metals Review
Mg and Its Alloys for Biomedical Applications: Exploring Corrosion and Its Interplay with Mechanical Failure Mirco Peron, Jan Torgersen * and Filippo Berto Department of Mechanical and Industrial Engineering, Norwegian University of Science and Technology (NTNU Norway), 7491 Trondheim, Norway;
[email protected] (M.P.);
[email protected] (F.B.) * Correspondence:
[email protected]; Tel.: +47-93-966576 Received: 6 June 2017; Accepted: 25 June 2017; Published: 5 July 2017
Abstract: The future of biomaterial design will rely on temporary implant materials that degrade while tissues grow, releasing no toxic species during degradation and no residue after full regeneration of the targeted anatomic site. In this aspect, Mg and its alloys are receiving increasing attention because they allow both mechanical strength and biodegradability. Yet their use as biomedical implants is limited due to their poor corrosion resistance and the consequential mechanical integrity problems leading to corrosion assisted cracking. This review provides the reader with an overview of current biomaterials, their stringent mechanical and chemical requirements and the potential of Mg alloys to fulfil them. We provide insight into corrosion mechanisms of Mg and its alloys, the fundamentals and established models behind stress corrosion cracking and corrosion fatigue. We explain Mgs unique negative differential effect and approaches to describe it. Finally, we go into depth on corrosion improvements, reviewing literature on high purity Mg, on the effect of alloying elements and their tolerance levels, as well as research on surface treatments that allow to tune degradation kinetics. Bridging fundamentals aspects with current research activities in the field, this review intends to give a substantial overview for all interested readers; potential and current researchers and practitioners of the future not yet familiar with this promising material. Keywords: Mg alloys; biocompatible; corrosion; crack growth; corrosion assisted cracking
1. Introduction In the last decades, concurrent with the increased lifetime in today’s world population, the amount of people undergoing surgical procedures involving the implantation of medical devices is continuously growing [1]. These implants are generally used for applications that ensure a substantial improvement in patients’ quality of life, such as orthopaedics, pacemakers, cardiovascular stents, defibrillators, neural prosthetics and drug delivery systems [2–5]. Among these, orthopaedic surgery is the most important [4], characterized by the highest annual growth rate. Indeed, according to Long and Rach [6], almost 90% of the population over 40 years is affected by degenerative joint diseases. Total hip replacements are predicted to represent half of the estimated total number of operations in 2030 [7]. For this reason, the research has centred on new biocompatible materials and on improving their performances [8–10]. Especially strength and corrosion resistance in human body fluid (HBF) has been a major issue of research [11]. Tailoring biomaterials to their specific use, many important properties must be considered. First, they must be bioinert or bioactive [12]; the former means that the organism has a coexistence with the material without noticeable change in the organism and/or exogenous material, whereas the latter implies that there is an interaction with and/or a response from living tissue, e.g., the formation of direct biochemical bonds with the surface of the material in bioceramics. Further, they must not Metals 2017, 7, 252; doi:10.3390/met7070252
www.mdpi.com/journal/metals
Metals 2017, 7, 252
2 of 41
evoke a sustained inflammatory response once implanted in the human body, have appropriate mechanical properties for their use and include appropriate permeability and processability for designed applications [13]. Several materials provide these properties and they can be divided into four main classes [4,14]: metals [15], polymers [16,17], ceramics [12,18] and composite materials [14,19]. It is worth to note that implant design, and therefore material choice, depends on the application of the devices and on the environment in which they are employed (Table 1). Table 1. Biomedical applications of biocompatible materials [4,12,14,15,17,19–21]. Application
Material
Spinal fixation
stainless steel; Ti, Ti alloys
Bone fixation (bone plate, screw, wire, bone nail, etc.)
stainless steel; Ti, Ti alloys; PMMA; PS; CF/PEEK; CF/Epoxy; CF/PS; Bio-Glass/PU; Bio-Glass/PS; PET/SR; PET/Hydrogel; CF/LCP; GF/PEEK; Bone particles/PMMA; Ti/PMMA; UHMWPE/PMMA; UHMWPE; GF/PMMA; CF/PMMA; KF/PMMA; PMMA/PMMA; Bio-Glass/Bis-GMA; CF/PP; CF/PS; CF/PLLA; CF/PLA; KF/PC; HA/PE; PLLA/PLDLA; PGA/PGA; Alumina; Zirconia; Tricalcium phosphate; Bio-Glass-Metal fiber composite; Bio-Glass
Artificial joint
Co-Cr-Mo alloy; Ti alloys; Alumina; Zirconia; PET/SR;CF/UHMWPE; PET/Hydrogel; CF/Epoxy; CF/C; CF/PS; CF/PEEK; CF/PTFE; CF/UHMWPE; CF/PE; UHMWPE/UHMWPE; HA/HDPE; Metal Bio-Glass coatings
Spinal spacer
stainless steel; Ti alloys
Dental implant and bridge
Ti; Ti alloys; Au; CF/C; SiC/C; Alumina; Bio-Glass; HA; Bio-Glass coated alumina; Bio-Glass coated vitallium; Au-Cu-Ag; Au-Cu-Ag-Pt-Pd; UHMWPE/PMMA; CF/PMMA; GF/PMMA; KF/PMMA;
Tendon/Ligament/Cartilage Replacement
PET/PU; PTFE/PU; CF/PTFE; CF/C; PET/PHEMA; KF/PMA; KF/PE; CF/PTFE; CF/PLLA; GF/PU; PP; ePTFE; PET/PET; PA; PU; PLGA
Despite the recent technological evolution of polymeric and ceramic materials, metals, such as stainless steel, cobalt-chromium , and titanium alloys, are still the most used materials in biomedical applications, covering 70% of the total production volume [15,22,23]. What makes these materials irreplaceable are their high: (1) (2) (3) (4) (5) (6)
Strength Fracture toughness Ductility Fatigue life Wear resistance and Corrosion resistance
It is widely reported [14,24] that bones are subjected to a stress of about 4 MPa on a daily basis and hip joints may experience a load as high as 10 times of the body weight in some cases, these implantable devices require high properties in terms of (1) and (2). Moreover, a high percentage of elongation at fracture (3) is needed in order to prevent brittle failure leading to the predominant choice of metallic implants as bioimplant materials. The human body also experiences dynamic loads: Ramakrishna et al. [14] reported that hip joints are subjected to loads applied for 1 × 106 cycles per year making (4) a requirement. This kind of joints are also affected by sliding issues potentially leading to the formation of wear debris if (5) is not high enough. This again causes acute immune response resulting in inflammation and/or fibrosis [4]. Finally, implanting materials in the human body necessitates great biocompatibility implying that they do not release any toxic substances into the body (6). However, since the mechanical properties of metals highly differ from those of human bone, they are not always suitable for medical devices (see Table 2): the bigger the difference in elastic modulus, the higher the occurrence of stress-shielding, which is a phenomenological consequence of
Metals 2017, 7, 252
3 of 41
Metals 2017, 7, 252 Metals 2017, 7, 252
3 of 40 3 of 40
stress distribution changes [25–33]. Bones accordingtotothe theWolff’s Wolff’s stress distribution changes [25–33]. Bonesadapt adapttotothe thechanged changed stress stress field field according stress distribution changes [25–33]. Bones adaptmore to theporous changed stress field accordingortothinner the Wolff’s law [34], resulting in the bone either becoming (internal remodelling) (external law [34], resulting in the bone either becoming more porous (internal remodelling) or thinner law [34], resulting in to thea higher bone either becoming more porous (internala hip remodelling) or thinnerupon remodelling), leading implant failure. Considering joint, foraexample, (external remodelling), leadingrisk to aofhigher risk of implant failure. Considering hip joint, for (external remodelling), leading to a higher risk of implant failure. Considering a hip joint, for the example, implantation of a respective prosthesis, the entire bone-biomaterial system can be abstracted upon the implantation of a respective prosthesis, the entire bone-biomaterial system can beas example, upon the implantation of a respective prosthesis, the entire bone-biomaterial system can be twoabstracted springs inasparallel (Figure 1). two springs in parallel (Figure 1). abstracted as two springs in parallel (Figure 1).
(a)
(a)
(b)
(b)
Figure 1. (a) Total hip joint representation and (b) schematic view of two springs in parallel, where k1 Figure 1. (a) (b)(b) schematic view of two springs in parallel, where k1 k1 Figure (a) Total Total hip hipjoint jointrepresentation representationand and schematic view of two springs in parallel, where and k2 are implant and bone stiffness, respectively. (From [35]). and k2 k2 are implant implant and [35]). and andbone bonestiffness, stiffness,respectively. respectively.(From (From [35]).
The load redistribution phenomenon is strictly related to the stiffness ratio according to: The load redistribution phenomenon is strictly related to the stiffness ratio according to: The load redistribution phenomenon is strictly related to the stiffness ratio according to: σ E1 σ1 1E= , (1) 1 = , σ σ 2 EE12 (1) σ21 = E2 , (1) σ2 E2 where σ and E is applied stress and material stiffness, respectively. Subscripts 1 and 2 refer to implant where σ and E is applied stress and material stiffness, respectively. Subscripts 1 and 2 refer to implant andσbone, As theand load on the bone decreases with the Subscripts implant, the1 bone become less where and Erespectively. is applied stress material stiffness, respectively. and 2will refer to implant and bone, respectively. As the load on the bone decreases with the implant, the bone will become less dense and thinner due to the insufficient stimuli for continued remodelling (Figure 2). and bone, Asthe theinsufficient load on the bone for decreases with the implant, the bone dense andrespectively. thinner due to stimuli continued remodelling (Figure 2). will become less dense and thinner due to the insufficient stimuli for continued remodelling (Figure 2).
Figure 2. Stress-shielding example in shoulder replacement surgery; (a) preoperative image; (b) image Figure 2. 2. Stress-shielding example shoulder replacement surgery; (a) preoperative image; (b) image Figure Stress-shielding examplein shoulder (a) preoperative image; (b) image immediately after the surgery; (c)inimage afterreplacement 7 years fromsurgery; the surgery; arrow indicates the region of immediately after the surgery; (c) image after 7 years from the surgery; arrow indicates the region of immediately after the surgery; (c) image after 7 years fromElsevier, the surgery; bone resorption. Reproduced with permission from [36], 2003. arrow indicates the region of bone resorption. Reproduced with permission from [36], Elsevier, 2003. bone resorption. Reproduced with permission from [36], Elsevier, 2003.
Metals 2017, 7, 252
4 of 41
Stress induced bone remodelling is widely described in literature [31,32,37] and several models have been proposed to explain this phenomenon. Van Rietbergen et al. [33] stated that the bone mass is regulated by the remodelling signal S, which is the strain energy (U) per unit of mass: S=
U , ρ
(2)
where ρ is the bone density. Bone remodelling as described in Equation (3) corresponds to the rate of net bone-mass turnover dM dt and happens when S differs from the characteristic reference value Sref describing the remodelling signal in non-operated conditions: dM = S − Sref dt
(3)
The theory further assumes that no remodeling takes place when S − Sref does not exceed a threshold. Huiskes et al. [38] agreed that local strain perturbation governs the remodelling process by means of an interaction with osteocytes, since they are widely reported to be transductor of mechanical signals [39–42]. They suggested that increased (or decreased) strain signals the osteocyte to transmit stimuli to the bone surface, where bone is formed (or removed) until the strain normalizes. Table 2. Comparison of the mechanical properties of natural bone with various implant materials [11,15,19,43–46]. Properties
Natural Bone
Stainless Steel
Ti Alloy
Co-Cr Alloy
Magnesium
Density (g/cm3 ) Elastic modulus (MPa) Tensile strength (MPa) Compressive yield strength (MPa) Elongation at failure (%) Fracture toughness (MPa m1/2 )
1.7–2.0 3–20 80–150 130–180 1–7 3–6
7.9–8.1 189–205 480–620 170–310 30–40 50–200
4.4–4.5 110–117 930–1140 758–1117 8–15 55–115
8.3–9.2 230 900–1540 450–1000 30–45 100
1.74–2.0 41–45 170–270 65–100 6–20 15–40
In certain biomedical applications, especially fixators such as bone plates, screws, pins and stents, biomedical devices are required to stay inside the human body only for a restricted period, i.e., as long as bones heal. Thus, materials that ideally degrade in the same manner and speed as natural bone heals [10,47,48] are widely required to be utilized for these temporary devices. Biodegradable materials allow circumventing cumbersome second surgeries involving the removal of the old implant. Besides complications and distress for the patient, this also increases insurance costs [49–55]. However, classic metallic materials as those above are not biodegradable and, if not removed, lead to long-term complications, as local inflammations due to the potential release of cytotoxic ions as a consequence of corrosion or wear processes [22,23,56–58]. It is therefore very attractive to identify materials that degrade in the same way as bone heals, dissolving entirely after complete healing. In this process, these materials must not cause any toxic, allergic, inflammatory or cancerous effects. In recent years, researchers and clinicians have employed a variety of compositions, properties and forms in which polymers are available to best match the specifications of the materials’ desired biomedical function [4,13,16,20,59]. However, they tend to be too flexible and too weak for load-bearing applications required in orthopaedic surgery. Moreover, they may also absorb liquids and swell, leach undesirable products such as monomers, fillers and antioxidants, and, furthermore, the sterilization process may affect their properties [14]. Thus, to overcome this drawback of polymeric materials, researchers’ and clinicians’ attention focused on metallic biodegradable materials. Several of them have been studied but most of the scientific efforts focus on Magnesium (Mg) and its alloys. Among metallic engineering materials, Mg possesses one of the best bio-compatibilities with human physiology and the best mechanical compatibility with human bone [60]. Its low density and elastic modulus best mimic the properties of
Metals 2017, 7, 252
5 of 41
natural bones (Table 2). Moreover, Mg is the fourth most abundant element in the human body (it is recommended that an adult receives 240–420 mg daily) and it is essential for the metabolism in many biological mechanisms (Mg is a cofactor for many enzymes [46]). Finally, Mg2+ ions, resulting from the degradation process (see Section 2), are reported to aid the healing and growth of tissue. Any excess of these ions is harmlessly excreted in the urine [43]. Despite the advantageous properties of bioactive Mg and its alloys, there are some limitations for their use in temporary implants. The major drawback is the high corrosion rate in the physiological environment that may lead to a loss of mechanical integrity before tissues have sufficient time to heal. Moreover, hydrogen as a corrosion product together with the generation of respective hydrogen pockets can influence the healing process or, if the pockets are large, they may cause death of patients through blocking of the blood stream. Finally, the simultaneous action of the corrosive human-body-fluid and the mechanical loading can cause the further complication of sudden fracture of implants due to corrosion-assisted cracking, such as stress corrosion cracking (SCC) and corrosion fatigue (CF) [61–63]. The aim of this work is to describe the state of the art regarding corrosion behaviour assessment and the developed methods for improving the corrosion resistance of Mg. First a short summary of the degradation mechanism of magnesium in human body fluid will be provided, and then a focus on the corrosion-assisted cracking phenomena (SCC and CF). Finally, an overview of how to improve corrosion properties will be given. 2. Corrosion The main drawback in the use of magnesium in orthopaedic implants is its high corrosion rate in the electrolytic physiological environment. Mg and its alloys are reactive metals and they corrode in aqueous environments according to the reactions given below [64,65]: Mg + 2H2 O → Mg(OH)2 + H2 ,
(4)
The corrosion reaction can be divided into the anodic and cathodic partial reaction, (5) and (6), respectively [66]: Mg → Mg2+ + 2e− , (5) 2H2 O + 2e− → H2 + 2(OH− ),
(6)
According to Equation (4), a film of Mg(OH)2 forms on the surface of Mg and its alloys, preventing further corrosion in water since it is slightly soluble there. Mg2+ + 2(OH− ) → Mg(OH)2 ,
(7)
However, if the corrosive medium contains any chlorides with concentration above 30 mmol/L [67], the magnesium hydroxide will be converted to magnesium chloride MgCl2 according to [43]: Mg(OH)2 + 2Cl− → MgCl2 + 2(OH− )
(8)
Magnesium chloride is highly soluble in aqueous solution, determining the further corrosion development of Mg and its alloys in the human body, where the chloride content is about 150 mmol/L [68–71]. It is worth to note that the chemistry involved in the corrosion, i.e., magnesium chloride solubility and hydrogen evolution, highly influences the in vivo performances of magnesium devices. Starting from the solubility of MgCl2 film, it is reported that it can affect the mechanical behaviour in two different ways. According to Ghali et al. [72], two different corrosion modes are possible, uniform and localized. Mg alloys with slow and constant degradation rate (lower than 0.5 mm/year [65]) are required to facilitate sufficient time to heal for the bone. Further, a uniform degradation mode is key as localized corrosion can lead to toxicity and early failure of the implant. However, uniform
Metals 2017, 7, 252
6 of 41
corrosion is not the main corrosion mode. Kirkland et al. [73] reported that 29 of 31 Mg alloys suffer of localized corrosion, i.e., pitting. The difference discriminating uniform from localized corrosion is mainly linked to the presence of second phases, precipitates or impurities on the surface of the Mg alloys, inhomogeneities generally widely present in Mg and its alloys. The most common impurities are Fe, Ni and Cu. They are most harmful in terms of corrosion resistance due to their very negative corrosion potential [65,74–76]. They are cathodic with respect to the Mg matrix, resulting in a faster corrosion rate and determining a higher and more severe susceptibility to pitting that quickly destroys the Mg(OH)2 protective film [60]. Once this film is removed, the surrounding corrosive fluid will have contact with the Mg matrix, causing further corrosion according to the just mentioned mechanism. If the inhomogeneities are not uniformly distributed [65] or above a tolerance limit [77], the material will corrode in a localized manner, acting as precursor for stress corrosion cracking (SCC) and corrosion fatigue (CF). These mechanisms highly affect the corrosion resistance of Mg and its alloys, hampering their application for biomedical implants and they will be described in detail in the next section. Hydrogen evolution due to the corrosion mechanism involved in Equation (4) influences the in vivo performances of Mg alloys in two ways [22,78–82]. First, when the corrosion rate is too high, human body cannot absorb the entire amount of developed hydrogen, thus generating gas pockets and bubbles. Though Witte et al. [70] found 0.068 mL/cm2 /day leading to no subcutaneous bubbles after 2–3 weeks on guinea pig, and even if Song [79] set 0.01 mL/cm2 /day as tolerated value in his work, no scientific evidence has been reported that these values can be used as a threshold for acceptable hydrogen amounts in the human body. In orthopaedic applications, where blood transport mechanisms are poor, these gas pockets are harmful since they occur around the device, causing risk of implant failure, tissue separation or, in the worst case, death of the patient due to blood clotting [79–82]. Second, hydrogen evolution is strictly connected to the corrosion-assisted cracking, i.e., SCC and CF, since their propagation is highly influenced by the hydrogen embrittlement (HE) phenomenon [22,78], which will be discussed in the next section. 3. Corrosion Assisted Cracking Phenomena From the middle of the 20th century, several metallic materials have been reported susceptible to the synergistic effect of mechanical loads and corrosive medium, leading to corrosion assisted cracking phenomena, i.e., SCC and CF [61–63]. These mechanisms result from a simultaneous combination of three factors, susceptible material, corrosive environment and tensile stress. Mg and its alloys have gained interest as materials for biomedical devices due to their degradation properties upon exposure to the human body fluid, and hence researchers and clinicians have started to investigate the susceptibility of these materials to corrosion assisted cracking mechanisms. This involves the consideration of whether the mechanical property reduction results from the combination of applied stresses and corrosive environment, i.e., SCC, or simply is a consequence of the reduction of implant cross-section because of corrosion mechanisms. Choudhary and Raman [83] performed slow strain rate tensile (SSRT) tests on AZ91D alloy in modified simulated body fluid (m-SBF) at different conditions: (a) strained in air; (b) strained in m-SBF; (c) strained in air after a pre-immersion in m-SBF solution for a time as long as the time to failure of case (a); (d) continuously cathodically charged and simultaneously pulled in m-SBF. Comparing the results (Figure 3), they found a considerable reduction in mechanical properties of specimens stressed in m-SBF, leading to the conclusion that the synergistic effect of corrosive environment and mechanical loads, rather than the corrosive environment itself, mostly affects the corrosion resistance of Mg and its alloys.
Metals 2017, 7, 7, 252 Metals 2017, 252
7 of 4041 7 of
Figure 3. Stress-time plot of AZ91D alloy tested ate strain rate of 2.2 × 10−7 s−1 under different Figure 3. Stress-time plot of AZ91D alloy tested ate strain rate of 2.2 × 10−7 s−1 under different environmental conditions. Reproduced with permission from [83], Elsevier, 2012. environmental conditions. Reproduced with permission from [83], Elsevier, 2012.
