obtained in a pin-on-disc set-up with 25 vol% bovine serum diluted in distilled water as lubricant was similar for the different multi-layer structures. The wear was reduced compared to the uncoated substrate, however there were no discernible differences between coatings in terms of wear performance. Noteworthy is that during wear testing, third body abrasion was the most prominent wear mechanism due to pull-out of carbides that acted as third body abrasive particles.
Investigation of the corrosion and wear resistance of tantalum oxide (TaO 2 ) deposited onto Ti6Al4V has shown improved corrosion properties with an increased corrosion potential, reduced anodic current and reduced Ti ion release (I corr of 6.77010 À8 A/cm 2 for TaO compared to an I corr of 2.56010 À7 A/cm 2 for Ti6Al4V). The wear volume of the coated samples was reduced compared to Ti6Al4V (2.22 mm 3 /Nm compared to 7.78 mm 3 /Nm for uncoated Ti6Al4V) [179]. Again, a limitation of the study was the comparison only to Ti6Al4V. Another study where TaN was deposited on CoCr by RF sputtering the coatings were shown to have comparable or lower wear rates of a PE counter surface to uncoated CoCr [180].
In summary, tantalum carbide and oxide coatings have been shown to be biocompatible, but their wear performance in the application requires further investigations.

Alumina based coatings
In addition to the previously discussed coatings alumina based coatings have been proposed as an option. These coatings include monolithic micron alumina (IDA), micron alumina yttria-stabilized zirconia (YSZ) composite coating (IDAZ), and nanostructured alumina titania/YSZ (IDZAT) deposited on Ti-6Al-4V alloy and have been shown to have better wear and corrosion compared to Ti6Al4V [181].

Discussion
This review has provided a comprehensive overview of both commercially available coated implants, as well as coatings currently being researched for potential application in joint implants. The coated implants currently available in the market are ZrO 2 coated Zr (Smith and Nephew), TiN coated CoCr or Ti6Al4V (Implantcast), Link orthopaedic), TiNbN coated CoCr (OHST medical technology) and TiN coated Ti6Al4V (Endotec). Reported follow up studies as well as case studies of retrieved implants have revealed revision rates comparable to traditional CoCr implants [104,105,[124][125][126]. The potential coating materials currently being investigated include SiN x , CrN, TaO, TaC, DLC, TiN, TiCN, ZrO 2 and ZrN. Looking at total knee replacements the use of coatings is more widespread and the retrieval data reveals the coated implants to be comparable to CoCr but there is not enough evidence to support a lower amount of osteolysis-caused failures [80,113,[116][117][118][119][120][121][122][123][124][125].
When comparing the specific wear rates of coatings currently being researched (Fig. 3b) it is evident that all coating types (silicon nitride based, titanium based, chromium nitride based, tantalum based and DLC) have the potential for low wear rates. However, it is difficult to distinguish between them since i) different set-ups have been used to assess the wear performance, and ii) different counter surfaces have been used in the tests, i.e. hard-on-hard (e.g. alumina) or hard-on-soft (a polymer). The available standards cover procedures for both types of material combinations, including suitable testing conditions. However, even when the testing is performed in accordance with the standards it might not reflect the full range of loading scenarios the implant will be exposed to. To better predict clinical outcome the current standardized approaches require further development. Important considerations for such an approach are of course the choice of material combinations, i.e. hard-onhard versus hard-on-soft as well as the effect of activities other than steady state gait. Common activities such as standing up from a chair and climbing stairs put a higher demand on the implant and will affect the implant performance.
The previous introduction to and subsequent removal from the market of DLC coated implants illustrates the need for more rigorous testing before clinical use. DLC coated implants were removed from the market after the revelation that exposure to synovial fluid over time led to a delamination of the coatings due to the instability of metal carbides in the interlayer, which in turn caused early revisions. Scratch and Rockwell tests typically evaluate the adhesion of as-deposited coatings, but typically do not consider corrosion or changes to the coating over time. To better predict the adhesion of the coating, the scratch or Rockwell tests can e.g. be performed on coatings exposed to liquid at different soaking times [52]. In this review, DLC coatings were included as materials that are not currently in the market but show potential for implant applications. Having been introduced to the market and suffered from early implant revisions, DLC coatings are often not considered an option. However, the mechanisms causing the delamination have since been investigated and are now better understood. By using interlayers, DLC coatings could be stable long-term, reduce wear and potentially increase implant longevity.
In summary, several of the investigated coatings show potential because of their ability to reduce wear and ion release. Different coatings carry different advantages, e.g. CrN on CoCr and TiN on Ti6Al4V can give enhanced adhesion. One material that has the potential to provide additional biological advantages is silicon nitride, which has demonstrated antibacterial and antiviral properties.
Revision and follow up studies of commercially available coatings have found revision rates to be comparable to, but not better than, conventional CoCr implants. This could be due to low revision rates or not long enough follow-up studies (Table 2) [55,80,96,. The signs of wear and damage of the PE countersurface has been found to be lower for oxidized zirconium knee implants compared with CoCr [120], which is promising as it speaks to the potential for reduced generation of PE debris. Since this debris is believed to be a major cause of revision its reduction could lead to a longer implant lifetime. Assuming the coating adheres well to the substrate it is possible to have a biocompatible surface that generates low amounts of wear debris and ion release, which would be ideal for articulating surfaces.
Whilst much of the coating work has focussed on the application of coatings to bearing surfaces to reduce wear and corrosion, there is now also a growing interest in translating such technologies to other interfaces, such as hip modular-tapers, where fretting-corrosion processes may dominate. The ability to develop novel multi-material systems also paves the way for a new generation of multi-functional coating technologies to combat the aforementioned issues as well as the emerging grand challenges within the area of orthopaedics (e.g. infection and treatment of metastatic cancers).

Conclusions
In conclusion, Coated implants are available in the marked. These implants include TiN, TiNbN, ZrN coatings and surface treated Zr resulting in an oxidized surface layer. Several candidate coating materials such as carbon-based, silicon nitride, chromium nitride, Ti-based, Zr-based, Ta-based and aluminabased are being researched. Coated implants exhibit comparable survival rates to uncoated implants, however the basis for assessment is limited due to generally low revision rates and short follow-up times. Coatings could be relevant for other surfaces such as modular interfaces.

Data availability
The data in this review consists of information found in published research papers.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.  [36,45,[51][52][53][54]134] (a) and specific wear rates from ball/pin-on-disc evaluations [46,49,[51][52][53][54]129,131,173,179,183] (b) found for coatings currently being researched. The reported values are divided into groups based on the main constituents of the coating.