Since orthopaedic devices are subjected different acute loadings during their use, i.e., tension, Since orthopaedic devices are subjected toto different acute loadings during their use, i.e., tension, compression, bending torsion depending on human activities and skeletal location [78] and compression, bending andand torsion depending on human activities and skeletal location [78] and since since these loadings are repetitive (just thinking to the daily activities of walking), SCC and CF are these loadings are repetitive (just thinking to the daily activities of walking), SCC and CF are of great of great the and use of and alloys for as material fordevices. biomedical devices. concern forconcern the usefor of Mg its Mg alloys asits material biomedical Moreover, MgMoreover, and its Mg and its alloys suffer pitting in human body fluid, due to the presence of chlorides and alloys suffer pitting in human body fluid, due to the presence of chlorides and impurities impurities [43,60,67,68,77,78,84–88], making these phenomena main limitations. Chemical induced [43,60,67,68,77,78,84–88], making these phenomena main limitations. Chemical induced pits and pits andsharp implants’ sharp contours required stress concentration forsuch the phenomena, onset of such implants’ contours provide the provide requiredthe stress concentration for the onset of phenomena, which are described Their most detrimental effect is implant failure at which are described below [60,89].below Their[60,89]. most detrimental effect is implant failure at stresses stresses considerably below yield and design stresses at mechanical conditions otherwise considerably below the yieldtheand design stresses [83], [83], i.e., i.e., at mechanical conditions otherwise consideredtotobebesafe safe[75]. [75].Such Such sudden failure produces serious issues thepatient, patient,such suchasasthe the considered sudden failure produces serious issues totothe need of a complicated and harmful removal of the failed implant and subsequential inflammatory need of a complicated and harmful removal of the failed implant and subsequential inflammatory responses,asaswell wellasasthe thenecessity necessityofofa anew newdevice deviceimplantation. implantation. responses, This section provides reader with a description of these phenomena. It should be noted This section provides thethe reader with a description of these phenomena. It should be noted that that the mechanistic understanding SCC CF derives mostly from derives from knowledge the mechanistic understanding of SCC of and CF and mostly knowledge obtainedobtained on Mg-Alon Mg-AlAs alloys. Al to is toxic to the human body, the mechanisms be entirely representative alloys. Al is As toxic the human body, the mechanisms mightmight not benot entirely representative of of biocompatible alloys, but it still provides a meaningful framework to describe the theory (see biocompatible alloys, but it still provides a meaningful framework to describe the theory (see Section Section 5.2.1). 5.2.1). 3.1. Stress Corrosion Cracking (SCC) 3.1. Stress Corrosion Cracking (SCC) Whenstatic statictensile tensilestresses stresses and and corrosive onon susceptible materials, they When corrosive environments environmentsact acttogether together susceptible materials, leadlead to the This mechanism is characterized by theby onset propagation of cracks, they to SCC. the SCC. This mechanism is characterized the and onset and propagation of resulting cracks, in the embrittlement of ductile Mg in the corrosive medium leading to a considerable elongation resulting in the embrittlement of ductile Mg in the corrosive medium leading to a considerable reductionreduction [60,83]. [60,83]. elongation The onset SCC requires a localized corrosion where the conditions for crack initiation The onset of of SCC requires a localized corrosion site,site, where the conditions for crack initiation are are met. It is widely reported that these cracks initiate at high stress locations, such as roots of a met. It is widely reported that these cracks initiate at high stress locations, such as roots of a corrosion pit or pre-existing flaws (surfaces defects) [60,75,78,83,90]. Localizedsites corrosion pitcorrosion or pre-existing microscopicmicroscopic flaws (surfaces defects) [60,75,78,83,90]. Localized corrosion take sites take place after the fracture of the magnesium hydroxide protective film as a consequence of the place after the fracture of the magnesium hydroxide protective film as a consequence of the chemical chemical reaction with chloride ions according to Equation (8). This proceeds locally because second reaction with chloride ions according to Equation (8). This proceeds locally because second phases phases and/or impurities the alloy underneath affect thefilm protective filmMoreover, resistance. and/or impurities of the alloyofunderneath affect the protective resistance. it isMoreover, widely it is widely reported [76] that the breakdown of the protective film also follows the applied stresses. reported [76] that the breakdown of the protective film also follows the applied stresses. Failure can derive from stress enhancing elements such as inclusions or notches in the implant geometries
Metals 2017, 7, 252
8 of 41
Failure can derive from stress enhancing elements such as inclusions or notches in the implant geometries enhancing stresses from daily activities or residual ones introduced during Metals 2017, 7, 252 8 of 40 device fabrication. Moreover, the presence of dynamic loading, i.e., CF, may lead to the formation of fatigue enhancing stressesa preferential from daily activities or residual ones introduced during device fabrication. cracking providing site for SCC [60,75,91]. Moreover, the presence of dynamic loading, CF, propagate may lead tountil the formation of fatiguethe cracking Once localized corrosion takes place, thei.e., cracks failure. However, propagation providing a preferential site for SCC [60,75,91]. phase is not well understood and literature distinguishes between several mechanisms. In their review, Once localized corrosion takes place, the cracks propagate until failure. However, the Winzer et al. [75] divided them into two groups of mechanisms:
propagation phase is not well understood and literature distinguishes between several mechanisms. their review, Winzer etmechanisms al. [75] divided them into two groups mechanisms: 1. In Anodic dissolution categorized in the: (a) of galvanic attack by film rupture model;
tunnelling model;mechanisms (c) preferential attack in model. 1. (b)Anodic dissolution categorized the: (a) galvanic attack by film rupture model; (b) tunnelling model; (c) preferential attack model. 2. Mechanical fracture mechanisms or cleavage-like fracture as described in the hydrogen 2. embrittlement Mechanical fracture model. mechanisms or cleavage-like fracture as described in the hydrogen embrittlement model The development of 1a started from the observation that SCC propagation in AZ31 alloys is The development of 1a started from the observation that SCC propagation in AZ31 alloys is halted once a cathodic applied.Logan Logan [92,93] proposed an electrochemical halted once a cathodiccurrent current is is applied. [92,93] proposed an electrochemical modelmodel for the for the crack propagation phase under corrosive media. He theorized that once an applied strain is sufficient crack propagation phase under corrosive media. He theorized that once an applied strain is sufficient to produce a breakdown of the Mg(OH) protective film, an electrochemical cell occurs between the to produce a breakdown of the Mg(OH)2 2protective film, an electrochemical cell occurs between the anodic film-free surfaceand and the the cathodic cathodic protective layer (Figure 4). 4). anodic film-free surface protective layer (Figure
Figure Continuous crack crack propagation to to thethe filmfilm rupture model. Reproduced with with Figure 4. 4.Continuous propagationaccording according rupture model. Reproduced permission from [75], John Wiley and Sons, 2005. permission from [75], John Wiley and Sons, 2005.
Moreover, he suggested that this electrochemical cell determines a localized corrosion that, Moreover, he suggested that this electrochemical cellThus determines a localized corrosion that, acting acting as a notch, leads to a region of stress concentration. the protective film does not reform, as aallowing notch, leads to a region ofpropagation. stress concentration. protective notforreform, allowing a continuous crack However,Thus somethe authors refutedfilm thisdoes model different reasons. Ebtehaj et propagation. al. [94] as well asHowever, Wearmouth et al.authors [95], for example, found a discontinuous crackreasons. a continuous crack some refuted this model for different growth taking by as anWearmouth alternating sequence of for filmexample, rupture and repassivation. A continuous Ebtehaj et al. [94]place as well et al. [95], found a discontinuous crack growth propagation mode is possible only when repassivation is limited. Logan proposed this taking place by an alternating sequence of film rupture and repassivation. A continuous propagation electrochemical model since he observed an improvement of corrosion resistance applying a cathodic mode is possible only when repassivation is limited. Logan proposed this electrochemical model current. However, he neglected that also hydrogen embrittlement models (which will be described since he observed an improvement of corrosion resistance applying a cathodic current. However, later in this section) require a film-free surface that can be influenced by a cathodic charge. he neglected that hydrogen models (which will be described later inA/cm this section) −6 m/s and 2 Furthermore, he also observed a crackembrittlement growth rate of 10 calculated a current density of 14 require a film-free surface can bePugh influenced a cathodic he observed needed to ensure such crackthat velocities. et al. [96]by considered thischarge. value tooFurthermore, high for a dissolution −6 m/s and calculated a current density of 14 A/cm2 needed to ensure a crack growth rate of et10al. model. Thus Winzer [75] considered the observations made by Logan being more closely such crack velocities. Pughthe et preferential al. [96] considered this value high for later. a dissolution model. described by another model, attack model, which willtoo be described The tunnelling model was proposed by Pickering and Swan [97] and is characterized by described the Thus Winzer et al. [75] considered the observations made by Logan being more closely formationmodel, of tubular due to theattack rupture of the which surface will film be at emerging steps (Figure 5a). At by another the pits preferential model, describedslip later. the beginning, the direction of these pits is considered to be governed by the potential by the The tunnelling model was proposed by Pickering and Swan [97]electrochemical and is characterized difference between the Mg matrix and the second phase that they assumed to be Mg17Al12. After the
formation of tubular pits due to the rupture of the surface film at emerging slip steps (Figure 5a). At the
Metals 2017, 7, 252
9 of 41
beginning, the direction of these pits is considered to be governed by the electrochemical potential Metals 2017, 7, 252 9 of 40 Metals 2017, 7, 252 the Mg matrix and the second phase that they assumed to be Mg Al 9. of 40 difference between 17 12 After the establishment of a of locallocal galvanic cellcell able to to prevent re-passivation, crack propagation governed by establishment galvanic able prevent re-passivation,crack crackpropagation propagationisisis governed establishment of aalocal galvanic cell able to prevent re-passivation, governed a ductile tearing of metal between narrow tunnels (Figure 5b). by a ductile tearing of metal between narrow tunnels (Figure 5b). by a ductile tearing of metal between narrow tunnels (Figure 5b).
Figure (a)Tubular Tubular pitsform form due film rupture onslip slipsteps; steps; (b)crack crack propagation occurs by Figure 5. (a) pitspits form duedue to film rupture on on slip steps; (b)(b) crack propagation occurs byby ductile Figure 5.5.Tubular (a) totofilm rupture propagation occurs ductiletearing tearingofofremaining remainingligaments. ligaments.Reproduced Reproducedwith withpermission permissionfrom from[75], [75],John JohnWiley Wileyand andSons, Sons, ductile tearing of remaining ligaments. Reproduced with permission from [75], John Wiley and Sons, 2005. 2005. 2005.
However, Winzer et et al. [75] refuted model observing that Pickering Swan proposed it However, Winzer etal. al. [75] refutedthis thismodel modelobserving observingthat that Pickering andand Swan proposed However, Winzer [75] refuted this Pickering and Swan proposed itit foraamaterial material (Mg-1Al) completely unlikely to form a second phase since the amount of aluminium for afor material (Mg-1Al) completely unlikely to form second phase since the amount of aluminium (Mg-1Al) completely unlikely to form a second phase since the amount of aluminium lower than itssolid solid solubility limit(a betterunderstanding understandingofofof this concept willwill beprovided provided is lower than its its solid solubility limit understanding this concept be provided isislower than solubility limit (a(abetter better this concept will be inin in Section 5). Section 5). 5). Section Thepreferential preferentialattack attackmodel modelhas hasbeen beenproposed proposedby byFairman andBray Bray[98] [98]and andititisiscommonly and TheThe preferential attack model has been proposed byFairman Fairman and Bray [98] and itcommonly is commonly believed to be the main dissolution mechanism. As a consequence of alloying operation, itisiswidely widely believed to be the main dissolution mechanism. As a consequence of alloying operation, it believed to be the main dissolution mechanism. As a consequence of alloying operation, it is widely reported [99] that second phases can precipitate on grain boundaries. Their presence enhances reported that second phases precipitate on grain boundaries. Their presence enhances reported [99] [99] thatconsequence second phases can can precipitate on grain boundaries. Their presence enhances stresses stressesasasaaconsequence differentelastic elasticproperties propertieswith withrespect respecttotothe thematrix matrixleading leadingtotoaasurface surface stresses ofofdifferent as a consequence of different elastic properties with respect to the matrix leading to a surface film filmrupture. rupture.Once Once theanodic anodic Mgmatrix matrix andthe the cathodic second phases (Mg17 17Al12 in Fairman and film the Mg and cathodic second phases (Mg Al12 in Fairman and rupture. Once the anodic Mg matrix and the cathodic second phases (Mg Al in Fairman andan Bray’s 17 12 Bray’s work) are in contact with the human body fluid (HBF), which acts as electrolyte, Bray’s work) are in contact with the human body fluid (HBF), which acts as electrolyte, an work) are in contact with the human body fluid (HBF), which acts as electrolyte, an electrochemical electrochemical cell develops, leading to an accelerated dissolution (Figure 6). electrochemical cell develops, leading to an accelerated dissolution (Figure 6). cell develops, leading to an accelerated dissolution (Figure 6).
Figure6.6.Preferential Preferentialcorrosion corrosionofofMg Mgmatrix matrixalong alongthe theprecipitates. Modifiedfrom from[75]. [75]. Figure Modified Figure 6. Preferential corrosion of Mg matrix along theprecipitates. precipitates. Modified from [75].
Dependingon onthe thedistribution distributionofofsecond secondphases, phases,crack crackpropagation propagationisisreported reportedcontinuous continuous[100], [100], Depending Depending on the distribution of second phases, crack propagation is reported continuous thereare arecontinuous continuous grainboundary boundary precipitates, discontinuous [77,78,83,98,101,102], second[100], ififthere grain precipitates, orordiscontinuous [77,78,83,98,101,102], ififsecond phases are non-uniformly non-uniformly distributed unfavourably oriented. [77,78,83,98,101,102], The discontinuous discontinuous crack crack if there are continuous grain boundary precipitates, or discontinuous if second phases are distributed oror unfavourably oriented. The propagation is affected by a slower galvanic dissolution, providing enough time for the formation phases are non-uniformly distributed or unfavourably oriented. The discontinuous crack propagation propagation is affected by a slower galvanic dissolution, providing enough time for the formation ofof thesurface surface protective film,and anddissolution, hencerendering rendering filmrupture rupture andrepassivation repassivation competingprocesses. processes. is affected by aprotective slower galvanic providing enough time for thecompeting formation of the surface the film, hence film and
protective film, and hence rendering film rupture and repassivation competing processes.
Metals 2017, 7, 252 Metals 2017, 7, 252
10 of 40 10 of 41
Concerning the other main group of propagation mechanisms, the mechanical fracture mechanisms or cleavage-like fracture, the model describing hydrogen embrittlement (HE) is reported the other main group ofdescribing propagation mechanisms, the mechanical fracture toConcerning be the most descriptive. Several models cleavage fracture without HE are available [75], mechanisms or cleavage-like fracture, describing hydrogen reported to but not satisfactory. Studying SCCthe of model a single-phase Mg-7.6 wt %embrittlement Al alloy, Pugh(HE) et al.is[96] suggested be the describing cleavage fracture without HEpropagation, are available [75], thatmost the descriptive. propagationSeveral phase models is discontinuous, alternating between crack due but to the notembrittlement satisfactory. Studying of asurface single-phase Mg-7.6ofwt % crack Al alloy, et al. [96] suggested thatductile the of a thinSCC oxide layer ahead the tip,Pugh and crack arrest, due to the propagation phase is discontinuous, alternating between crack propagation, due to the embrittlement region ahead of the embrittled layer. However, Fairman and Bray [98] argued against Pugh et al.’s of amodel thin oxide surface layer of theoxide crack layers tip, andoncrack arrest, due to the ductile region aheadbrittle of as they either do ahead not detect the fracture surfaces or do not observe thefracture embrittled However, Fairman andsolutions. Bray [98] argued Pugh et al.’s model as they eitherthat onlayer. Mg-Al alloys in chloride Severalagainst authors [77,78,83,86,101,102] agreed do cleavage not detectfracture oxide layers the fracture of surfaces or doinduced not observe on Mg-Al Presence alloys in of is a on consequence hydrogen Mg brittle matrixfracture embrittlement. chloride solutions. Several authors agreed thatincleavage fracture is a consequence hydrogen is a consequence of [77,78,83,86,101,102] the Mg corrosion mechanism aqueous solutions (see Section 2); of hydrogen Mg matrix embrittlement. a consequence of the according induced to the cathodic partial reaction, one Presence molecule of of hydrogen hydrogen is gas is produced for eachMg atom corrosion mechanism solutions Section according toatthe partial reaction, of Mg dissolved in in theaqueous human body, but (see hydrogen is 2); also produced ancathodic anodic potential due to the onenegative molecule of hydrogen is produced forSection each atom of Mg dissolved in the human body, that but H2 difference effectgas (NDE, detailed in 4). Chakrapani and Pugh [103] suggested hydrogen is also produced at an anodic potential due to the negative difference effect (NDE, detailed is involved in brittle fracture since it can be related to the formation of brittle hydrides (MgH2) in or in Section 4). Chakrapani and Pugh suggested thatproposed H2 is involved in brittlemodel fracture can be decohesion phenomena. Some[103] authors [104] also an adsorption insince orderitto explain related to the formationwhere of brittle hydrides in decohesion phenomena. authors [104] to the HE mechanism, hydrogen is (MgH adsorbed the solution at the crackSome tip or transported 2 ) orfrom alsothe proposed adsorption model in order to explain the HEBursle mechanism, where[105] hydrogen is adsorbed crackedanregion by dislocation motion. However, and Pugh argued against the from the solution at theThey crackstated tip orthat, transported to the cracked region by dislocation motion. However, adsorption model. according to dislocation motion, the adsorbed hydrogen would Bursle and Pugh argued against adsorption model. They stated that, according dislocation embrittle only[105] a few atomic layersthe ahead of the crack tip, which is smaller than thetoobserved crack motion, adsorbed would embrittle only athey fewobserved atomic layers ahead crack tip, arrestthe markings of hydrogen the fractography. Furthermore, a 1 μm layerofofthe MgH 2 on thewhich fracture is smaller the observed arrest markingsthe of the fractography. Furthermore, they observed surfacethan of their specimenscrack further disproving model. Bursle and other authors [106] stated athat 1 µm layer of MgH fracture surface of of their specimens further the Bursle embrittlement occurs due to the formation brittle hydrides aheaddisproving of the crack tipmodel. favouring crack 2 on the andpropagation other authors [106]suppressing stated that embrittlement the formation while plasticization occurs arounddue theto crack region. of brittle hydrides ahead Figure briefly describes the mechanical based/cleavage likearound fracture mechanism: after of the crack tip 7favouring crack propagation whilefracture suppressing plasticization the crack region. the formation ofdescribes the brittle ahead of thebased/cleavage crack tip, the like applied stresses determine Figure 7 briefly theregion mechanical fracture fracture mechanism: afterthe crack inside theofbrittle area.tip, The ductiledetermine area prevents further crack thepropagation formation of of thethis brittle region ahead the crack thefollowing applied stresses the propagation propagation. Repassivation occursThe ahead of the crack tiparea leading to hydrogen of this crack inside the brittle area. following ductile prevents further embrittlement. crack propagation. Repassivation occurs ahead of the crack tip leading to hydrogen embrittlement.
Figure 7. Model for for mechanical induced cracking. Reproduced with permission from [75], John Wiley Figure 7. Model mechanical induced cracking. Reproduced with permission from [75], John Wiley andand Sons, 2005. Sons, 2005.
Finding a solid mechanism SCC propagation phase (anodic dissolution mechanisms Finding a solid mechanism for for SCC propagation phase (anodic dissolution mechanisms or or cleavage-like fracture) is subject to intensive research. However, the anodic dissolution described cleavage-like fracture) is subject to intensive research. However, the anodic dissolution described by by preferential sufficiently describes the inter-granular specific inter-granular stresscracking corrosion thethe preferential attackattack modelmodel sufficiently describes the specific stress corrosion
Metals 2017, 7, 252
11 of 41
Metals 2017, 7, 252
11 of 40
(IGSCC). Trans-granular stress corrosion cleavage-like fracture due to HE. cracking (IGSCC). Trans-granular stress cracking corrosion(TGSCC) crackingfollows (TGSCC) follows cleavage-like fracture Thus, in order to obtain a more detailed understanding of the problem, fracture surface micrographs due to HE. Thus, in order to obtain a more detailed understanding of the problem, fracture surface relate anodic dissolution anddissolution cleavage-like fracture to inter-granular and trans-granular (TG)transcrack micrographs relate anodic and cleavage-like fracture (IG) to inter-granular (IG) and growth, respectively. Since anodic dissolution requires the presence of second phases, the propagation granular (TG) crack growth, respectively. Since anodic dissolution requires the presence of second mode strongly influenced byismicrostructure. In fact, and WestIn[107] single phases,is the propagation mode strongly influenced byFairman microstructure. fact,tested Fairman andphase West Mg-7Al-1Zn and observing a TG propagation mode asince no Mg17 Al12mode has been from their [107] tested single phase Mg-7Al-1Zn and observing TG propagation sincefound no Mg 17Al12 has microstructural observations, whereas Farman and Bray [98] reported that the Mg-Al alloys with Al been found from their microstructural observations, whereas Farman and Bray [98] reported that the content higherwith thanAl 6 wt % failhigher in a IGthan mode due%tofail theinlarge ofto second phases. Moreover, also Mg-Al alloys content 6 wt a IG presence mode due the large presence of second grain size influences the propagation phase. In fact, increasing the grain size leads to a transition to phases. Moreover, also grain size influences the propagation phase. In fact, increasing the grain size TGSCC. Priest et al. [100] reported the maximum grain size for prevailing IG to be 28 µm, regardless of leads to a transition to TGSCC. Priest et al. [100] reported the maximum grain size for prevailing IG the presence second phases. Bray [98] reported that Mg-Al-Zn and Mg-Al alloys to be 28 μm, of regardless of the Fairman presenceand of second phases. Fairman and Bray [98] reported that with Mgfine grains (at about 5 µm) fail in a IG manner; Pardue et al. [100] observed a TG crack propagation in Al-Zn and Mg-Al alloys with fine grains (at about 5 μm) fail in a IG manner; Pardue et al. [100] their Mg-6Al-1Zn alloy with a coarse grain size (about 80 µm). Similarily, Pugh et al. [96] reported a observed a TG crack propagation in their Mg-6Al-1Zn alloy with a coarse grain size (about 80 μm). TG crack propagation in their high purity binary magnesium alloy wtpurity % Al content Similarily, Pugh et al.mode [96] reported a TG crack propagation mode in with their 7.6 high binary with grain size of 70 µm. All results have been obtained on thermally treated specimens. Without magnesium alloy with 7.6 wt % Al content with grain size of 70 μm. All results have been obtained heat treatment, a combined IG and TG mechanism (anodic dissolutionIGand fracture) is on thermally treated specimens. Without heat treatment, a combined andcleavage-like TG mechanism (anodic underlying SCC propagation [83,101]. In Choudhary and Raman’s work [83], fractographic analyses of dissolution and cleavage-like fracture) is underlying SCC propagation [83,101]. In Choudhary and AZ91D alloy tested m-SBF reveal both and inter-granular cracking. Moreover, Raman’s work [83], under fractographic analyses oftransAZ91D alloy tested under m-SBF reveal both specimens trans- and were also testedcracking. in m-SBFMoreover, under a continuous improvement in corrosion inter-granular specimens cathodic were alsocharge, tested showing in m-SBFan under a continuous cathodic resistance compared to the ones tested without any charge (Figure 3). From micrographic analyses, charge, showing an improvement in corrosion resistance compared to the ones tested without any the authors suggested a cathodic analyses, charge prevents the anodic dissolution mechanism, to charge (Figure 3). Fromthat micrographic the authors suggested that a cathodic chargeleading prevents TGSCC only (indicatedmechanism, by arrows inleading Figure to 8).TGSCC only (indicated by arrows in Figure 8). the anodic dissolution
Figure 8. Fractographic analyses analyses of AZ91D alloy sample continuously cathodically charged and 8. Fractographic and simultaneously tested testedininmodified modifiedsimulated simulated body fluid (m-SBF). Reproduced permission body fluid (m-SBF). Reproduced withwith permission from from [83], Elsevier, [83], Elsevier, 2012.2012.
Until now, now, the the influence influence of of applied applied stresses stresses was was not not considered, considered, it it was was simply simply mentioned mentioned that that Until stresses derive from daily activities or are remaining as residuals after fabrication. The susceptibility stresses derive from daily activities or are remaining as residuals after fabrication. The susceptibility of of SCC applied stress levels was notyet yetdetailed. detailed.ItItisiswidely widelyknown knownthat thatthe theoccurrence occurrence of of SCC SCC SCC to to thethe applied stress levels was not is linked linked to to aa stress stress above above aa certain certainthreshold, threshold,called calledσσSCC.. Thus, in the design phase, a stress below is SCC Thus, in the design phase, a stress below σ SCC should be targeted to avoid the onset of SCC. However, production processes can introduce σSCC should be targeted to avoid the onset of SCC. However, production processes can introduce flaws or ordefects, defects,such such cracks inside the material negatively affecting its strength. Since these flaws as as cracks inside the material negatively affecting its strength. Since these process process limitations are todifficult to overcome, a damage-tolerant approach has into to be taken into limitations are difficult overcome, a damage-tolerant approach has to be taken account. It is account. It is hence important to characterize the conditions for onset and propagation of cracks on hence important to characterize the conditions for onset and propagation of cracks on defects to predict defects to predict the lifetime of a device in a corrosive medium. Since propagate inSCC, a brittle the lifetime of a device in a corrosive medium. Since cracks propagate in cracks a brittle manner in the manner in SCC, the linear elastic fracture mechanics approach (LEFM) is widely used in literature linear elastic fracture mechanics approach (LEFM) is widely used in literature [75,83,94,95], providing providing a design tooltofor predicting the time to failure of an implant. a[75,83,94,95], design tool for predicting the time failure of an implant. LEFM is an important, widely studied and accepted failure prediction criterion. A point criterion insufficiently describes the stress field in the presence of a singularity; a proper description requires a field based criterion. In the proximity of the crack tip, the stress intensity factor in mode I (KI) can be described according to Gross and Mendelson [108]:
Metals 2017, 7, 252
12 of 41
LEFM is an important, widely studied and accepted failure prediction criterion. A point criterion insufficiently describes the stress field in the presence of a singularity; a proper description requires a field based criterion. In the proximity of the crack tip, the stress intensity factor in mode I (KI ) can be Metals 2017, 7, 252 12 of 40 described according to Gross and Mendelson [108]:
√ KKII = α ·σg √πa, πa, α·σ
(9) (9)
where length of of the the initial initial defect, defect, “α” “α” is is aa geometric geometric parameter parameter and “σg”” is the applied where “a” “a” is is the the length and “σ g is the applied stress. of the the main mainunderlying underlyingassumptions assumptionsofof LEFM Paris curve Figure 9, line) red line) stress. One One of LEFM is is thethe Paris curve (see(see Figure 9, red that that describes the fatigue crack propagation rate (da/dN) as a function of the applied stress intensity describes the fatigue crack propagation rate (da/dN) as a function of the applied stress intensity factor factor range range (∆K ). (ΔKI). I
Figure 9. Paris curve (red line); Paris-like curve for Stress Corrosion Cracking (SCC) (black line). Figure 9. Paris curve (red line); Paris-like curve for Stress Corrosion Cracking (SCC) (black line).
The Paris Paris curve curve is characterized characterized by by three three different regions: region region II has has no no crack crack growth growth and and The stress intensity factors below ΔK ∆KIth (threshold stress intensity factor for fatigue crack growth), region stress (threshold stress intensity factor for fatigue crack growth), region Ith III results in sudden failure with loads exceeding the fracture toughness (K ). The The region region labelled labelled II II III (KIC IC). describes a linearly increasing growth rate with the stress intensity factor. describes factor. With respect implant research and and application, a Paris-like curve may alsomay describe With respecttotobiomedical biomedical implant research application, a Paris-like curve also SCC (Figure 9, black9,line). ∆KIth ΔK becomes KISCC (threshold stress intensity crack describe SCC (Figure black line). Ith becomes KISCC (threshold stress intensityfactor factorfor for SCC SCC crack growth) and and the the crack crack growth growth rate is now based on time. To To prevent prevent failure failure due due to SCC, stresses growth) below K KISCC aretargeted, targeted,which which represents a main parameter alloy based implant design. below represents a main parameter forfor MgMg alloy based implant design. Its ISCCare Its calculation is inspired fracture mechanics,testing testingfatigue fatigueair airpre-cracked pre-cracked and and circumferential circumferential calculation is inspired byby fracture mechanics, notched specimens. tensile (CNT) tests are are carried out at different stress stress levels notched specimens. Circumferential Circumferentialnotched notched tensile (CNT) tests carried out at different in corrosive environments aiming at best representing body-like environments [94]. K is calculated at levels in corrosive environments aiming at best representing body-like environments [94]. KI is I each applied Equation (9) and then the(9) corresponding to failure is time extracted. Finally, calculated at load each from applied load from Equation and then thetime corresponding to failure is experimental dataexperimental are plotted indata a time failurein(taf )-K graph (Figure 8). Time to failure increases extracted. Finally, are to plotted time to failure (t f )-K I graph (Figure 8). Time to I exponentially decreasing the applied stress, the i.e., applied the stressstress, intensity and intensity the KISCCfactor, can be and defined failure increases exponentially decreasing i.e., factor, the stress the asISCC thecan horizontal asymptote in the tf -Kasymptote 10).tf-KI plot (Figure 10). K be defined as the horizontal in the I plot (Figure
Metals 2017, 7,7,252 Metals2017, 2017,7, 252 Metals 252
13 of 41 1313ofof4040
Figure 10. for determination ofofK KK with permission from [83], Elsevier, 2012. Figure10. 10.ttff-K t-K f-K Iplot plotfor fordetermination determinationof ISCC.Reproduced Reproducedwith withpermission permissionfrom from[83], [83],Elsevier, Elsevier,2012. 2012. ISCC. Figure II plot ISCC.
Severalworks workshave havebeen beenperformed performedon onevaluating evaluatingthe theMg Mgthreshold thresholdstress stressintensity intensityfactor factorfor for Several Several works performed evaluating the SCC crack growth in corrosive environment, as reported by Winzer in his review [75]. Since the use SCC crack growth growth in in corrosive corrosiveenvironment, environment,asasreported reportedbybyWinzer Winzer review [75]. Since inin hishis review [75]. Since thethe useuse of of Mg and its alloys in biomedical field is a relative new issue, very few consider a corrosive medium of Mg and its alloys in biomedical field is a relative new issue, very few consider a corrosive medium Mg and its alloys in biomedical field is a relative new issue, very few consider a replicatinghuman humanbody bodyconditions. conditions.Choudhary Choudharyand andRaman Raman[83] [83]studied studiedthe thecorrosion corrosionbehaviour behaviourofof replicating Choudhary and Raman 1/2 1/2 AZ91Dalloy alloyin m-SBFand andthey theyreport report aKKISCC ISCC value of 5.18 MPam m1/2. ..Testing Testingspecimens specimensat different AZ91D ininm-SBF m-SBF and they report aaK Testing specimens atatdifferent different ISCC value of 5.18 MPa m 1/2 stress intensity factors, they reported no failure after 1000 h CNT testing at a K I value of 4.8 MPa 1/21/2 stress intensity factors, factors, they they reported reportedno nofailure failureafter after1000 1000hhCNT CNTtesting testingatataaKKI Ivalue valueof of4.8 4.8MPa MPam mm ,, , below the threshold limit that they propose. below the threshold limit that they propose. However,ininsome somecircumstances circumstancesthe thethreshold threshold value KISCC ISCCloses losesits itsmeaning. meaning.For Forsmall smallcracks cracksoror However, some circumstances the threshold value value KKISCC loses its meaning. For small cracks flaws, LEFM is not applicable as it becomes underrating (for a → 0, K I → ∞, see Figure 11). TheSCC SCC flaws, LEFM is not applicable as it becomes becomes underrating underrating (for (for aa → → 0, 0, KKII → → ∞, ∞, see Figure 11). The threshold σ SCC (mentioned above) determined like K ISCC but on smooth specimens, replaces K ISCC threshold σSCC (mentioned ISCC. . . (mentionedabove) above)determined determinedlike likeKKISCC but butononsmooth smoothspecimens, specimens,replaces replacesKK SCC
ISCC
ISCC
Figure11. 11.Definition Definitionofofa**a..*. Figure
The minimumcrack crack lengthfor for LEFMcan canbe becalculated calculatedfrom fromthe thecrack cracklength length“a*” “a*”implying implyingKK I= The I= The minimum minimum cracklength length forLEFM LEFM can be calculated from the crack length “a*” implying KISCCwith withan anapplied appliedstress stressequal equaltotoσσSCC SCC K :: KISCC I = KISCC with an applied stress equal to σSCC : (10) α·σSCC ·√π * ,* , √ (10) ISCC α·σ SCC KKISCC ·√π KISCC = α·σSCC · πa∗ , (10) obtaining: obtaining: obtaining:
**
2
KISCC K2ISCC 22 2 , , Kπα 2 σ2σSCC πα ISCC
(11) (11) SCC , a∗ = (11) 2 σ2 παis SCC Atthe thebest bestofofthe theauthors’ authors’knowledge, knowledge,there there limited informationon onthe theupper upperthreshold thresholdsize size At is limited information *At that may lead to stress corrosion cracking according to LEFM. In the field of Mg and its alloys of a * the best of the authors’ knowledge, there is limited information on the upper threshold size ofas of a that may lead to stress corrosion cracking according to LEFM. In the field of Mg and its alloys as * that may lead biomedical implant materials, it is a potential topic of interest as it provides a powerful design tool a to stress corrosion cracking according to LEFM. In the field of Mg and its alloys as biomedical implant materials, it is a potential topic of interest as it provides a powerful design tool forimplant implantimplant compliance prediction, determining whether theas implant shallabe be designed employing biomedical materials, it is a determining potential topic of interest it provides powerful design tool for for compliance prediction, whether the implant shall designed employing aa solid mechanics approach or fracture mechanistic considerations. solid mechanics approach or fracture mechanistic considerations.
Metals 2017, 7, 252
14 of 41
Metals 2017, 7, 252 implant compliance
of 40 prediction, determining whether the implant shall be designed employing a14solid mechanics approach or fracture mechanistic considerations. Taking into account account σσSCC and KISCC depending on the length of the flaws, a safe region can be Taking into SCC and KISCC depending on the length of the flaws, a safe region can be defined with respect to SCC (Figure defined with respect to SCC (Figure 12) 12) [75]: [75]: for for stresses stresses belonging belonging to to region region I,I, no no SCC SCC will will occur, occur, whereas crack growth duedue to stress andand corrosion. The upper limit whereas region regionIIIIisisaffected affectedbybycontinuous continuous crack growth to stress corrosion. The upper of this region is defined by K IC and by the material’s fracture tensile stress for small cracks. For applied limit of this region is defined by KIC and by the material’s fracture tensile stress for small cracks. stresses above the threshold III),(region suddenIII), failure willfailure occurwill withoccur loadswith exceeding material For applied stresses above the(region threshold sudden loads exceeding strength. Region IRegion is desired in temporary biomedical applications as as implant material strength. I is desired in temporary biomedical applications implantfailure failureisis only only characterized by homogeneous corrosion, avoiding an accelerated failure due to SCC. characterized by homogeneous corrosion, avoiding an accelerated failure due to SCC.
Figure 12. Influence of applied stress on the material behaviour in corrosive environment. Figure 12. Influence of applied stress on the material behaviour in corrosive environment.
3.2. Corrosion Corrosion Fatigue Fatigue (CF) (CF) 3.2. Orthopaedic implants implants experience experience severe severe dynamic dynamic loads loads due due to to walking, walking, running running and and other other body body Orthopaedic 6 walking cycles per year. The simultaneous effect movements. Taylor [109] reported a mean of 2 × 10 movements. Taylor [109] reported a mean of 2 × 106 walking cycles per year. The simultaneous effect of cyclic cyclic loading loading and and corrosive corrosive environment environment leads reduced stresses stresses resistance. Gu et et al. al. [91] [91] found found of leads to to aa reduced resistance. Gu 7 cycles, is that the fatigue limit for Mg high strength extruded alloy WE43, i.e., fatigue strength at 10 7 that the fatigue limit for Mg high strength extruded alloy WE43, i.e., fatigue strength at 10 cycles, is much lower forfor those tested in in airair (40(40 MPa vs. much lower for for specimens specimenstested testedin insimulated simulatedbody bodyfluid fluid(SBF) (SBF)than than those tested MPa 7 cycles, whereas 110 MPa) and that the fatigue strength for die-cast AZ91D tested in air is 50 MPa at 10 7 vs. 110 MPa) and that the fatigue strength for die-cast AZ91D tested in air is 50 MPa at 10 cycles, 6 cycles for 20 MPa at2010MPa specimens tested intested SBF. They defined reduction rate of fatigue whereas at 106 cycles for specimens in SBF. They the defined the reduction rate ofstrength fatigue (RRFS) for evaluating effects ofeffects the corrosive environment. RRFS isRRFS the extension of another index strength (RRFS) for evaluating of the corrosive environment. is the extension of another proposed by Bhuiyan [110] and it is defined as: index proposed by Bhuiyan [110] and it is defined as: σ -σ RRFS σ air−σSBF, (12) RRFS = air σair SBF , (12) σair where σair is the fatigue strength for specimens tested in air and σSBF is the fatigue strength in SBF where σair isnumber the fatigue strength specimens testedRRFS in airmay andrange σSBF isfrom the fatigue strength in SBF at the same of cycles. For for different Mg alloys, 30 to 70%, showing the at the same number of cycles. For different Mg alloys, RRFS may range from 30 to 70%, showing significant impact of the corrosive medium on the fatigue resistance of Mg and its alloys [86,91]. the significant the corrosive medium on the fatigue resistance of Mg and its alloys fatigue [86,91]. (CF) Failureimpact due toofdynamic loading in a corrosive environment is related to corrosion Failure due to dynamic loading in a corrosive environment is related to corrosion fatigue (CF) and and relations can be drawn from SCC. Although CF is considered among the major concerns in relations canapplications be drawn from SCC. Although CF is considered the major in biomedical biomedical causing device failure [111,112], among it is still vastly concerns unexplored. Current applicationsrelates causing failure it is stilltovastly unexplored. Current knowledge knowledge thedevice reduction in [111,112], fatigue strength the nucleation of corrosion pits (as forrelates SCC) the reduction in fatigue strength to the nucleation of corrosion pits (as for SCC) acting as preferential acting as preferential sites for fatigue crack initiation. Jafari et al. [86] investigated as-cast AZ91D sites for fatigue crack initiation. Jafari et al. [86] investigated as-cast AZ91D specimens in modified SBF specimens in modified SBF under different electrochemical conditions, open circuit potential (OCP), under different electrochemical conditions, open circuit potential (OCP), cathodic and anodic charging cathodic and anodic charging condition, reporting that reducing pitting size increases fatigue condition, reporting strength (Figure 13). that reducing pitting size increases fatigue strength (Figure 13).
Metals 2017, 2017, 7, 7, 252 252 Metals
40 15 of 41
Figure 13. S-N curve for AZ91D magnesium alloy under different electrochemical conditions. Figure 13. S-N curve for AZ91D magnesium alloy under different electrochemical conditions. Reproduced with permission from [86], Copyright Elsevier, 2015. Reproduced with permission from [86], Copyright Elsevier, 2015.
The anodic potential. potential. The pitting pitting of of Mg Mg and and its its alloys alloys under under cathodic cathodic potential potential is is lower lower than than under under anodic This This lower lower sensitivity sensitivity mainly mainly affects affects fatigue fatigue behaviour behaviour at at low low stress stress amplitudes, amplitudes, where where pit pit depths depths are are not enough for fatigue crack initiation improving fatigue resistance. not enough for fatigue crack initiation improving fatigue resistance. As As for for SCC, SCC, researchers researchers argue argue between between anodic anodic dissolution dissolution and and hydrogen hydrogen embrittlement embrittlement as as the the main propagation mechanism for CF. Zeng et al. [113] reported that AZ61 alloys mainly suffer of HE main propagation mechanism for CF. Zeng et al. [113] reported that AZ61 alloys mainly suffer of HE in show aa crack-tip crack-tip dissolution dissolution propagation propagation phenomenon phenomenon [114]. [114]. in wet wet air, air, whereas whereas VMD10 VMD10 and and IMV7-1 IMV7-1 show Yet, Uematsu [102] referred to HE as primary and anodic dissolution as the secondary mechanism for Yet, Uematsu [102] referred to HE as primary and anodic dissolution as the secondary mechanism CF propagation [86]. Figure 13 shows that specimens tested under applied potential for crack CF crack propagation [86]. Figure 13 shows that specimens testedanunder ananodic applied anodic are characterized by lower fatigue strength, due to the concurrent action of anodic dissolution and potential are characterized by lower fatigue strength, due to the concurrent action of anodic hydrogen dissolutionembrittlement. and hydrogen embrittlement. Moreover, Moreover, crack crack growth growth rate rate can can be be further further increased increased by by aa simultaneous simultaneous SCC SCC [60]. [60]. Vasudevan Vasudevan and Sadanada [115] distinguished three different combined roles of SCC and CF in crack propagation, and Sadanada [115] distinguished three different combined roles of SCC and CF in crack propagation, depending ) compared to KISCC(Figure (Figure12). 12). depending on on the the applied applied stress stress intensity intensity factor factor (K (Kmax max) compared to KISCC Figure 14a describes what is considered as true CF behaviour since the environment playsa Figure 14a describes what is considered as true CF behaviour since the environment plays aharmful harmfulrole rolefor forall allthe thestress stressintensity intensityfactor factorvalues: values:the thecorrosive corrosiveenvironment environment reduces reduces the the maximum maximum stress intensity factor value for fatigue crack propagation K , leading to a higher crack growth max,th stress intensity factor value for fatigue crack propagation Kmax,th, leading to a higher crack growth rate rate until K approaches the fracture toughness K , where the crack growth rate curve for corrosive max until Kmax approaches the fracture toughness KIC, IC where the crack growth rate curve for corrosive environment In the environment merges merges with with the the one one for for the the inert inert fluid. fluid. In the second second case case (Figure (Figure 14b), 14b), there there is is no no environmental contribution until K reaches K , after which superposition of SCC and CF occurs max ISCC environmental contribution until Kmax reaches KISCC, after which superposition of SCC and CF occurs implying dependant process. The last (Figure 14c) is characterized by the combination implying a astress stress dependant process. Thescenario last scenario (Figure 14c) is characterized by the of the two previously described processes. combination of the two previously described processes. Since areare time-dependent, frequency of Since corrosion corrosionassisted assistedcracking crackingphenomena, phenomena,i.e., i.e.,SCC SCCand andCF, CF, time-dependent, frequency loading (that range from 1 Hz to 3toHz, seesee Table 3) strongly affects fatigue strength. of loading (that range from 1 Hz 3 Hz, Table 3) strongly affects fatigue strength. Table 3. Frequency range for orthopaedic applications. Modified from [60]. Implants
Frequency (Hz)
Activity
Orthopaedic Orthopaedic
1–3 0.5–1.5
Normal walking (vertical direction) Normal walking (lateral direction)
Metals2017, 2017,7,7,252 252 Metals
16of of40 41 16
Figure 14. 14.Schematic Schematic representation crack growth under different combinations of static Figure representation of of crack growth raterate under different combinations of static and and dynamic loading in inert corrosive environments. Crack growth ratedepending dependingon on stress stress dynamic loading in inert andand corrosive environments. Crack growth rate intensity factor factor in in (a) time-stress dependent SCC andand CF intensity (a) pure pure corrosion corrosionfatigue fatigue(CF), (CF),(b) (b)stress stressand and(c)(c) time-stress dependent SCC combinations. Reproduced with permission from [115], Elsevier, 2009. CF combinations. Reproduced with permission from [115], Elsevier, 2009.
Table increase 3 Frequency forcrack orthopaedic Modified fromtime [60].for corrosion and Low frequencies therange fatigue growthapplications. rate, providing enough allowing localized dissolution and/or H to embrittle the matrix. Moreover, the effect of frequency Implants Frequency (Hz) Activity depends on the applied stress, being more pronounced in the low stress regime. The fatigue crack Orthopaedic 1–3 Normal walking (vertical direction) growth rate decreases with frequency due to the Normal crack closure [116]. Several studies on strain Orthopaedic 0.5–1.5 walkingeffect (lateral direction) rate effects on SCC show that intermediate strain rates provide maximum susceptibility for corrosion assisted [75,87,94,95,117]. At low strain rate, filmrate, integrity is preserved, hydrogen Lowcracking frequencies increase the fatigue crack growth providing enough avoiding time for corrosion to enter and localized embrittledissolution the Mg matrix. Increasing strain rate, the repassivation effect is limited, and allowing and/or H to embrittle the matrix. Moreover, the effect of frequency allowingon thethe hydrogen facilitate themore crackpronounced propagation. strainregime. rates there insufficient depends appliedto stress, being in At thehigher low stress The is fatigue crack time forrate hydrogen embrittlement minimizing thiscrack synergistic rendering determined growth decreases with frequency due to the closureinfluence effect [116]. Severalstrain studies on strain cracking dominating. rate effects on SCC show that intermediate strain rates provide maximum susceptibility for corrosion assisted cracking [75,87,94,95,117]. At low strain rate, film integrity is preserved, avoiding hydrogen 4. Negative Effect to enter andDifference embrittle the Mg(NDE) matrix. Increasing strain rate, the repassivation effect is limited, allowing the hydrogen to facilitate the crack propagation. At higher strain rates there is insufficient The NDE is an unusual phenomenon affecting especially magnesium dissolution. From an time for hydrogen embrittlement minimizing this synergistic influence rendering strainand determined electrochemical point of view, corrosion reactions can be either cathodic or anodic, they are cracking inverselydominating. related. For traditional metals (i.e., copper and steel) an increase in applied anodic/cathodic
potential leads to an increase in the anodic/cathodic reaction rate, increasing/decreasing 4. Negativeevolution. Difference Effect (NDE) hydrogen Looking 15, thephenomenon solid lines labelled as Iespecially the normal anodicFrom and the a and Ic represent The NDEatisFigure an unusual affecting magnesium dissolution. an cathodic partial reaction, respectively, both assumed to follow Tafel kinetics. Increasing the applied electrochemical point of view, corrosion reactions can be either cathodic or anodic, and they are potential from Ecorr totraditional Eappl leadsmetals to a decrease in cathodic reaction rate (from I0 to IH,eanodic/cathodic ) and an increase inversely related. For (i.e., copper and steel) an increase in applied in anodicleads reaction rate (from Iin ). However, Mg and itsrate, alloys behave in a different manner, 0 to potential to an increase theIMg,e anodic/cathodic reaction increasing/decreasing hydrogen anodic polarisation increases the amount of hydrogen evolution, i.e., an increase in cathodic reaction evolution. rate Looking according IH curve. Increasing thelabelled appliedaspotential Ecorr to , the anodic actual cathodic atto Figure 15, the solid lines Ia and Icfrom represent theEappl normal and the corrosionpartial rate corresponds to IH,m , which is assumed greater than the expected current IIncreasing also the H,e . Moreover, cathodic reaction, respectively, both to follow Tafel kinetics. the applied Mg corrosion rate is increased. According to the I curve, anodic corrosion corresponds to IMg,m , Mg potential from Ecorr to Eappl leads to a decrease in cathodic reaction rate (from I0 to IH,e) and an increase greater than IMg,e . rate (from I0 to IMg,e). However, Mg and its alloys behave in a different manner, in anodic reaction
anodic polarisation increases the amount of hydrogen evolution, i.e., an increase in cathodic reaction rate according to IH curve. Increasing the applied potential from Ecorr to Eappl, the actual cathodic corrosion rate corresponds to IH,m, which is greater than the expected current IH,e. Moreover, also the
Metals 2017, 7, 252 Metals 2017, 7, 252
17 of 40 17 of 40
Mg corrosion rate is increased. According to the IMg curve, anodic corrosion corresponds to IMg,m, Metals 2017, 7, 252 17 of 41 Mg corrosion greater than IMg,erate . is increased. According to the IMg curve, anodic corrosion corresponds to IMg,m, greater than IMg,e.
Figure representation of negative different effect. Reproduced with permission from Figure15. 15.Schematic Schematic representation of negative different effect. Reproduced with permission Figure 15. Schematic representation of negative different effect. Reproduced with permission from [118], Elsevier, 1997. from [118], Elsevier, 1997. [118], Elsevier, 1997.
Several authors have tried to explain the NDE phenomenon, developing four different models Several authors have tried to explain the NDE phenomenon, developing four different models Several authors shortly described here:have tried to explain the NDE phenomenon, developing four different models shortly shortly described described here: here: i. Partially protective surface film i. i. Partially protective surface protective surface film ii. Monovalent magnesium ionfilm model ii.ii. Monovalent magnesium ion magnesium ion model model iii. Particle undermining model undermining model iii. Particle undermining model iv. Magnesium hydride (MgH2) model iv. (MgH22) model iv. Magnesium hydride (MgH For a deeper insight, the readers are referred to [77]. For aa deeper deeper insight, insight,the thereaders readersare arereferred referredto to[77]. [77]. For 4.1. Partially Protective Surface Film 4.1. Partially Partially Protective Protective Surface Surface Film Film 4.1. It is well known that a passivating film forms naturally on the surface of Mg, preventing further is well well known that passivating film forms naturally onrupture the surface surface of Mg, Mg, preventing preventing further ItIt is aa passivating film naturally the of corrosion. Thisknown modelthat suggests that NDE is forms attributed to theon of the protective layer further when corrosion. This model suggests that NDE is attributed to the rupture of the protective layer when corrosion. attributed rupture of the protective layer when current is flowing through the interfacial electrical doublethe layer formed between the magnesium currentand is flowing flowing throughmedium. the interfacial interfacial electrical double layer formeddivalent between the magnesium magnesium current is through the double layer formed between the surface the electrolytic Once electrical the protective film is broken, magnesium ions surface and and the electrolytic medium. Once the the protective film is broken, broken, divalent magnesium ions surface electrolytic medium. Once is magnesium ions dissolve andthe undergo hydrolysis lowering theprotective pH of thefilm solution and divalent increasing the parasitic dissolve and undergo hydrolysis lowering the pH of the solution and increasing the parasitic dissolve and hydrolysis lowering the pH the solution and increasing the parasitic corrosion corrosion rate.undergo The higher the current density or of potential is, the greater the removal of protective corrosion rate. The higher the current density or is,the theremoval greater of theprotective removal of protective rate. The be higher the 16). current density or potential is, potential the greater layer will be layer will (Figure layer will (Figure 16).be (Figure 16).
(a) (b) (a) (b) Figure 16. Model of partially protective surface film at low (a) and high (b) applied potential (E) or Figure 16.Modified Model of partially protective surface film at low (a) and high (b) applied potential (E) or current [85].protective Figure (I). 16. Model of from partially surface film at low (a) and high (b) applied potential (E) or current (I). Modified from [85]. current (I). Modified from [85].
Metals 2017, 7, 252
Metals 2017, 7, 252
18 of 40
18 of 41
4.2. Monovalent Magnesium Ion Model
4.2. Monovalent Magnesium Ion Model This model proposes that a transient Mg+ monovalent ion may be involved in the NDE as
This model proposes that a transient Mg+ monovalent ion may be involved in the NDE as illustrated in Figure 17. illustrated in Figure 17. Metals 2017, 7, 252
18 of 40
4.2. Monovalent Magnesium Ion Model This model proposes that a transient Mg+ monovalent ion may be involved in the NDE as illustrated in Figure 17.
Figure 17. Monovalent magnesium ion model. Modified from [85].
Figure 17. Monovalent magnesium ion model. Modified from [85].
Mg+ monovalent ion is produced electrochemically according to:
Mg+ monovalent ion is produced electrochemically according to: + − Mg → Mg + e ,
(13)
then chemically reacting with two protons toMg yield →hydrogen Mg+ + e−gas , via: Figure 17. Monovalent magnesium ion model. Modified from [85]. 2Mg + 2H → 2Mg + H2, +
+
2+
then chemically reacting with two protons to yield hydrogen gas to: via: Mg+ monovalent ion is produced electrochemically according
(13) (14)
providing a chemical rather than electrochemical explanation of hydrogen production. Mg → Mg+ + e−,
(13)
2Mg+ + 2H+ → 2Mg2+ + H2 ,
4.3. Particle Model with two protons to yield hydrogen gas via: then Undermining chemically reacting
(14)
+ + 2H+explanation 2+ + H This amodel suggests NDE to be explained development of production. second phases (14) or providing chemical ratherthe than electrochemical 2Mg → by 2Mgthe 2,of hydrogen
impurities, especially at high anodic current densities or potentials. Second phases and impurities are
providing a chemical rather than electrochemical explanation of hydrogen production. 4.3. Particle Undermining Model more cathodic with respect to the Mg matrix, resulting in high localized galvanic corrosion at their
boundaries. The Undermining phases fall Model out due to an increased mass loss weakening the material. This Particle This4.3. model suggests the NDE to be explained by the development of second phases or impurities, phenomena increases with higher current densities (Figure 18).
especially atThis high anodic current potentials. Second phases of and impurities model suggests the densities NDE to beorexplained by the development second phases are or more impurities, especially at high anodic current densities or potentials. Second phases and impurities are cathodic with respect to the Mg matrix, resulting in high localized galvanic corrosion at their moreThe cathodic with respect to the matrix, resulting in high localized galvanic corrosion their boundaries. phases fall out due to Mg an increased mass loss weakening the material. Thisatphenomena boundaries. The phases fall out due to an increased mass loss weakening the material. This increases with higher current densities (Figure 18). phenomena increases with higher current densities (Figure 18).
(a)
(b)
Figure 18. Particle undermining model, at low (a) and high (b) current density or potential. Modified from [77].
(a)
4.4. Magnesium Hydride (MgH2) Model
(b)
Figure 18. Particle undermining model, at low (a) and high (b) current density or potential. Modified
This model buildsundermining on the formation of aat MgH layer onhigh the specimens’ surface [119], which reacts Figure 18. Particle model, low2 (a) and (b) current density or potential. Modified from [77]. with H 2 O to produce H 2 : from [77]. 4.4. Magnesium Hydride (MgH 2) Model MgH 2 +2H2O → Mg2+ + 2OH− + 2 H2,
(15)
4.4. Magnesium (MgH 2 ) Model ThisHydride builds formation of a MgH2 layer on the specimens’ surface [119], which reacts This can bemodel significant inonanthe anodic condition (Figure 19). with H2O to produce H2:
This model builds on the formation of a MgH2 layer on the specimens’ surface [119], which reacts MgH2 +2H2O → Mg2+ + 2OH− + 2 H2, (15) with H2 O to produce H2 : 2+ − +2H2condition O → Mg(Figure + 2OH (15) This can be significantMgH in an 2anodic 19). + 2H2 , This can be significant in an anodic condition (Figure 19).
Metals 2017, 7, 252
19 of 41
Metals 2017, 7, 252
19 of 40
Figure 19. Magnesium hydride MgH2 model, modified from [77] Figure 19. Magnesium hydride MgH2 model, modified from [77]
Song et al. [77] combined model 1 and 2, providing one of the most applicable models. Increasing Song al. [77] combined 1 and 2, providing of the mostmaterial applicable models.film Increasing applied et potential or currentmodel density, a larger surfaceone area of the becomes free, Mg applied potential or current density, a larger surface area of the material becomes film free, Mg + produces monovalent Mg (Equation (13)), which in turn leads to hydrogen generation via: produces monovalent Mg+ (Equation (13)), which in turn leads to hydrogen generation via: 1 Mg+ + H2O → Mg+ + OH− + H2, (16) 2 1 + + − Mgof applied + H2 O potential/high → Mg + OHcurrent + H2densities, the model considers (16) For strong negative values the 2 film surface intact, leading to low anodic dissolution, whereas cathodic hydrogen evolution is still For strong negative values of appliedpotential. potential/high current densities, theismodel considers the allowed but decreasing with increased Once the pitting potential reached, the surface film surface intact, leading to low anodic dissolution, whereas cathodic hydrogen evolution is still film breaks down, favouring hydrogen evolution and Mg ions dissolution, which, of course, increases allowed but decreasing with increased potential. Once the pitting potential is reached, the surface film with applied potential/current density. breaks down, favouring hydrogen evolution and Mg ions dissolution, which, of course, increases with applied potential/current density. 5. Corrosion Improvements The importance of temporary biomedical devices is continuously increasing. Mg-based 5. Corrosion Improvements materials are emerging as attractive metallic materials in this field, because of bone like mechanical The importance of temporary biomedical devices is continuously increasing. Mg-based materials properties and chemical compatibility (Mg is a trace element of the human body). The main limitation are emerging as attractive metallic materials in this field, because of bone like mechanical properties is their low corrosion resistance leading to implant failure before bone healing is sufficient. and chemical compatibility (Mg is a trace element of the human body). The main limitation is their Researcher efforts start to grow in this field, aiming to improve Mg and its alloys’ corrosion low corrosion resistance leading to implant failure before bone healing is sufficient. Researcher efforts behaviour. In the last years, some methods have been developed, acting either on the bulk material start to grow in this field, aiming to improve Mg and its alloys’ corrosion behaviour. In the last years, or on the surface. These will be discussed in detail below. some methods have been developed, acting either on the bulk material or on the surface. These will be discussed in detail below. 5.1. High-Purity Mg and Mg-Alloys 5.1. High-Purity Mgpoor and Mg-Alloys Because of molten metal handling processes and as a consequence of the natural composition of the raw Mg, some undesirable impurities, as iron (Fe), (Ni) and copper Because of poor molten metal handling processes and as asuch consequence of thenickel natural composition are incorporated into the material. Moreover, theiron amount impurities is inversely related of(Cu), the raw Mg, some undesirable impurities, such as (Fe), of nickel (Ni) and copper (Cu), areto the efficiency of the refining process. The effect of impurities on Mg based materials’ corrosion incorporated into the material. Moreover, the amount of impurities is inversely related to the efficiency has thus gained greatofinterest. Inclusions are thematerials’ main issue affecting corrosionhas andthus thus ofbehaviour the refining process. The effect impurities on Mg based corrosion behaviour have great been interest. intensively investigated recent [120,121], especially conjunction with the gained Inclusions are the in main issueyears affecting corrosion and thusinhave been intensively leaching of potential toxic (Ni) elements into the surrounding causing inflammations [68,85,122,123]. investigated in recent years [120,121], especially in conjunction with the leaching of potential toxic (Ni) The corrosion is reportedcausing to increase 10–100 times if the concentration of impurities rises beyond elements into therate surrounding inflammations [68,85,122,123]. The corrosion rate is reported to the solid solubility toleranceof maximum [124]. solidthe solubility limit is defined as the increase 10–100 timesdetermined if the concentration impurities risesThis beyond solid solubility determined extent to which an[124]. element baseismaterials in thistocase) without forming tolerance maximum This will soliddissolve solubilityinlimit defined as(Mg the extent which an element will a different phase [85]. Thus solid solubility limita different leads to phase more [85]. homogenously dispersed dissolve in base materials (Mgainhigher this case) without forming Thus a higher solid elements, whereas a lower one increases the amount of separate phases within the Mg-matrix. The solubility limit leads to more homogenously dispersed elements, whereas a lower one increases the tolerance limits are usually low and the limit at which segregation in pure metal (Fe) or Mgamount of separate phases within the Mg-matrix. The tolerance limits are usually low and the limit phasesin (Ni, Cu)metal occurs, as low as 35–50phases ppm for Fe,Cu) 100–300 ppm andas 20– atintermetallic which segregation pure (Fe)can or be Mg-intermetallic (Ni, occurs, canfor beCu, as low 50 ppm for Ni [68]. 35–50 ppm for Fe, 100–300 ppm for Cu, and 20–50 ppm for Ni [68]. In these segregations, a galvanic corrosion is higher as the standard reduction potential is greater than that of Mg. The formation of this corrosion cell induces a preferential dissolution of Mg acting
Metals 2017, 7, 252
Metals 2017, 7, 252
20 of 41
20 of 40
In these segregations, a galvanic corrosion is higher as the standard reduction potential is greater than thatmaterial, of Mg. The formation of this cell material induces awill preferential acting as as anodic while the mass of corrosion the cathodic remain. dissolution Moreover, of theMg galvanic anodic material, whilephenomenon the mass of the cathodic material willformation remain. Moreover, galvanicthe corrosion corrosion is a localized also increasing pitting and thus the enhancing SCC is a localized phenomenon also increasing pitting formation and thus enhancing the SCC and CF and CF sensitivity. Several authors reported that SCC susceptibility increases with Fe concentration sensitivity. Several authors reported susceptibility with Fe concentration due to emerging corrosion cells betweenthat theSCC anodic matrix and increases cathodic impurities [100,125]. due to emerging corrosion to cells between the anodic matrix and cathodiclevels, impurities [100,125].their activity It is fundamental keep these elements under their tolerance or to moderate It is fundamental to keep these elements under their tolerance levels, or tois moderate utilizing alloying element (discussed later). Achieving higher corrosion resistance linked to their an activity utilizing alloying element (discussed later). Achieving higher corrosion resistance is linked amount of impurities (Fe, Ni and Cu) under the tolerance limits. Shi et al. [126] discovered better to an amount of impurities Ni and Cu) under thecomparing tolerance limits. Shihigh et al.purity [126] Mg discovered better corrosion resistance of high(Fe, purity Mg and its alloys low and and Mg1Al corrosion resistance of high purity Mg and its alloys comparing low and high purity Mg and Mg1Al alloys (Figure 20) in salt immersion test (SIT) and salt spray test (SST). Hofstetter et al. [127] stated alloys (Figure 20) in saltZX50 immersion testshows (SIT) and salt spray test (SST). Hofstetter et al.to [127] stated that that an ultrahigh purity Mg alloy greater corrosion resistance compared high purity an ultrahigh purity ZX50 Mg alloy shows greater corrosion resistance compared to high purity alloys alloys (almost three times) and to standard purity alloy (over an order of magnitude). Li et al. [128] (almostthat threespecimens times) andmade to standard purity (over anhave orderno of mass magnitude). Li et al. [128] reported reported from 99.99 wtalloy % pure Mg loss even after 180 days of that specimens immersion in SBF.made from 99.99 wt % pure Mg have no mass loss even after 180 days of immersion in SBF.
Figure Comparison corrosion rates high purity (HP) with low purity (CP) Figure 20.20. Comparison of of thethe corrosion rates of of thethe high purity (HP) with low purity MgMg (CP) in in thethe salt immersion test (3% NaCl SIT) and salt spray test (SST). Reproduced with permission from [126], salt immersion test (3% NaCl SIT) and salt spray test (SST). Reproduced with permission from [126], Elsevier, 2006. Elsevier, 2006.
Tolerance Level Tolerance Level The tolerance level represents a threshold value, below which corrosion rate limited, and The tolerance level represents a threshold value, below which thethe corrosion rate is is limited, and once the amount of impurities exceeds this threshold, the corrosion rapidly increases. Obviously, once the amount of impurities exceeds this threshold, the corrosion rapidly increases. Obviously, each element has own solid solubility determined tolerance limit. same concentration, each element has itsits own solid solubility determined tolerance limit. AtAt thethe same concentration, thethe detrimental effect impurities decreases follows:NiNi Cu. However, a typical Mg alloy detrimental effect of of impurities decreases asasfollows: >> FeFe >> Cu. However, a typical Mg alloy incorporates several alloying elements, all of which cross-influence their tolerance levels. An element incorporates several alloying elements, all of which cross-influence their tolerance levels. An element specific impurity threshold level is thus possible define. best authors’ knowledge, specific impurity threshold level is thus notnot possible to to define. ToTo thethe best of of thethe authors’ knowledge, the first study on the critical level of impurities has been performed by Hanawalt et al. [129]. They have the first study on the critical level impurities has been performed by Hanawalt et al. [129]. They carried outout corrosion tests on on MgMg alloys immersed in 3% NaCl solution and compared it to pure Mg. have carried corrosion tests alloys immersed in 3% NaCl solution and compared it to pure ForFor pure Mg, they reported thethe tolerance level forfor Fe,Fe, NiNi and CuCu to to bebe 0.017 wtwt %,%, 0.0005 wtwt %% and Mg. pure Mg, they reported tolerance level and 0.017 0.0005 0.10.1 wt wt %, %, respectively. Increasing the amount of Manganese (Mn) (Mn) from 0.2 to 20.2 wtto%,2 the and respectively. Increasing the amount of Manganese from wt Ni %, tolerance the Ni limit grows 0.001 to 0.001 0.015 to wt0.015 % (Figure tolerance limitfrom grows from wt % 21). (Figure 21).
Metals 2017, 7, 252
21 of 41
Metals 2017, 7, 252
21 of 40
Figure Figure 21. 21. Tolerance Tolerance limits limits for for Nickel Nickel in in pure pure Mg Mg and and Mg-Zn Mg-Zn and and Mg-Mn Mg-Mn alloys. alloys. Modified Modified from from [129]. [129].
Othershave havestudied studiedNiNi tolerance levels, revealing not third only elements third elements the Others tolerance levels, revealing that that not only but alsobut thealso casting casting has method has influences on the [77,130]. solubilitySand [77,130]. Sand and permanent mould have a method influences on the solubility and permanent mould casted have acasted significantly significantly lower nickel solid solubility (10 ppm) as high pressure die casted AZ91 alloys (50 ppm) lower nickel solid solubility (10 ppm) as high pressure die casted AZ91 alloys (50 ppm) [77]. [77]. Hanawalt et al. [129] studied the solid solubility limits of the three main impurities on binary Hanawalt al.Mg-Al [129] studied the solid solubility limits of the three impurities binary alloys, especiallyeton alloys. They reported a significant influence of main Al content on Fe on tolerance alloys, especially on Mg-Al alloys. They reported a significant influence of Al content on Fe tolerance level, in contrast to none observed on Cu and Ni. 7 wt % of Al in the alloy drops the tolerance level of level, in 0.017 contrast none observed Cureason and Ni. 7 wtformation % of Al inofthe alloy drops(FeAl the tolerance level Fe from wt to % to 0.0005 wt %. on The is the Fe-Al phases 3 ) being even of Fe from 0.017 wt % to 0.0005 wt %. The reason is the formation of Fe-Al phases (FeAl 3 ) being even more cathodic than iron particles. moreFurthermore, cathodic thanthey ironstudied particles. also ternary alloys, such as Mg-Al-Mn, reporting that an amount of they studied also ternary Mg-Al-Mn, reporting that an amount of 0.2 wtFurthermore, % of Mn rendering Fe tolerance levelsalloys, not tosuch dropasbelow 0.002 wt %, meaning that they are 0.2 wt % of Mn rendering Fe tolerance levels not to drop below 0.002 wt %, meaning that they are stable over a wide range of Al contents. stable over a Mn wideasrange of Alelement contents. Adding alloying to Mg-Al improves corrosion resistance by increasing the Fe Adding Mn as alloying A element Mg-Al improves corrosion resistance the Fe tolerance level [77,130,131]. weighttoratio Fe/Mn of 0.032 is widely definedby to increasing be the threshold tolerance level [77,130,131]. A weight ratio Fe/Mn of 0.032 is widely defined to be the threshold above above which the corrosion rate highly increases [77,130,131]. Mercer and Hillis [132] found that the which the corrosion highly increases [77,130,131]. Mercer Hillis [132] found that thethe Fe Fe tolerance levels forrate AE42 and AM60 alloys are different, but and comparable when considering tolerance levels for AE42 and AM60 alloys are different, but comparable when considering the Fe/Mn Fe/Mn ratio. ratio.The copper tolerance level is also influenced by other alloying elements. A small amount of copperfor tolerance level is also by other alloyingFor elements. A small amount of of Cu Cu Cu isThe beneficial creep strength, butinfluenced strongly affects corrosion. example, the addition is beneficial for creep strength, but strongly affects corrosion. For example, the addition of Cu in Mgin Mg-Al-Zn alloys has a detrimental effect due to the incorporation of Cu in the eutectic phase as Al-Zn alloys has a The detrimental effectlevel due is tohighly the incorporation of Cu the eutectic phase Mg(Cu, Mg(Cu, Zn) [106]. Cu tolerance influenced by Zn.in Song and Atrens [77]asreported [106]. The Cu tolerance is highly influenced by Zn. Songwith andHanawalt Atrens [77] reported a higher aZn) higher Cu tolerance if Zn islevel above 0.4 wt %. These results agree [129] reporting that Cu tolerance if Zn is above 0.4 wt %. These results agree with Hanawalt [129] reporting that the the addition of 3 wt % Zn increases Cu corrosion tolerance in a Mg-Al-Mn (0.2 wt %) system, whereas addition of are 3 wtobserved % Zn increases Cu%corrosion tolerance in a Mg-Al-Mn (0.2 wt %) system, whereas no no changes at 0.5 wt Zn (Figure 22). changes are observed at 0.5 wt % Zn (Figure 22).
Metals 2017, 7, 252 Metals 2017, 7, 252
22 of 41 22 of 40
Figure 22. 22. Effect Effect of of Cu Cu in in aa Mg-Al-Mn-Zn Mg-Al-Mn-Zn alloy. alloy. Modified Modified from from [129]. Figure [129].
5.2. Alloying Alloying 5.2. No alloying outperforms the corrosion resistance of ultra-high purity Mg [133]. Mg Increasing No alloyingstrategy strategy outperforms the corrosion resistance of ultra-high purity [133]. alloying elements content increases second phases, promoting the establishment of a local galvanic Increasing alloying elements content increases second phases, promoting the establishment of a local cell at interfaces. For example, Shi et al. the corrosion rate ofrate high Mg with galvanic cell at interfaces. For example, Shi[126] et al.compared [126] compared the corrosion of purity high purity Mg high a purity alloy containing 5 wt % Al. They found a higher corrosion rate for the latter due with high a purity alloy containing 5 wt % Al. They found a higher corrosion rate for the latter due to the the microgalvanic cathodic to microgalvanic corrosion corrosion acceleration acceleration of of the the Mg Mg matrix matrix by by the the adjacent adjacent Mg Mg1717Al Al12 12 cathodic phases. However, cannot provide the mechanical properties required for biomedical devices. phases. However,pure pureMg Mg cannot provide the mechanical properties required for biomedical Therefore, the use of alloying elements has also gained interest for increasing mechanical properties devices. Therefore, the use of alloying elements has also gained interest for increasing mechanical (strength, elastic modulus, at fracture,atetc.). Moreover, the addition of certainofalloying properties (strength, elastic elongation modulus, elongation fracture, etc.). Moreover, the addition certain elementselements is reported to improve the corrosion resistance of unalloyed Mg with conventional purity. alloying is reported to improve the corrosion resistance of unalloyed Mg with conventional The improvement in the corrosion behaviour can be achieved by three different ways: purity. The improvement in the corrosion behaviour can be achieved by three different ways:
••
Since grain grain boundaries boundaries are are characterized characterized by higher Refining the grain size through alloying. Since imperfection and andhigher higherinternal internalenergy energycompared compared matrix, corrosive attack imperfection to to thethe MgMg matrix, any any corrosive attack acts acts preferentially on grain boundaries. Segregation of alloying elements and second preferentially on grain boundaries. Segregation of alloying elements and second phasesphases occur occur on boundaries these boundaries to an accelerated cathodic of the surrounding Mg on these leadingleading to an accelerated cathodic activityactivity of the surrounding Mg matrix. matrix. This would normally coarse grains, however, such segregationsare are continuously continuously This would normally favourfavour coarse grains, however, such segregations distributed in Mg alloys with finer grains leading to a more homogeneous corrosion behaviour corrosion barrier [60,68]. Furthermore, Furthermore, fine fine grain grain sizes sizes improve improve the the corrosion corrosion assisted assisted acting as a corrosion inhibit crack initiation and dislocation motion and lead to an increase cracking resistance resistancesince sincethey they inhibit crack initiation and dislocation motion and lead to an in the number barriers crack propagation. increase in the of number of to barriers to crack propagation. •• Surrounding theMgMg matrix continuously with passivating second phases allows the Surrounding the matrix continuously with passivating second phases allows the development development of an oxidative film protecting the Mg matrix and acts as barrier of an oxidative film protecting the Mg matrix and acts as barrier to hamper corrosion.to hamper • corrosion. Adding elements reduces precipitation of second phases at grain boundaries or balances the • Adding precipitation of second phases at grain microgalvanic boundaries orcorrosion. balances the potentialelements differencereduces between matrix and second phases to decrease potential difference between matrix and second phases to decrease microgalvanic corrosion. Several alloying strategies allow achieving these improvements, it is however necessary to identify Several alloying strategies achieving theseused improvements, it is however necessary to nontoxic ones. In this section, we allow provide a list of most alloying elements and their effects. It has identify nontoxic In of this section, we provide a list related of mosttoused alloying elements andadded. their to be noted that theones. effects alloying elements are strictly the system where they are effects. It has to be noted that the effects of alloying elements are strictly related to the system where they added. 5.2.1.are Aluminium Al is the most common addition to Mg, it is relatively cheap, light, soluble, and improves strength 5.2.1. Aluminium considerably (i.e., from 170 to 250 MPa considering AZ91) [134]. Furthermore, it is passivating and Al is corrosion the most resistance. common addition to behaviour Mg, it is relatively light, and improves improves Corrosion studies of cheap, Mg alloys aresoluble, almost exclusively on strength considerably from to [99] 250tested MPa pure considering AZ91) [134]. Furthermore, it is Mg-Al alloys, especially(i.e., AZ91. Song170 et al. Mg and different Mg-Al alloys in chloride passivating and improves Corrosion behaviour studies of Mg alloys are almost solution finding that purecorrosion Mg has aresistance. higher anodic dissolution rate than AZ21. The surface film of exclusively Mg-Al especially AZ91. Song et an al. Al [99] tested pure Mg and different Mg-Al specimens inon this studyalloys, consists of three different layers: O inner layer, a MgO middle layer and 2 3 alloys in chloride solution finding that pure Mg has a higher anodic dissolution rate than AZ21. The surface film of specimens in this study consists of three different layers: an Al2O3 inner layer, a MgO
Metals 2017, 7, 252 Metals 2017, 2017, 7, 7, 252 252 Metals
23 of 40 23 of of 41 40 23
middle layer and a Mg(OH)2 outer layer (Figure 23). The higher corrosion resistance of AZ21 is middle layer and a Mg(OH)2 outer layer (Figure 23). The higher corrosion resistance of AZ21 is related to the emergence of passivating Al2O3. arelated Mg(OH) outer layer (Figure 23). The higher to 2the emergence of passivating Al2O3.corrosion resistance of AZ21 is related to the emergence of passivating Al2 O3 .
Figure Figure 23. 23. Corrosion Corrosion interface interface of of AZ21 AZ21 alloy. alloy. Reproduced Reproducedwith withpermission permissionfrom from[99], [99],Elsevier, Elsevier,1998. 1998. Figure 23. Corrosion interface of AZ21 alloy. Reproduced with permission from [99], Elsevier, 1998.
Compared to AZ21, AZ21, AZ91 alloys alloys havehigher higher corrosionrates rates dueto to increasedcathodic cathodic second Compared Compared to to AZ21, AZ91 AZ91 alloys have have higher corrosion corrosion rates due due to increased increased cathodic second second phases along grain boundaries. Al is soluble to almost 12 wt % in Mg, depending on the temperature phases phases along along grain grain boundaries. boundaries. Al Al is is soluble soluble to to almost almost 12 12 wt wt % % in Mg, depending depending on on the temperature temperature (Figure 24). (Figure (Figure 24). 24).
Figure 24.Mg-Al Mg-Al phasediagram. diagram. Reprinted with with permissionfrom from [135]. Figure Figure 24. 24. Mg-Al phase phase diagram. Reprinted Reprinted with permission permission from [135]. [135].
Gusieva et al. [136] reported alloys of higher Al content than AZ31 to have second phases Gusieva et et al. al. [136] [136] reported reported alloys alloys of of higher higher Al Al content content than than AZ31 AZ31 to to have have second second phases phases Gusieva (Mg17Al12) implying that an amount of Al above 3 wt % lowers the corrosion resistance. Some authors (Mg17 17Al12 12)) implying resistance. Some authors authors (Mg implying that that an an amount amount of Al Al above above 3 wt wt % % lowers lowers the corrosion resistance. contradicted this hypothesis stating that an increase in Al steadily rises corrosion resistance [137,138]. contradicted this this hypothesis hypothesis stating stating that that an increase in Al steadily rises corrosion resistance [137,138]. contradicted Lunder et al. [138], for example, proposed the anodic dissolution to be further decreased with Al Lunderetetal. al.[138], [138],forfor example, proposed anodic dissolution be further decreased with Al Lunder example, proposed the the anodic dissolution to beto further decreased with Al above above 10 wt %. Winzer et al. [75] finally resumed these studies observing two influences of Mg17Al12 above 10 wt %. Winzer et al. [75] finally resumed these studies observing two influences of Mg 17 Al 10 wt %. Winzer et al. [75] finally resumed these studies observing two influences of Mg17 Al12 phases12 phases on corrosion, they are a (1) barrier and (2) galvanic cathode influences, depending on the phases on corrosion, are a (1) barrier and (2)cathode galvanicinfluences, cathode influences, on the on corrosion, they are they a (1) barrier and (2) galvanic dependingdepending on the amount of amount of second phases and on their distribution. Mg17Al12 acts as a galvanic cathode and accelerates amountphases of second on their distribution. Mgas 17Al acts as acathode galvanicand cathode and accelerates second andphases on theirand distribution. Mg17 Al12 acts a 12galvanic accelerates corrosion corrosion at low volume fractions, whereas when forming an interconnected network at high corrosion at low volume fractions, whereas forming an network interconnected high at low volume fractions, whereas when formingwhen an interconnected at high network fractions,at reduce fractions, reduce corrosion acting as a barrier through the passivating properties of Al (Figure 25). fractions, acting reduceascorrosion as athe barrier through the passivating properties corrosion a barrieracting through passivating properties of Al (Figure 25). of Al (Figure 25).
Metals 2017, 7, 252 Metals 2017, 7, 252
24 of 41 24 of 40
Figure 25. Schematic illustration of the role of Mg17 Al12 phases in the Mg matrix whether their Figure 25. Schematic illustration of the role of Mg17Al12 phases in the Mg matrix whether their distribution is continuous (a) or discontinuous (b). Reproduced with permission from [139], distribution is continuous (a) or discontinuous (b). Reproduced with permission from [139], Elsevier, Elsevier, 2014. 2014.
These findings findings are interesting interesting from from aa mechanistic mechanistic perspective perspective and and might might be be generalizable generalizable for alloying elements elements with with similar similar electronegativity electronegativity (I (Inn). Yet the extensive knowledge obtained with Al Al alloys Long term effects of of exposure to Al reveals Al alloys is is not notdirectly directlyapplicable applicablefor forbiomedical biomedicalimplants. implants. Long term effects exposure to Al reveals to affecting the reproductive ability [140], inducing dementia [141] and leading to Alzheimer’s Albetotoxic, be toxic, affecting the reproductive ability [140], inducing dementia [141] and leading to disease [142,143]. Alzheimer’s disease [142,143]. 5.2.2. 5.2.2. Manganese Manganese As As aa binary binary addition addition to to Mg, Mg, Mn Mn shows shows no no significant significant increase increase on on corrosion corrosion for for concentrations concentrations of of up to 5 wt % [129]. It increases the Fe tolerance level in Mg-Al alloys when keeping the Fe/Mn ratio up to 5 wt % [129]. It increases the Fe tolerance level in Mg-Al alloys when keeping the Fe/Mn ratio at at 0.032. 0.032. The The most most likely likely mechanism mechanism explaining explaining the the moderation moderation of of Fe Fe is is that that Fe Fe is is incorporated incorporated in in an an intermetallic AlMnFe compound, which is less active as local cathode, and thus reduces microgalvanic intermetallic AlMnFe compound, which is less active as local cathode, and thus reduces coupling [136,144]. However, Mn in concentrations than 10 µmol/L in the10blood, hasinbeen microgalvanic coupling [136,144]. However, Mn in higher concentrations higher than μmol/L the shown induce a neurological disorder similar to Parkinson’s diseaseto [145]. blood, to has been “Manganism”, shown to induce “Manganism”, a neurological disorder similar Parkinson’s
disease [145]. 5.2.3. Zinc 5.2.3.Zn Zinc causes solid solution strengthening increasing the strength of Mg up to 280 MPa adding 6 wt % of zinc [11]. Moreover, Zn is an essential trace mineral to hundreds of biological enzymes, being Zn causes solid solution strengthening increasing the strength of Mg up to 280 MPa adding 6 wt required by human body at 15 mg/day [146,147]. The main drawback regarding Zn’s biocompatibility % of zinc [11]. Moreover, Zn is an essential trace mineral to hundreds of biological enzymes, being is the reaction of Zn2+ with hydrochloric acid (HCl). Zn2+ evolves from the oxidation reaction of Zn required by human body at 15 mg/day [146,147]. The main drawback regarding Zn’s biocompatibility used as alloying material, is the reaction of Zn2+ with hydrochloric acid (HCl). Zn2+ evolves from the oxidation reaction of Zn Zn → Zn2+ + 2e− , (17) used as alloying material, and HCl reduced according to: Zn → Zn2+ + 2e−, (17) 2HCl + 2e− → H2 + 2Cl− , (18) and HCl reduced according to: 2HCl + 2e− → H2 + 2Cl−,
(18)
Metals 2017, 7, 252 Metals 2017, 7, 252
25 of 40 25 of 41
leading then to the formation of ZnCl2, known to damage stomach parietal cells [148]. In Mg alloys, Zn enhances the tolerance limit and reduces effects of the three main impurities (Fe, Cu, Ni) when leading then to the formation of ZnCl , known to damage stomach parietal cells [148]. In Mg alloys, their solid solubility limits have been 2exceeded. 1 wt % of Zn in Mg raises the tolerance limit for Ni Zn enhances the tolerance limit and reduces effects of the three main impurities (Fe, Cu, Ni) when their [129]. Song and Atrens [77] reported the Ni tolerance limit to be increased up to 20 ppm in Mg-Alsolid solubility limits have been exceeded. 1 wt % of Zn in Mg raises the tolerance limit for Ni [129]. Mn alloys with an addition of 3 wt % Zn, and it can further reduce the corrosion rate of ternary alloys Song and Atrens [77] reported the Ni tolerance limit to be increased up to 20 ppm in Mg-Al-Mn alloys when Ni and Fe tolerance levels are reached. At Zn concentration above 3 wt %, second phases form with an addition of 3 wt % Zn, and it can further reduce the corrosion rate of ternary alloys when Ni and the corrosion resistance lowers [67] leading to localized corrosion [149]. Another advantage of and Fe tolerance levels are reached. At Zn concentration above 3 wt %, second phases form and the Zn alloying is the decrease in hydrogen evolution along with the decrease in solubility of the Mg corrosion resistance lowers [67] leading to localized corrosion [149]. Another advantage of Zn alloying matrix. Both, Mg2+ ions and Zn2+ ions bind with free OH− anions forming Zn(OH)2 and reducing the is the decrease in hydrogen evolution along with the decrease in solubility of the Mg matrix. Both, amounts of free H2 [85]. However, in crystalline Mg, the solubility of alloying elements is limited, Mg2+ ions and Zn2+ ions bind with free OH− anions forming Zn(OH)2 and reducing the amounts hydrogen evolution can hence only be slightly reduced. Mg-based glasses, in contrast, offer increased of free H [85]. However, in crystalline Mg, the solubility of alloying elements is limited, hydrogen solubility2 for alloying elements, allowing to significantly lower corrosion. Zberg et al. [80] evolution can hence only be slightly reduced. Mg-based glasses, in contrast, offer increased solubility investigated glassy Mg60+xZn35−xCa5 alloys (x = 0, 3, 6, 7, 9, 12, 14, 15) in SBF. They found that an for alloying elements, allowing to significantly lower corrosion. Zberg et al. [80] investigated glassy increase in Zn reduces hydrogen evolution, with a distinct drop in hydrogen release at a Zn content Mg60+x Zn35−x Ca5 alloys (x = 0, 3, 6, 7, 9, 12, 14, 15) in SBF. They found that an increase in Zn reduces of 28 wt % (Figure 26). hydrogen evolution, with a distinct drop in hydrogen release at a Zn content of 28 wt % (Figure 26).
Figure 26. Hydrogen evolution evolution measurement measurement for for MgZnCa MgZnCa alloys alloys with with different different content content of of Zn. Zn. Figure 26. Hydrogen Reproduced with permission from [80], Nature Publishing Group, 2009. Reproduced with permission from [80], Nature Publishing Group, 2009.
5.2.4. Zirconium Zirconium 5.2.4. Zirconium (Zr) (Zr) can can refine refine the the grain Li et et al. al. reported reported the the microstructure microstructure of of Mg-Zr Mg-Zr binary binary Zirconium grain size. size. Li alloys to be finer than pure Mg [150]. The grain size can be reduced to 50 µm. Zr alloys also have alloys to be finer than pure Mg [150]. The grain size can be reduced to 50 μm. Zr alloys also have good corrosion resistance due to improved castability. Zr reacts with the impurities of molten Mg good corrosion resistance due to improved castability. Zr reacts with the impurities of molten Mg lowering their their levels levels through through precipitations. precipitations. For For example, example, Fe Fe combines combines with with Zr Zr to to form form Fe Fe22Zr, Zr, which, which, lowering by gravity, settles to the bottom of the melt improving the purity of the cast [77]. Song and St. John by gravity, settles to the bottom of the melt improving the purity of the cast [77]. Song and St. [151] John compared the corrosion resistance of rare of earth alloys (Mg-REZn-Zr) with andwith without [151] compared the corrosion resistance rare(RE)-containing earth (RE)-containing alloys (Mg-REZn-Zr) and Zr addition, labelled labelled with MEZ MEZ (Figure 27). r and u , respectively without Zr addition, with MEZ r and MEZu, respectively (Figure 27). They related relatedthe thehigher highercorrosion corrosion resistance of MEZ to dissolution the dissolution ofimpurities. iron impurities. They resistance of MEZ r to rthe of iron From From composition analysis of the two they alloys, theyMEZ found MEZ(0.005 % Zr) to contain u alloys composition analysis of the two alloys, found u alloys wt(0.005 % Zr) wt to contain 0.0013 wt 0.0013 wt % Fe, whereas alloys characterized by 0.6 wt % Zr have lower Fe (0.004 wt %). However, % Fe, whereas alloys characterized by 0.6 wt % Zr have lower Fe (0.004 wt %). However, purification purification cannot the in difference in corrosion centres andofedges MEZR cannot explain the explain difference corrosion resistanceresistance between between centres and edges MEZof R grains grains (Figure (Figure 28). 28). Song provided threethree possible explanations of theseofobservations. First, theyFirst, attributed Song and andSt.St.John John provided possible explanations these observations. they the higherthe corrosion of the grain to acentre highertoamount Zr withofrespect torespect the grain attributed higher resistance corrosion resistance of centre the grain a higherofamount Zr with to boundaries. Comparing MEZU andMEZ MEZ , they found the onset of corrosion of to the grain boundaries. Comparing U Rand MEZ R, they foundand thepropagation onset and propagation be slower in the latter suggesting that Zr in solid solution improves the resistance via solubility corrosion to be slower in the latter suggesting that Zr in solid solution improves the resistance via reduction reduction of Mg in the sameinmedia. Second, they observed decreased anumber of precipitated solubility of Mg the same media. Second, theya observed decreased number of
Metals 2017, 7, 252 Metals 2017, 7, 252 Metals 2017, 7, 252
26 of 40 26 of 40 26 of 41
precipitated particles producing hydrogen in MEZR than in MEZu suggesting a reduction of the precipitated particles producing hydrogen in MEZR than in MEZu suggesting a reduction of the cathodic activity of intermetallic precipitates containing Zr. Third, the authors stated that the grain cathodicproducing activity ofhydrogen intermetallic precipitates Zr. Third, the authors stated that the grain particles in MEZ MEZu suggesting a reduction of the cathodic activity R than incontaining refinement effect of Zr provides more continuous layers of corrosion resistant intermetallic RE phases effect of Zr provides more continuous layers of corrosion resistant intermetallic RE phases ofrefinement intermetallic precipitates containing Zr. Third, the authors stated that the grain refinement effect around grain boundaries, decelerating corrosion between grains. grain boundaries, decelerating corrosion between grains. ofaround Zr provides more continuous layers of corrosion resistant intermetallic RE phases around grain boundaries, decelerating corrosion between grains.
Figure 27. 27. Weight loss loss rate for for MEZuu and and MEZrr specimens. Modified from [151]. Figure Figure 27.Weight Weight lossrate rate forMEZ MEZu andMEZ MEZrspecimens. specimens.Modified Modifiedfrom from[151]. [151].
28. Optical Figure 28. Optical micrographs micrographs of of MEZ MEZR surface surface after after immersion immersion in in chloride chloride solution solution for for 33 h. h. Figure 28. Optical micrographs of MEZRR surface after immersion in chloride solution for 3 h. permission from from [151], [151], Elsevier, Elsevier,2002. 2002. Reproduced with permission Reproduced with permission from [151], Elsevier, 2002.
5.2.5. Calcium 5.2.5. Calcium in in human bones and and is beneficial for bone and growth Ca is is aamajor majorcomponent component human bones is beneficial for healing bone healing and Ca is a major component in human bones and is beneficial for bone healing and growth [127,139,152]. It is widely utilized as alloying element for for biodegradable Mg growth [127,139,152]. It is widely utilized as alloying element biodegradable Mgalloys. alloys.When WhenCa Ca is [127,139,152]. It is widely utilized as alloying element for biodegradable Mg alloys. When Ca is present, Mg Mg alloys alloysdevelop developaahydroxyapatite hydroxyapatite(HA) (HA)surface surfacelayer, layer,improving improving biocompatibility. present, itsits biocompatibility. CaCa is present, Mg alloys develop a hydroxyapatite (HA) surface layer, improving its biocompatibility. Ca is also a grain refiner improving mechanical properties. and[154] Yang [154] reported also a grain refiner [153][153] improving mechanical properties. Zhang Zhang and Yang reported a decreasea is also a grain refiner [153] improving mechanical properties. Zhang and Yang [154] reported a decrease in grain from 175 in to Mg-Zn-Mn 51 μm in Mg-Zn-Mn alloys increasing theofamount Catofrom in grain size from size 175 to 51 µm alloys increasing the amount Ca fromof0.3 1 wt0.3 %. decrease in grain size from 175 to 51 μm in Mg-Zn-Mn alloys increasing the amount of Ca from 0.3 to 1 solubility wt %. Thelimit solubility limit of Ca only is however wt %,the after which the corrosion resistance The of Ca is however 1 wt %,only after1which corrosion resistance drops due to to 1 wt %. The solubility limit of Ca is however only 1 wt %, after which the corrosion resistance drops due to the development of Mg 2Ca second [155]. Yet et al. [156] assumedof the the development of Mg2 Ca second phases [155]. phases Yet Bornapour et Bornapour al. [156] assumed the presence a drops due to the development of Mg2Ca second phases [155]. Yet Bornapour et al. [156] assumed the presence of a second phase as being desirable.the Comparing corrosionofbehaviour of that pureofMg to that second phase as being desirable. Comparing corrosionthe behaviour pure Mg to different presence of a second phase as being desirable. Comparing the corrosion behaviour of pure Mg to that of different types alloys with Ca andMg-0.6Ca, Sr (Mg-0.5Sr, Mg-0.6Ca, Mg-0.5Sr-0.6Ca, types of alloys withofCa and Sr (Mg-0.5Sr, Mg-0.5Sr-0.6Ca, Mg-0.3Sr-0.3Ca),Mg-0.3Sr-0.3Ca), they found that of different types of alloys with Ca and Sr (Mg-0.5Sr, Mg-0.6Ca, Mg-0.5Sr-0.6Ca, Mg-0.3Sr-0.3Ca),
Metals 2017, 7, 252
27 of 41
Metals 2017, 7, 252
27 of 40
they found that Mg-0.3Sr-0.3Ca alloys have the highest corrosion resistance, despite the presence of
Mg-0.3Sr-0.3Ca alloys have the highest corrosion resistance, despite the presence of Ca/Sr-rich second Ca/Sr-rich second phases (Figure 29). phases (Figure 29).
◦ C. Reproduced with FigureFigure 29. Corrosion raterate of pure Mg, binary in SBF SBFatat3737°C. 29. Corrosion of pure Mg, binaryand andternary ternary alloys alloys in Reproduced with permission from from [156],[156], Elsevier, 2014. permission Elsevier, 2014.
They related the improvement in the corrosionbehaviour behaviour to these phases in both They related the improvement in the corrosion to the thepresence presenceofof these phases in both the grain boundaries and in the interior aligning the corrosion potential of the matrix and the grain the grain boundaries and in the interior aligning the corrosion potential of the matrix and the grain boundaries and hence lowering dissolution. Moreover they state that small amounts of Sr and Ca boundaries and hence lowering dissolution. Moreover they state that small amounts of Sr and Ca result in stable HA surface layers, whereas large amounts may result in a non-adherent HA surface result layer in stable HAeasily surface layers, whereas large amounts may result in a non-adherent HA surface that can detach losing its electrochemical barrier function. layer that can easily detach losing its electrochemical barrier function. 5.2.6. Rare Earth (RE) Elements
5.2.6. Rare Earth (RE) Elements
RE are mostly present as additions to Zr containing alloys to improve their mechanical
RE are mostly present[157] as reported additions to Zrresistance containing alloys improve theirofmechanical properties [75]. Rokhlin a higher against SCCto with the addition Cd and Nd to[75]. Mg-Zn-Zr alloys. A lower sensitivity to SCC using Ndagainst has beenSCC obtained et al.of[158]. properties Rokhlin [157] reported a higher resistance withby theKannan addition Cd and In their study, EV31A, composed by 2.35 wt % Nd and 1.3 wt % Gd, has a higher SCC susceptibility Nd to Mg-Zn-Zr alloys. A lower sensitivity to SCC using Nd has been obtained by Kannan et al. [158]. index (ISCC ) than AZ80 in chloride solution when ratesa higher of 10−6 SCC s−1 and 10−7 s−1, In their study, EV31A, composed by 2.35 wt % Nd andtested 1.3 wtat%strain Gd, has susceptibility SCC susceptibility index is calculated based on particular index respectively (ISCC ) than (Table AZ80 4). in The chloride solution when tested at strain rates of 10−6 s−1 mechanical and 10−7 s−1 , properties, e.g. ultimate tensile strength (UTS) and elongation to failure (εf), measured in a SSRT test respectively (Table 4). The SCC susceptibility index is calculated based on particular mechanical in corrosive environments and compared to its corresponding value in air. A low ISCC index properties, e.g. ultimate strength (UTS) andthis elongation to failureunity, (εf ), measured a SSRT test in corresponds to hightensile SCC susceptibility. When index approaches the alloy isinmeant to be corrosive environments and compared to its corresponding value in air. A low I index corresponds SCC highly resistant to SCC in that particular medium. to high SCC susceptibility. When this index approaches unity, the alloy is meant to be highly resistant Table 4. Imedium. SCC indices for different alloys at different strain rate. Modified from [158]. to SCC in that particular 10−6 s−1 Strain Rate
10−7 s−1 Strain Rate
Table 4. ISCC indices for different alloys Alloy ISCC at different strain ISCCrate. Modified from [158].
Alloy
εf UTS 1 Strain Rate 10−6 s−0.35 AZ80 0.83 ZE41 0.35 0.80 ISCC QE22 0.12 0.82 εf 0.70 UTS EV31A 0.85
εf 0.0910−7 0.17 0.05 0.67 εf
UTS
−1 Strain Rate s0.62
0.66 ISCC 0.65 0.85
UTS
AZ80 0.35 0.83 0.09 0.62 The improved corrosion behaviour of EV31A development of0.66 a film of mixed oxides ZE41 0.35 0.80is due to the 0.17 of Nd and Gd, which film of 0.05 RE free alloys. 0.65 However, despite that QE22 is more stable 0.12 than the surface 0.82 EV31A 0.70 0.85 0.67 0.85
The improved corrosion behaviour of EV31A is due to the development of a film of mixed oxides of Nd and Gd, which is more stable than the surface film of RE free alloys. However, despite that both
Metals 2017, 7,7, 252 Metals 2017, 252 Metals 2017, 7, 252
2828 ofof 4040 28 of 41
both bothZE41 ZE41and andQE22 QE22contain containRE REelements, elements,they theyhave havelow lowcorrosion corrosionresistance. resistance.Kannan Kannanetetal.al.[158] [158] attributed their lower corrosion properties to other elements, such as high Zn content (ZE41), and attributed theircontain lower corrosion properties to other elements,resistance. such as high Zn content (ZE41), andAg Ag ZE41 and QE22 RE elements, they have low corrosion Kannan et al. [158] attributed with higher noble potential than Mg (QE22). withlower higher noble potential than (QE22). their corrosion properties to Mg other elements, such as high Zn content (ZE41), and Ag with higher Nd isisalso Nd alsoreported reported toimprove improvethe thecorrosion corrosionresistance resistanceofofMg-Al. Mg-Al.Zhang Zhangetetal.al.[159] [159]studied studiedthe the noble potential than Mg to (QE22). effect of Nd on the corrosion behaviour of AZ91 finding a drop in corrosion rate when 1.5 wt % of effect ofisNd onreported the corrosion behaviour of AZ91 resistance finding a drop in corrosion 1.5 wt % Nd also to improve the corrosion of Mg-Al. Zhang etrate al. when [159] studied theof Nd 30). They this improvement Nd isis added (Figure 30).behaviour They related related this improvement to a a modification modification of the the%alloy’s alloy’s effect ofadded Nd on (Figure the corrosion of AZ91 finding a drop into corrosion rate whenof 1.5 wt of Nd microstructure, Al-Mn phases minimizing ofofgalvanic sinceThey Al-Mn phases arereplaced replacedby byaAl-Nd Al-Nd minimizing theeffect effect galvanic ismicrostructure, added (Figuresince 30). related thisare improvement to modification of thethe alloy’s microstructure, coupling. coupling. since Al-Mn phases are replaced by Al-Nd minimizing the effect of galvanic coupling.
Figure Figure30. 30.Corrosion Corrosionrate rateof ofAZ91 AZ91with withand andwithout withoutNd. Nd.Reproduced Reproducedwith withpermission permissionfrom from[159], [159], Figure 30. Corrosion rate of AZ91 with and without Nd. Reproduced with permission from [159], Elsevier, 2011. Elsevier, Elsevier,2011. 2011.
When Al, RE reduce ductility and strength due formation ofofAlAl xRE y, ywhich Whencombined combinedwith with and strength dueto tothe the RE , which When combined with Al, Al,RE REreduce reduceductility ductility and strength due to formation the formation xof Al x REy , isishowever connected totoananimprovement inincorrosion resistance. For example, binary addition ofofLaLa however connected improvement corrosion resistance. For example, binary addition which is however connected to an improvement in corrosion resistance. For example, binary addition isisdetrimental for corrosion resistance since Mg 12La phases above the detrimental forthe the resistance sinceitsince itforms forms Mg 12Lacathodic cathodic phases aboveabove thesolid solid of La is detrimental forcorrosion the corrosion resistance it forms Mg12 La cathodic phases the solubility limit. As elemental addition to Mg-Al, however, it improves corrosion resistance. Liu etetal.al. solubility limit. As elemental addition to Mg-Al, however, it improves corrosion resistance. Liu solid solubility limit. As elemental addition to Mg-Al, however, it improves corrosion resistance. [160] reported a adrop rate 0.5 is added to AZ91 (Figure 31), [160] reported dropinincorrosion ratewhen whenrate 0.5wt wt%%LaLa added (Figure 31),which whichthey they Liu et al. [160] reported a corrosion drop in corrosion when 0.5iswt % LatoisAZ91 added to AZ91 (Figure 31), related to a modification of the alloy’s microstructure. The formation of needle-like Al-La compounds related to a modification of the alloy’s microstructure. The formation of needle-like Al-La compounds which they related to a modification of the alloy’s microstructure. The formation of needle-like Al-La and of Mg 17Al 12 12 from discontinuous totocontinuous andthe thealteration alteration Mg 17Al from continuousisisobserved. observed. compounds and theofalteration of Mgdiscontinuous 17 Al12 from discontinuous to continuous is observed.
Figure Figure31. 31.Corrosion Corrosioncurrent currentdensities densities(i(iH(i)H))ofof ofAZ91 AZ91alloy alloyvaried variedwith withdifferent differentimmersion immersiontime. time. Figure 31. Corrosion current densities AZ91 alloy varied with different immersion time. H Reproduced with permission from [160], John Wiley and Sons, 2009. Reproduced with permission from [160], John Wiley and Sons, 2009. Reproduced with permission from [160], John Wiley and Sons, 2009.
Zhou Zhouetetal.al.[161] [161]studied studiedthe theeffect effectofofadding adding0.24 0.24wt wt%%Ho Hoand and0.44 0.44wt wt%%Ho HototoAZ91D AZ91Dalloys, alloys, Zhou et al. [161] studied the the effect of(Figure adding32). 0.24 wt % Ho and 0.44 wt % Ho to AZ91D alloys, which also significantly decreases rate which also significantly decreases the rate (Figure 32). which also significantly decreases the rate (Figure 32).
Metals Metals 2017, 2017, 7, 7, 252 252
29 29 of of 41 40
Figure from [161], Figure 32. 32. Corrosion Corrosion rate rate of of AZ91D AZ91D with with and and without without Ho. Ho. Reproduced Reproduced with with permission permission from [161], Elsevier, Elsevier, 2006. 2006.
Both Both Fe Fe and and Mg17Al12 Mg17Al12 volume volume fractions are reduced by Ho Ho due due to to the the formation formation of of Al-Ho Al-Ho intermetallic a lower microgalvanic corrosion through a loweradifference in potential. intermetallic phases phasesinducing inducing a lower microgalvanic corrosion through lower difference in The corrosive of Ho containing alloys is more protective to aprotective higher Al concentration. Finally, potential. Thefilm corrosive film of Ho containing alloys isdue more due to a higher Al Yao et al. [162] reported the Sc addition to AZ91Ethe alloys, which refines the microstructure by means concentration. Finally, Yao et al. [162] reported Sc addition to AZ91E alloys, which refines the of Al3 Sc formation suppressing and reducing corrosion rate fromthe 8 µA cm−2 microstructure by means of Al3ScMg formation suppressing Mg17Al12the phases and reducing corrosion 17 Al12 phases 2 with to 2 µA cm8−μA 0.12wt Sc−2addition. This that the addition of RE beneficial toof reduce rate from cm−2a to μA%cm with a 0.1 wt shows % Sc addition. This shows thatisthe addition RE is corrosion relatedcorrosion mechanical thefailures, synthesis of RE elements is expensive leading beneficialand to reduce andfailures, related however, mechanical however, the synthesis of RE elements is expensive leading to high production costs, to high production costs, which limits their use.which limits their use. 5.3. 5.3. Surface Surface Treatment Treatment There There are are two two possible possible ways ways to to improve improvethe thecorrosion corrosionresistance resistanceof ofMg Mgand andits itsalloys: alloys: (1) Tailor composition (1) Tailor composition and and microstructure; microstructure; (2) Treat the surface/apply coatings. (2) Treat the surface/apply coatings. (1) (1) can can be be achieved achieved through through alloying. alloying. Here Here we we therefore therefore focus focus on on (2) (2) via via biomedical biomedical coatings coatings and and their influence on corrosion. Due to Mg’s high chemical reactivity, coating is a viable strategy their influence on corrosion. Due to Mg’s high chemical reactivity, coating is a viable strategy for for both both increased increased biocompatibility biocompatibility and and corrosion corrosion protection, protection, allowing allowing aa wide wide range range of of possible possible non-toxic non-toxic and and fully degradable surface modifications. Mg alloys tend to passivate naturally forming a protective fully degradable surface modifications. Mg alloys tend to passivate naturally forming a protective layer MgO (2.5 nm) andand an outer Mg(OH) (2.2 nm) [163]. This layer is not stable 2 layer layer stack stackofofananinner inner MgO (2.5 nm) an outer Mg(OH) 2 layer (2.2 nm) [163]. This layer is not in chloride solution.solution. Increasing the thickness the passivating layer through thermal treatments is stable in chloride Increasing the of thickness of the passivating layer through thermal reported to be an effective method to improve the corrosion behaviour. Hanzi et al. [164] evaluated treatments is reported to be an effective method to improve the corrosion behaviour. Hanzi et al. the hydrogen evolution rate of WE43 alloy rate under surface conditions, obtained means of [164] evaluated the hydrogen evolution ofdifferent WE43 alloy under different surfaceby conditions, various treatments (Figureheat 33):treatments (Figure 33): obtainedheat by means of various
•• •• ••
◦ C for 6 h and then they were grinded and polished (labelled Samples heat-treated at 525 Sampleswere werefirst first heat-treated at 525 °C for 6 h and then they were grinded and polished as SHT); (labelled as SHT); ◦ ◦ C for Samples 6h then they were artificially aged at 250 Samples were werefirst firstheat-treated heat-treatedatat525 525C°Cfor for 6 and h and then they were artificially aged at 250 °C 16 h before being grinded and polished (T6); for 16 h before being grinded and polished (T6); ◦ C, being covered by an oxide layer during Specimens Specimens were were first first polished, polished, then then annealed annealed at at 500 500 °C, being covered by an oxide layer during the the annealing. annealing. The The oxidation oxidation in in air air during during the the heat heat treatment treatment was was carried carried out out for for various various times, times, i.e., 1 (ox1), 8 (ox8), 24 (ox24), 48 (ox48), 168 h (ox168). i.e., 1 (ox1), 8 (ox8), 24 (ox24), 48 (ox48), 168 h (ox168).
After After the the third third procedure, procedure,an anoxide oxidelayer, layer,made madeby byMgO MgOand andYY22O O33,, covers covers the the surface surface because because of of the heat treatment, and the latter growths with increasing oxidation duration, from 500 nm after 1 h to the heat treatment, and the latter growths with increasing oxidation duration, from 500 nm after 1 h 2700 nmnm after 168168 h. h. They reported that to 2700 after They reported thatthe thehigher higherthe thethickness thicknessofofthe theprotective protectivefilm filmis, is,the the slower slower and more homogeneous the degradation will be. and more homogeneous the degradation will be.
Metals 2017, 2017, 7, 7, 252 252 Metals
30 of of 40 41 40 30
◦ C after Figure 33. Hydrogen Hydrogen evolution of WE43 specimens in SBF at 37 °C after various heat treatment. Figure Figure 33. Hydrogenevolution evolutionof of WE43 WE43specimens specimensin inSBF SBF at at 37 37 °C after various various heat heat treatment. treatment. Reproduced with permission from [164], Elsevier, 2009. Reproduced with permission from [164], Elsevier, 2009.
Interestingly, the degradation rate highly increases once the protective surface layer is Interestingly, the degradation rate highly increases once the protective surface layer is penetrated penetrated or removed, (Figure 34), compared to heat treated and polished (SHT) and heat treated, or removed, (Figure 34), compared to heat treated and polished (SHT) and heat treated, aged and aged and polished (T6) samples. polished (T6) samples.
Figure 34. Maximum hydrogen evolution rate for different WE43 samples. Reproduced with Figure 34. 34. Maximum Maximum hydrogen hydrogen evolution evolution rate rate for for different different WE43 WE43 samples. samples. Reproduced Reproduced with with Figure permission from [164], Elsevier, 2009. permission from from [164], [164], Elsevier, Elsevier, 2009. 2009. permission
The reason can oxidation results in can be The be found found in inaacomposition compositionchange changedue duetotothe theoxidation. oxidation.Short Short oxidation results Y-depletion underneath the the oxide layer,layer, thus lowering the amount of Y inof Mg. an effective in Y-depletion underneath oxide thus lowering the amount Y However, in Mg. However, an coating strategy preserve the desired of the material Thus Calcium effective coating must strategy must preserve theproperties desired properties of the underneath. material underneath. Thus Phosphate (Ca-P such as dicalcium phosphate (DCPD) (DCPD) and hydroxyapatite (HA)) have been studied. Calcium Phosphate (Ca-P such as dicalcium phosphate and hydroxyapatite (HA)) have been They haveThey gained largegained interest in biomedical due toapplications their intrinsicdue biocompatibility related studied. have large interestapplications in biomedical to their intrinsic to their analogy to the inorganic component of natural bones that aidof thenatural bone growth [165]. biocompatibility related to their analogy to the inorganic component bones that aidCa-P the coatings can be obtained by means of several coating methods, either using conversion or deposition bone growth [165]. Ca-P coatings can be obtained by means of several coating methods, either using methods (for review andmethods a detailed description these methods the readersofare referred to [166]) conversion oradeposition (for a review of and a detailed description these methods the and improve corrosion resistance and biocompatibility. Song et al. [167] fabricated three different readers are referred to [166]) and improve corrosion resistance and biocompatibility. Song et al. [167] kinds of Ca-P coatings onkinds Mg-Zn i.e., DCPD, HA andalloys, fluoridated hydroxyapatite (FHA) via fabricated three different of alloys, Ca-P coatings on Mg-Zn i.e., DCPD, HA and fluoridated electrodeposition and compared their effectsand on compared the degradation behaviour indegradation m-SBF (Figure 35). hydroxyapatite (FHA) via electrodeposition their effects on the behaviour in m-SBF (Figure 35).
Metals Metals2017, 2017,7,7,252 252
31 31of of40 41
Figure 35. Hydrogen evolution of Mg-Zn alloys with and without Ca-P coatings in m-SBF. Reproduced Figure 35. Hydrogen evolution of Mg-Zn alloys with and without Ca-P coatings in m-SBF. with permission from [167], Elsevier, 2010. Reproduced with permission from [167], Elsevier, 2010.
Theresults resultsshow showthat thatcoatings coatingscan cansignificantly significantlydecrease decreasethe thedegradation degradationrate rateof ofMg-Zn Mg-Znalloys alloys The actingas asan an electrochemical electrochemicalbarrier barrierdelaying delayingthe theMg Mgalloys’ alloys’corrosion. corrosion.Wang Wangetetal. al.[168] [168] compared compared acting the effect of a Ca-deficient hydroxyapatite coating on the degradation behaviour of Mg-Zn-Ca alloys the effect of a Ca-deficient hydroxyapatite coating on the degradation behaviour of Mg-Zn-Ca alloys − 5 −5 inSBF. SBF.Performing Performingslow slowstrain strainrate rate tensile (SSRT) tests at 2.16 a beneficial in tensile (SSRT) tests at 2.16 × 10× 10 mm/s,mm/s, a beneficial effect ofeffect the of the coating SCC resistance has been shown, increasing ultimate tensile strength (UTS) coating on SCC on resistance has been shown, increasing the ultimatethe tensile strength (UTS) and time of and time of fracture (TOF) by 5.6% and 16.6%, respectively. Zhu et al. [169] studied the corrosion fracture (TOF) by 5.6% and 16.6%, respectively. Zhu et al. [169] studied the corrosion resistance of a resistance of a hydroxyapatite/aminated hydroxyethyl cellulose (HA/AHEC) alloy hydroxyapatite/aminated hydroxyethyl cellulose (HA/AHEC) coated AZ31 alloycoated in SBF.AZ31 Uncoated in SBF. Uncoated coated samples were days immersed for sevenevolution days andwas hydrogen evolution and coated samplesand were immersed for seven and hydrogen monitored. They was monitored. reported that the coating greatly improves theofcorrosion resistance of AZ31 reported that theThey coating greatly improves the corrosion resistance AZ31 alloys, leading to a 2/day. alloys, leading a reduction in therelease average hydrogen rate from of about i.e.,mL/cm from 0.92 to reduction in thetoaverage hydrogen rate of aboutrelease 65%, i.e., 0.9265%, to 0.31 2 /day. Moreover, they also studied the cytocompatibility of the coating, investigating the 0.31 mL/cm Moreover, they also studied the cytocompatibility of the coating, investigating the MC3T3-E1 cellular MC3T3-E1reporting cellular response, reporting coated no alloy to present no cytotoxic reaction to MC3T3-E1 response, the coated alloy tothepresent cytotoxic reaction to MC3T3-E1 cells and to cells and to significantly enhance their proliferation rate. Yang chemically et al. [170] chemically coated significantly enhance their proliferation rate. Yang et al. [170] coated AZ31 rodsAZ31 with rods Cawith Ca-P implanting them into the thighbone of rabbits to assess the changes in biocompatibility and P implanting them into the thighbone of rabbits to assess the changes in biocompatibility and biodegradation provided After 8 weeks of implantation, coated samples show ashow slower biodegradation provided by bythe thecoating. coating. After 8 weeks of implantation, coated samples a biodegradation than uncoated while inducing fast formation new boneofaround the implants. slower biodegradation than uncoated while ainducing a fast of formation new bone around the Chiu et al. [171] studied the effect of fluoride containing (MgF2 ) coatings on pure Mg (99.96 wt %) implants. immersed Hank’s balanced salt solution, whichcontaining mimics ph(MgF values of the human body. means Chiu etinal. [171] studied the effect of fluoride 2) coatings on pure Mg By (99.96 wt of a conversion treatment, they cover the Mg samples surface with a 1.5 µm thick MgF protective %) immersed in Hank’s balanced salt solution, which mimics ph values of the human body.2 By means layer. After 18 treatment, days of immersion, reported an surface averagewith corrosion ratethick of 1.01 mm/year for of a conversion they coverthey the Mg samples a 1.5 μm MgF 2 protective coated specimens, 3.70 mm/year uncoated Mg, respectively. Witte al. mm/year [172] confirmed this layer. After 18 days and of immersion, they for reported an average corrosion rate ofet1.01 for coated studying implants LAE442 with andMg, without MgF2 coatings, into the this medial femur specimens, and 3.70 from mm/year for uncoated respectively. Witte etimplanted al. [172] confirmed studying condyle of rabbits. Thewith coating (Figure 36) and, into untilthe it disappears, thecondyle releaseof of implants from LAE442 and reduces without mass MgF2 loss coatings, implanted medial femur alloyingThe elements. rabbits. coating reduces mass loss (Figure 36) and, until it disappears, the release of alloying elements.
Metals Metals 2017, 2017, 7, 7, 252 252
32 32 of of 41 40
Figure Figure36. 36 Implant volume of LAE442 and and MgF MgF22-coated LAE442 at different postoperative intervals. Reproduced with permission from [172], Elsevier, Reproduced with permission from [172], Elsevier,2010. 2010.
All reported reported implants implants are are affected affected by by localized localized corrosion corrosion leading leading to to pitting pitting once once the the protective protective All layer breaks down. Further, the fluoride-containing coatings seem to irritate the local synovial tissue layer breaks down. Further, the fluoride-containing coatings seem to irritate the local synovial tissue during their theirdissolution. dissolution. Organic coatings, especially biopolymers, a viable during Organic coatings, especially organicorganic biopolymers, provide aprovide viable alternative alternative offering functionalization possibilities with organic biomolecules. Polylactic-co-glycolic offering functionalization possibilities with organic biomolecules. Polylactic-co-glycolic acid (PLGA) acid (PLGA) great demonstrates great and cell proliferation adhesion andproperties proliferation [13]. performed Li et al. [173] demonstrates cell adhesion [13]. properties Li et al. [173] cell performed cell attachment tests utilizing mouse osteoblast-like cells MC3T3-E1 on Mg-6Zn alloys attachment tests utilizing mouse osteoblast-like cells MC3T3-E1 on Mg-6Zn alloys with and without with and without PLGA coating. The coated alloys possess a significantly ability for PLGA coating. The coated Mg alloys possessMg a significantly enhanced abilityenhanced for cell attachment cell attachment compared to the uncoated one (Figure 37). compared to the uncoated one (Figure 37).
Figure Figure 37. 37. SEM SEM micrographs micrographs of of cell cell morphology morphology after after various various culture culture times times on on polylactic-co-glycolic polylactic-co-glycolic acid (PLGA) coated Mg alloys after (a) 1 day; (b) 2 days and (c) 3 days culture acid (PLGA) coated Mg alloys after (a) 1 day; (b) 2 days and (c) 3 days culture and and uncoated uncoated PLGA PLGA after days culture. culture. Reprinted Reprinted with with permission permission from after (d) (d) 11 day; day; (e) (e) 22 days days and and (f) (f) 33 days from [173], [173], Springer, Springer, 2010. 2010.
Moreover, they they also also reported reported the the PLGA PLGA coating coating to to offer offer protection protection against against corrosion corrosion since since itit Moreover, significantly reduces reduces the the degradation degradationrate rateof ofthe theMg Mgalloy alloy(Table (Table5). 6). significantly
Metals 2017, 7, 252
33 of 41
Table 5. Average degradation rate of naked and coated Mg6Zn alloys after 72 h and 144 h of immersion in 0.9% NaCl solution at 37 ◦ C [173].
Immersion Time (h) 72 144
Average Degradation Rate (mg/cm2 /h) Uncoated
Coated
0.063 0.161
2.44 × 10−4 3.23 × 10−4
6. Conclusions and Outlook In some biomedical applications, such as plates, screws and wires, temporary devices have continuously gained interest in the last years to render cumbersome second surgeries after bone healing unnecessary. Among all biodegradable materials, Mg has attracted great research interest as material for temporary implants since it possesses one of the best mechanical compatibilities with human bone. Its degradation products (Mg2+ ions) are essential to the human metabolism. However, the mechanical properties of Mg are too low to make it suitable for implants and hence researchers and clinicians focus on Mg alloys, although related clinical applications are still limited due to several factors. One of the main limitations is the low corrosion resistance inducing a loss of mechanical integrity in human body fluid before bone and surrounding tissues have sufficient time to heal. Moreover, since implants are subjected to different acute loadings during their use, Mg alloys must possess adequate resistance to cracking under the simultaneous action of the corrosive human body fluid and the tensile or cyclic test, i.e., resistance to stress corrosion cracking and corrosion fatigue. Few biomedical relevant studies of corrosion assisted cracking resistance are available since almost all results were obtained studying Mg-Al alloys, not suitable for biomedical devices due to the harmful effects of Al during long term exposure. To push Mg and its alloys to clinics, we will require proper evaluation on their corrosion assisted cracking. Finally, Mg dissolution in aqueous solutions results in hydrogen evolution that has a detrimental effect on biocompatibility. If the corrosion rate is too high, the amount of hydrogen produced cannot be absorbed by the human body leading to toxic gas bubbles and ultimate implant failures. Moreover, the formation of hydroxide ions (OH− ) involved in the corrosion of Mg alloys in aqueous solutions leads to an increase in the pH of the surrounding solution, resulting in a detrimental effect on cell proliferation. Improvements for the corrosion resistance of Mg alloys have been widely studied recently. Researchers can tune the morphology of the metal, reducing impurities or utilize alloying elements that benefit mechanical and chemical properties. Further, they can employ surface modification strategies, coating Mg alloys with a corrosion resistant yet bioactive material, favouring also bone formation (e.g., Ca-P coatings). However, few reports exist on such coatings. The most employed, Ca-P, respectively HA, has poor mechanical properties and low adhesion to the Mg alloy leading to delamination [174]. This highlights the need for biocompatible surfaces and new coating methods in the future. The authors envision that key research will centre around the investigation of Mg and its alloys’ corrosion behaviour and corrosion assisted cracking resistance coated with new biocompatible coatings such as TiO2 , TiC and TiN [175]. Further, extensive research will centre on device production processes and their effect on corrosion resistance. Grain refinement might be such a strategy. Song and Atrens [77] showed that finer grains allow a nearly continuous distribution of second phases, leading to a corrosion delay as corrosion products remain near the Mg matrix, and act as barrier. Furthermore, fine grains improve the corrosion assisted cracking resistance inhibiting crack initiation and dislocations motion leading to an increase in the number of barriers to crack propagation [91]. Related production processes that lead to finer grains are hot deformation and rapid solidification. The authors expect that these will be investigated in detail, exploring also the property of rapid solidification processes to increase the solubility of the alloying elements.
Metals 2017, 7, 252
34 of 41
Acknowledgments: All authors acknowledge funding from the Faculty of Engineering at the Norwegian University of Science and Technology. Author Contributions: Mirco Peron wrote the entire manuscript. Jan Torgersen and Filippo Berto provided scientific guidance, proof reading and suggestions. Conflicts of Interest: The authors declare no conflict of interest.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
Ginebra, M.P.; Traykova, T.; Planell, J.A. Calcium phosphate cements as bone drug delivery systems: A review. J. Control. Release 2006, 113, 102–110. [CrossRef] [PubMed] Regar, E.; Sianos, G.; Serruys, P.W. Stent development and local drug delivery. Br. Med. Bull. 2001, 59, 227–248. [CrossRef] [PubMed] Greatbatch, W.; Holmes, C.F. History of implantable devices. IEEE Eng. Med. Biol. Mag. 1991, 10, 38–41. [CrossRef] [PubMed] Long, P.H. Medical Devices in Orthopedic Applications. Toxicol. Pathol. 2008, 36, 85–91. [CrossRef] [PubMed] Khan, W.; Muntimadugu, E.; Jaffe, M.; Domb, A.J. Implantable medical devices. In Focal Controlled Drug Delivery; Springer: New York, NY, USA, 2014; pp. 33–59. Long, M.; Rack, H. Titanium alloys in total joint replacement—A materials science perspective. Biomaterials 1998, 19, 1621–1639. [CrossRef] Gultepe, E.; Nagesha, D.; Sridhar, S.; Amiji, M. Nanoporous inorganic membranes or coatings for sustained drug delivery in implantable devices. Adv. Drug Deliv. Rev. 2010, 62, 305–315. [CrossRef] [PubMed] Andrade, A.P. Towards New Biocompatible Materials: From Biological Polyesters to Synthetic Copoly (ester-urethanes) for Biomedical Applications. Ph.D. Thesis, University of Bern, Bern, Switzerland, January 2003. Lutz, J.-F. Polymerization of oligo(ethylene glycol) (meth)acrylates: Toward new generations of smart biocompatible materials. J. Polym. Sci. Part A Polym. Chem. 2008, 46, 3459–3470. [CrossRef] Schinhammer, M.; Hänzi, A.C.; Löffler, J.F.; Uggowitzer, P.J. Design strategy for biodegradable Fe-based alloys for medical applications. Acta Biomater. 2010, 6, 1705–1713. [CrossRef] [PubMed] Zhang, L.-N.; Hou, Z.-T.; Ye, X.; Xu, Z.-B.; Bai, X.-L.; Shang, P. The effect of selected alloying element additions on properties of Mg-based alloy as bioimplants: A literature review. Front. Mater. Sci. 2013, 7, 227–236. [CrossRef] Dubok, V.A. Bioceramics—Yesterday, Today, Tomorrow. Powder Metall. Met. Ceram. 2000, 39, 381–394. [CrossRef] Ulery, B.D.; Nair, L.S.; Laurencin, C.T. Biomedical Applications of Biodegradable Polymers. J. Polym. Sci. B. Polym. Phys. 2011, 49, 832–864. [CrossRef] [PubMed] Ramakrishna, S.; Mayer, J.; Wintermantel, E.; Leong, K.W. Biomedical applications of polymer-composite materials: A review. Compos. Sci. Technol. 2001, 61, 1189–1224. [CrossRef] Hanawa, T. Overview of metals and applications. In Metals for Biomedical Devices; Elsevier: Amsterdam, The Netherlands, 2010; pp. 3–24. Lendlein, A.; Rehahn, M.; Buchmeiser, M.R.; Haag, R. Polymers in Biomedicine and Electronics. Macromol. Rapid Commun. 2010, 31, 1487–1491. [CrossRef] [PubMed] Pruitt, L.; Furmanski, J. Polymeric biomaterials for load-bearing medical devices. JOM J. Miner. Met. Mater. Soc. 2009, 61, 14–20. [CrossRef] Hench, L.L. Bioceramics: From Concept to Clinic. J. Am. Ceram. Soc. 1991, 74, 1487–1510. [CrossRef] Wang, M. Bioactive ceramic-polymer composites for tissue replacement. In Engineering Materials for Biomedical Applications; World Scientific: Singapore, 2004; pp. 8-1–8-29. Maitz, M.F. Applications of synthetic polymers in clinical medicine. Biosurf. Biotribol. 2015, 1, 161–176. [CrossRef] Thamaraiselvi, T.V.; Rajeswari, S. Biological evaluation of bioceramic materials—A review. Carbon 2004, 24, 172. Pound, B.G. Corrosion behavior of metallic materials in biomedical applications. II. Stainless steels and Co-Cr alloys. Corros. Rev. 2014, 32, 21–41. [CrossRef] Pound, B.G. Corrosion behavior of metallic materials in biomedical applications. I. Ti and its alloys. Corros. Rev. 2014, 32, 1–20. [CrossRef] Black, J. Biological Performance of Materials: Fundamentals of Biocompatibility; CRC Press: Boca Raton, FL, USA, 2006. Bauer, T.W.; Schils, J. The pathology of total joint arthroplasty. II. Mechanisms of implant failure. Skelet. Radiol. 1999, 28, 483–497. [CrossRef]
Metals 2017, 7, 252
26. 27. 28. 29.
30. 31. 32.
33. 34. 35.
36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50.
35 of 41
Dujovne, A.R.; Bobyn, J.D.; Krygier, J.J.; Miller, J.E.; Brooks, C.E. Mechanical compatibility of noncemented hip prostheses with the human femur. J. Arthroplast. 1993, 8, 7–22. [CrossRef] Engh, C.; Bobyn, J.; Glassman, A. Porous-coated hip replacement. The factors governing bone ingrowth, stress shielding, and clinical results. Bone Jt. J. 1987, 69, 45–55. Engh, C.A.; Bobyn, J.D. The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin. Orthop. Relat. Res. 1988, 231, 7–28. [CrossRef] Kerner, J.; Huiskes, R.; van Lenthe, G.H.; Weinans, H.; van Rietbergen, B.; Engh, C.A.; Amis, A.A. Correlation between pre-operative periprosthetic bone density and post-operative bone loss in THA can be explained by strain-adaptive remodelling. J. Biomech. 1999, 32, 695–703. [CrossRef] Sumner, D.R.; Galante, J.O. Determinants of stress shielding: Design versus materials versus interface. Clin. Orthop. Relat. Res. 1992, 274, 202–212. [CrossRef] Sumner, D.R.; Turner, T.M.; Igloria, R.; Urban, R.M.; Galante, J.O. Functional adaptation and ingrowth of bone vary as a function of hip implant stiffness. J. Biomech. 1998, 31, 909–917. [CrossRef] Turner, T.M.; Sumner, D.R.; Urban, R.M.; Igloria, R.; Galante, J.O. Maintenance of proximal cortical bone with use of a less stiff femoral component in hemiarthroplasty of the hip without cement. An investigation in a canine model at six months and two years. J. Bone Jt. Surg. Am. 1997, 79, 1381–1390. [CrossRef] Van Rietbergen, B.; Huiskes, R.; Weinans, H.; Sumner, D.R.; Turner, T.M.; Galante, J.O. The mechanism of bone remodeling and resorption around press-fitted THA stems. J. Biomech. 1993, 26, 369–382. [CrossRef] Wolff, J. The Law of Bone Remodelling; Springer: Berlin/Heidelberg, Germany, 1986. Joint-IMplant-Surgeons-Hip-Replacement-Imgae.png (215 × 360). Available online: http://www.jointimplantsurgeons.com/wp-content/uploads/2015/09/Joint-IMplant-SurgeonsHip-replacement-imgae.png (accessed on 5 May 2017). Nagels, J.; Stokdijk, M.; Rozing, P.M. Stress shielding and bone resorption in shoulder arthroplasty. J. Shoulder Elb. Surg. 2003, 12, 35–39. [CrossRef] [PubMed] Huiskes, R.; Weinans, H.; van Rietbergen, B. The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials. Clin. Orthop. Relat. Res. 1992, 274, 124–134. [CrossRef] Huiskes, R.; Ruimerman, R.; van Lenthe, G.H.; Janssen, J.D. Effects of mechanical forces on maintenance and adaptation of form in trabecular bone. Nature 2000, 405, 704–706. [CrossRef] [PubMed] Mullender, M.G.; Huiskes, R. Proposal for the regulatory mechanism of Wolff’s law. J. Orthop. Res. 1995, 13, 503–512. [CrossRef] [PubMed] Lanyon, L.E. Using functional loading to influence bone mass and architecture: Objectives, mechanisms, and relationship with estrogen of the mechanically adaptive process in bone. Bone 1996, 18, S37–S43. [CrossRef] Burger, E.H.; Klein-Nulend, J. Mechanotransduction in bone—Role of the lacuno-canalicular network. FASEB J. 1999, 13, S101–S112. [PubMed] Noble, B.S.; Stevens, H.; Loveridge, N.; Reeve, J. Identification of apoptotic changes in osteocytes in normal and pathological human bone. Bone 1997, 20, 273–282. [CrossRef] Staiger, M.P.; Pietak, A.M.; Huadmai, J.; Dias, G. Magnesium and its alloys as orthopedic biomaterials: A review. Biomaterials 2006, 27, 1728–1734. [CrossRef] [PubMed] DeGarmo, E.P.; Black, J.T.; Kohser, R.A. DeGarmo’s Materials and Processes in Manufacturing; John Wiley & Sons: Hoboken, NJ, USA, 2011; p. 1184. Gibson, L.J.; Ashby, M.F. Cellular Solids: Structure & Properties; Pergamon Press: Oxford, UK, 1988. Hänzi, A.C.; Sologubenko, A.S.; Uggowitzer, P.J. Design strategy for new biodegradable Mg–Y–Zn alloys for medical applications. Int. J. Mater. Res. 2009, 100, 1127–1136. [CrossRef] Vojtˇech, D.; Kubásek, J.; Šerák, J.; Novák, P. Mechanical and corrosion properties of newly developed biodegradable Zn-based alloys for bone fixation. Acta Biomater. 2011, 7, 3515–3522. [CrossRef] [PubMed] Kroeze, R.J.; Helder, M.N.; Govaert, L.E.; Smit, T.H. Biodegradable Polymers in Bone Tissue Engineering. Materials 2009, 2, 833–856. [CrossRef] Gunde, P.; Hänzi, A.C.; Sologubenko, A.S.; Uggowitzer, P.J. High-strength magnesium alloys for degradable implant applications. Mater. Sci. Eng. A 2011, 528, 1047–1054. [CrossRef] Granchi, D.; Ciapetti, G.; Stea, S.; Savarino, L.; Filippini, F.; Sudanese, A.; Zinghi, G.; Montanaro, L. Cytokine release in mononuclear cells of patients with Co}Cr hip prosthesis. Biomaterials 1999, 20, 1079–1086. [CrossRef]
Metals 2017, 7, 252
51.
52. 53.
54.
55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.
71.
72. 73. 74.
36 of 41
Niki, Y.; Matsumoto, H.; Suda, Y.; Otani, T.; Fujikawa, K.; Toyama, Y.; Hisamori, N.; Nozue, A. Metal ions induce bone-resorbing cytokine production through the redox pathway in synoviocytes and bone marrow macrophages. Biomaterials 2003, 24, 1447–1457. [CrossRef] Haynes, D.R.; Boyle, S.J.; Rogers, S.D.; Howie, D.W.; Vernon-Roberts, B. Variation in cytokines induced by particles from different prosthetic materials. Clin. Orthop. Relat. Res. 1998, 352, 223–230. [CrossRef] Wang, J.Y.; Wicklund, B.H.; Gustilo, R.B.; Tsukayarna, D.T. Titaniurn, chromium and cobalt ions modulate the release of bone-associated cytokines by human monocytes/macrophages in vitro. Biomoterials 1996, 17, 2233–2240. [CrossRef] Bi, Y.; van de Motter, R.R.; Ragab, A.A.; Goldberg, V.M.; Anderson, J.M.; Greenfield, E.M. Titanium particles stimulate bone resorption by inducing differentiation of murine osteoclasts. J. Bone Jt. Surg. Am. 2001, 83, 501–508. [CrossRef] Allen, M.J.; Myer, B.J.; Millett, P.J.; Rushton, N. The effects of particulate cobalt, chromium and cobalt-chromium alloy on human osteoblast-like cells in vitro. J. Bone Jt. Surg. Br. 1997, 79, 475–482. [CrossRef] Jacobs, J.J.; Gilbert, J.L.; Urban, R.M. Corrosion of metal orthopaedic implants. J. Bone Jt. Surg. Am. 1998, 80, 268–282. [CrossRef] Jacobs, J.J.; Hallab, N.J.; Skipor, A.K.; Urban, R.M. Metal degradation products: A cause for concern in metal-metal bearings? Clin. Orthop. Relat. Res. 2003, 417, 139–147. Beech, I.B.; Sunner, J.A.; Arciola, C.R.; Cristiani, P. Microbially-influenced corrosion: Damage to prostheses, delight for bacteria. Int. J. Artif. Organs 2006, 29, 443–452. [PubMed] Williams, D.F. On the nature of biomaterials. Biomaterials 2009, 30, 5897–5909. [CrossRef] [PubMed] Raman, R.K.S.; Jafari, S.; Harandi, S.E. Corrosion fatigue fracture of magnesium alloys in bioimplant applications: A review. Eng. Fract. Mech. 2015, 137, 97–108. [CrossRef] Sanderson, G.; Scully, J.C. Room Temperature Stress Corrosion Cracking of Titanium Alloys. Nature 1966, 211, 179. [CrossRef] Mori, K.; Takamura, A.; Shimose, T. Stress Corrosion Cracking off Ti and Zr in HCl-Methanol Solutions. Corrosion 1966, 22, 29–31. [CrossRef] Sutcliffe, J.M.; Fessler, R.R.; Boyd, W.K.; Parkins, R.N. Stress Corrosion Cracking of Carbon Steel in Carbonate Solutions. Corrosion 1972, 28, 313–320. [CrossRef] Mueller, W.-D.; Nascimento, M.L.; de Mele, M.F.L. Critical discussion of the results from different corrosion studies of Mg and Mg alloys for biomaterial applications. Acta Biomater. 2010, 6, 1749–1755. [CrossRef] [PubMed] Chen, K.; Dai, J.; Zhang, X. Improvement of corrosion resistance of magnesium alloys for biomedical applications. Corros. Rev. 2015, 33, 101–117. [CrossRef] Atrens, A.; Liu, M.; Abidin, N.I.Z. Corrosion mechanism applicable to biodegradable magnesium implants. Mater. Sci. Eng. B 2011, 176, 1609–1636. [CrossRef] Shaw, B.A. Corrosion Resistance of Magnesium Alloys. In ASM Handbook; ASM International Handbook Committee: Metals Park, OH, USA, 2003; Volume 13A. Witte, F.; Hort, N.; Vogt, C.; Cohen, S.; Kainer, K.U.; Willumeit, R.; Feyerabend, F. Degradable biomaterials based on magnesium corrosion. Curr. Opin. Solid State Mater. Sci. 2008, 12, 63–72. [CrossRef] Xu, L.; Yu, G.; Zhang, E.; Pan, F.; Yang, K. In vivo corrosion behavior of Mg-Mn-Zn alloy for bone implant application. J. Biomed. Mater. Res. Part A 2007, 83, 703–711. [CrossRef] [PubMed] Witte, F.; Kaese, V.; Haferkamp, H.; Switzer, E.; Meyer-Lindenberg, A.; Wirth, C.J.; Windhagen, H. In vivo corrosion of four magnesium alloys and the associated bone response. Biomaterials 2005, 26, 3557–3563. [CrossRef] [PubMed] Hänzi, A.C.; Gerber, I.; Schinhammer, M.; Löffler, J.F.; Uggowitzer, P.J. On the in vitro and in vivo degradation performance and biological response of new biodegradable Mg-Y-Zn alloys. Acta Biomater. 2010, 6, 1824–1833. [CrossRef] [PubMed] Ghali, E.; Dietzel, W.; Kainer, K.-U. General and Localized Corrosion of Magnesium Alloys: A Critical Review. J. Mater. Eng. Perform. 2004, 13, 7–23. [CrossRef] Kirkland, N.T.; Lespagnol, J.; Birbilis, N.; Staiger, M.P. A survey of bio-corrosion rates of magnesium alloys. Corros. Sci. 2010, 52, 287–291. [CrossRef] Manivasagam, G.; Suwas, S. Biodegradable Mg and Mg based alloys for biomedical implants. Mater. Sci. Technol. 2014, 30, 515–520. [CrossRef]
Metals 2017, 7, 252
75. 76. 77. 78. 79. 80. 81.
82. 83.
84.
85. 86. 87. 88. 89. 90. 91.
92. 93. 94. 95. 96.
97. 98. 99.
37 of 41
Winzer, N.; Atrens, A.; Song, G.; Ghali, E.; Dietzel, W.; Kainer, K.U.; Hort, N.; Blawert, C. A critical review of the Stress Corrosion Cracking (SCC) of magnesium alloys. Adv. Eng. Mater. 2005, 7, 659–693. [CrossRef] Song, G.; Atrens, A.; Dargusch, M. Influence of microstructure on the corrosion of diecast AZ91D. Corros. Sci. 1998, 41, 249–273. [CrossRef] Song, G.-L.; Atrens, A. Corrosion Mechanisms of Magnesium Alloys. Adv. Eng. Mater. 1999, 1, 11–33. [CrossRef] Jafari, S.; Harandi, S.E.; Raman, R.K.S. A review of stress-corrosion cracking and corrosion fatigue of magnesium alloys for biodegradable implant applications. JOM 2015, 67, 1143–1153. [CrossRef] Song, G. Control of biodegradation of biocompatable magnesium alloys. Corros. Sci. 2007, 49, 1696–1701. [CrossRef] Zberg, B.; Uggowitzer, P.J.; Löffler, J.F. MgZnCa glasses without clinically observable hydrogen evolution for biodegradable implants. Nat. Mater. 2009, 8, 887–891. [CrossRef] [PubMed] Matias, T.B.; Asato, G.H.; Ramasco, B.T.; Botta, W.J.; Kiminami, C.S.; Bolfarini, C. Processing and characterization of amorphous magnesium based alloy for application in biomedical implants. J. Mater. Res. Technol. 2014, 3, 203–209. [CrossRef] Kirkland, N.T. Magnesium biomaterials: Past, present and future. Corros. Eng. Sci. Technol. 2012, 47, 322–328. [CrossRef] Choudhary, L.; Raman, R.K.S. Acta Biomaterialia Magnesium alloys as body implants: Fracture mechanism under dynamic and static loadings in a physiological environment. Acta Biomater. 2012, 8, 916–923. [CrossRef] [PubMed] Hofstetter, J.; Martinelli, E.; Weinberg, A.M.; Becker, M.; Mingler, B.; Uggowitzer, P.J.; Löffler, J.F. Assessing the degradation performance of ultrahigh-purity magnesium in vitro and in vivo. Corros. Sci. 2015, 91, 29–36. [CrossRef] Persaud-Sharma, D.; McGoron, A. Biodegradable Magnesium Alloys: A Review of Material Development and Applications. J. Biomim. Biomater. Tissue Eng. 2012, 12, 25–39. [CrossRef] [PubMed] Jafari, S.; Raman, R.K.S.; Davies, C.H.J. Corrosion fatigue of a magnesium alloy in modified simulated body fluid. Eng. Fract. Mech. 2015, 137, 2–11. [CrossRef] Song, G.; Atrens, A. Understanding Magnesium Corrosion—A Framework for Improved Alloy Performance. Adv. Eng. Mater. 2003, 5, 837–858. [CrossRef] Potzies, C.; Kainer, K.U. Fatigue of magnesium alloys. Adv. Eng. Mater. 2004, 6, 281–289. [CrossRef] Zheng, Y.F.; Gu, X.N.; Witte, F. Biodegradable metals. Mater. Sci. Eng. R Rep. 2014, 77, 1–34. [CrossRef] Maragoni, L.; Carraro, P.A.; Peron, M.; Quaresimin, M. Fatigue behaviour of glass/epoxy laminates in the presence of voids. Int. J. Fatigue 2017, 95, 18–28. [CrossRef] Gu, X.N.; Zhou, W.R.; Zheng, Y.F.; Cheng, Y.; Wei, S.C.; Zhong, S.P.; Xi, T.F.; Chen, L.J. Corrosion fatigue behaviors of two biomedical Mg alloys—AZ91D and WE43—In simulated body fluid. Acta Biomater. 2010, 6, 4605–4613. [CrossRef] [PubMed] Logan, H.L. Mechanism of Stress-Corrosion Cracking in the AZ31B Magnesium Alloy. J. Res. Natl. Bur. Stand. (1934) 1958, 61, 503–508. [CrossRef] Logan, H.L. Film-Rupture Mechanism of Stress Corrosion. J. Res. Natl. Bur. Stand. (1934) 1952, 48, 99–105. [CrossRef] Ebtehaj, K.; Hardie, D.; Parkins, R.N. The influence of chloride-chromate solution composition on the stress corrosion cracking of a Mg-Al alloy. Corros. Sci. 1988, 28, 811–821. [CrossRef] Wearmouth, W.R.; Dean, G.P.; Parkins, R.N. Role of Stress in the Stress Corrosion Cracking of a Mg-Al Alloy. Corrosion 1973, 29, 251–260. [CrossRef] Pugh, E.N.; Green, J.A.S.; Slattery, P.W. On the propagation of stress-corrosion cracks in magnesium-aluminum alloy. In Proceedings of the 2nd International Conference on Fracture, London, UK, April 1969; Chapman & Hall Limited: London, UK, 1969. Pickering, H.W.; Swann, P.R. Electron Metallography of Chemical Attack Upon Some Alloys Susceptible to Stress Corrosion Cracking. Corrosion 1963, 19, 373t–389t. [CrossRef] Fairman, L.; Bray, H.J. Intergranular Stress-Corrosion Crack Propagation in Magnesium Aluminium Alloys. Br. Corros. J. 1971, 6, 170–174. [CrossRef] Song, G.; Atrens, A.; Wu, X.; Zhang, B. Corrosion behaviour of AZ21, AZ501 and AZ91 in sodium chloride. Corros. Sci. 1998, 40, 1769–1791. [CrossRef]
Metals 2017, 7, 252
38 of 41
100. Pardue, W.M.; Beck, F.H.; Fontana, M.G. Propagation of stress-corrosion cracking in a magnesium-base alloy as determined by several techniques. Trans. Am. Soc. Met. 1961, 54, 539–548. 101. Raman, R.K.S.; Harandi, S.E. Understanding corrosion-assisted cracking of magnesium alloys for bioimplant applications. In Magnesium Technology; Springer International Publishing: Cham, Switzerland, 2016; pp. 343–346. 102. Uematsu, Y.; Kakiuchi, T.; Nakajima, M.; Nakamura, Y.; Miyazaki, S.; Makino, H. Fatigue crack propagation of AZ61 magnesium alloy under controlled humidity and visualization of hydrogen diffusion along the crack wake. Int. J. Fatigue 2014, 59, 234–243. [CrossRef] 103. Chakrapani, D.G.; Pugh, E.N. The transgranular SCC of a Mg-Al alloy: Crystallographic, fractographic and acoustic-emission studies. Metall. Trans. A 1975, 6, 1155–1163. [CrossRef] 104. Lynch, S.P.; Trevena, P. Stress Corrosion Cracking and Liquid Metal Embrittlement in Pure Magnesium. Corrosion 1988, 44, 113–124. [CrossRef] 105. Bursle, A.; Pugh, E. On the Mechanism of transgranular stress-corrosion cracking. In Mechanisms of Environment Sensitive Cracking of Materials; Materials Society: London, UK, 1977. 106. Makar, G.L.; Kruger, J. Corrosion of magnesium. Int. Mater. Rev. 1993, 38, 138–153. [CrossRef] 107. Fairman, L.; West, J.M. Stress corrosion cracking of a magnesium aluminium alloy. Corros. Sci. 1965, 5, 711–716. [CrossRef] 108. Gross, B.; Mendelson, A. Plane elastostatic analysis of V-notched plates. Int. J. Fract. Mech. 1972, 8, 267–276. [CrossRef] 109. Taylor, D. Fatigue of bone and bones: An analysis based on stressed volume. J. Orthop. Res. 1998, 16, 163–169. [CrossRef] [PubMed] 110. Bhuiyan, M.S.; Mutoh, Y.; Murai, T.; Iwakami, S. Corrosion fatigue behavior of extruded magnesium alloy AZ61 under three different corrosive environments. Int. J. Fatigue 2008, 30, 1756–1765. [CrossRef] 111. Sivakumar, M.; Rajeswari, S. Investigation of failures in stainless steel orthopaedic implant devices: Pit-induced stress corrosion cracking. J. Mater. Sci. Lett. 1992, 11, 1039–1042. [CrossRef] 112. Antunes, R.A.; de Oliveira, M.C.L. Corrosion fatigue of biomedical metallic alloys: Mechanisms and mitigation. Acta Biomater. 2012, 8, 937–962. [CrossRef] [PubMed] 113. Zeng, R.C.; Han, E.H.; Ke, W. Effect of Temperature and Relative Humidity on Fatigue Crack Propagation Behavior of AZ61 Magnesium Alloy. Mater. Sci. Forum 2007, 546–549, 409–412. [CrossRef] 114. Shiplov, S.A. Mechanisms for corrosion fatigue crack propagation. Fatigue Fract. Eng. Mater. Struct. 2002, 25, 243–259. [CrossRef] 115. Vasudevan, A.K.; Sadananda, K. Classification of environmentally assisted fatigue crack growth behavior. Int. J. Fatigue 2009, 31, 1696–1708. [CrossRef] 116. Rozali, S.; Mutoh, Y.; Nagata, K. Effect of frequency on fatigue crack growth behavior of magnesium alloy AZ61 under immersed 3.5 mass% NaCl environment. Mater. Sci. Eng. A 2010, 528, 2509–2516. [CrossRef] 117. Chakrapani, D.G.; Pugh, E.N. Hydrogen Embrittlement in a Mg-Al Alloy. Metall. Trans. A 1976, 7, 173–178. [CrossRef] 118. Song, G.; Atrens, A.; Stjohn, D.; Nairn, J.; Li, Y. The electrochemical corrosion of pure magnesium in 1 N NaCl. Corros. Sci. 1997, 39, 855–875. [CrossRef] 119. Perrault, G.G. Potentiostatic study of the magnesium electrode in aqueous solution. J. Electroanal. Chem. Interfac. Electrochem. 1970, 27, 47–58. [CrossRef] 120. Murray, R.W.; Hillis, J.E. Magnesium Finishing: Chemical Treatment and Coating Practices. SAE Technical Paper 900791, 1990. 121. Hillis, J.E. The Effects of Heavy Metal Contamination on Magnesium Corrosion Performance. SAE Technical Paper 830523, 1983. 122. Ahamed, M. Toxic response of nickel nanoparticles in human lung epithelial A549 cells. Toxicol. In Vitro 2011, 25, 930–936. [CrossRef] [PubMed] 123. Milavec-Pureti´c, V.; Orli´c, D.; Marusi´c, A. Sensitivity to metals in 40 patients with failed hip endoprosthesis. Arch. Orthop. Trauma Surg. 1998, 117, 383–386. [CrossRef] [PubMed] 124. Hillis, J.; Murray, R. Finishing alternatives for high purity magnesium alloys. In Proceedings of the SDCE 14th International Die Casting Congress and Exposition, Toronto, ON, Canada, 11–14 May 1987. 125. Pelensky, M.A.; Gallaccio, A. Stress corrosion of magnesium alloys—Environmental factors. In Stress Corrosion Testing, ASTM STP425; American Society for Testing and Materials: West Conshohocken, PA, USA, 1967.
Metals 2017, 7, 252
39 of 41
126. Shi, Z.; Song, G.; Atrens, A. Corrosion resistance of anodised single-phase Mg alloys. Surf. Coat. Technol. 2006, 201, 492–503. [CrossRef] 127. Hofstetter, J.; Becker, M.; Martinelli, E.; Weinberg, A.M.; Mingler, B.; Kilian, H.; Pogatscher, S.; Uggowitzer, P.J.; Löffler, J.F. High-strength low-alloy (HSLA) Mg-Zn-Ca alloys with excellent biodegradation performance. JOM 2014, 66, 566–572. [CrossRef] 128. Li, L.; Gao, J.; Wang, Y. Evaluation of cyto-toxicity and corrosion behavior of alkali-heat-treated magnesium in simulated body fluid. Surf. Coat. Technol. 2004, 185, 92–98. [CrossRef] 129. Hanawalt, J.D.; Nelson, C.E.; Peloubet, J.A. Corrosion studies of magnesium and its alloys. Trans AIME 1942, 47, 273–299. 130. Reichek, K.N.; Clark, K.J.; Hillis, J.E. Controlling the Salt Water Corrosion Performance of Magnesium AZ91 Alloy. SAE Technical Paper 850417, 1985. 131. Lunder, O.; Aune, T.K.; Nisancioglu, K. Effect of Mn Additions on the Corrosion Behavior of Mould-Cast Magnesium ASTM AZ91. Corrosion 1987, 43, 291–295. [CrossRef] 132. Mercer, W.E.; Hillis, J.E. The Critical Contaminant Limits and Salt Water Corrosion Performance of Magnesium AE42 Alloy. SAE Technical Paper 920073, 1992. 133. Avedesian, M.M.; Baker, H. Magnesium and Magnesium Alloys; ASM International Handbook Committee: Novelty, OH, USA, 1999. 134. Kainer, K.U.; Srinivasan, P.B.; Blawert, C.; Dietzel, W. Corrosion of magnesium and its alloys. Shreir’s Corros. 2010, 3, 2011–2041. 135. Mezbahul-Islam, M.; Mostafa, A.O.; Medraj, M. Essential Magnesium Alloys Binary Phase Diagrams and Their Thermochemical Data. J. Mater. 2014, 2014, 704283. [CrossRef] 136. Gusieva, K.; Davies, C.H.J.; Scully, J.R.; Birbilis, N. Corrosion of magnesium alloys: The role of alloying. Int. Mater. Rev. 2015, 60, 169–194. [CrossRef] 137. Hermann, F.; Sommer, F.; Jones, H.; Edyvean, R.G.J. Corrosion inhibition in magnesium-aluminium-based alloys induced by rapid solidification processing. J. Mater. Sci. 1989, 24, 2369–2379. [CrossRef] 138. Lunder, O.; Lein, J.E.; Aune, T.K.; Nisancioglu, K. The Role of Mg 17 Al 12 Phase in the Corrosion of Mg Alloy AZ91. Corrosion 1989, 45, 741–748. [CrossRef] 139. Jeong, Y.S.; Kim, W.J. Enhancement of mechanical properties and corrosion resistance of Mg–Ca alloys through microstructural refinement by indirect extrusion. Corros. Sci. 2014, 82, 392–403. [CrossRef] 140. Domingo, J.L. Reproductive and developmental toxicity of aluminum: A review. Neurotoxicol. Teratol. 1995, 17, 515–521. [CrossRef] 141. Venugopal, B.; Luckey, T.D. Metal Toxicity in Mammals. Volume 2. Chemical Toxicity of Metals and Metalloids; Plenum Press: Berlin, Germany, 1978. 142. Flaten, T.P. Aluminium as a risk factor in Alzheimer’s disease, with emphasis on drinking water. Brain Res. Bull. 2001, 55, 187–196. [CrossRef] 143. El-Rahman, S.S.A. Neuropathology of aluminum toxicity in rats (glutamate and GABA impairment). Pharmacol. Res. 2003, 47, 189–194. [CrossRef] 144. Robinson, H.A.; George, P.F. Effect of Alloying and Impurity Elements In Magnesium Alloy Cast Anodes. Corrosion 1954, 10, 182–188. [CrossRef] 145. Culotta, V.C.; Yang, M.; Hall, M.D. Manganese Transport and Trafficking: Lessons Learned from Saccharomyces cerevisiae. Eukaryot. Cell 2005, 4, 1159–1165. [CrossRef] [PubMed] 146. Brandt, E.G.; Hellgren, M.; Brinck, T.; Bergman, T.; Edholm, O. Molecular dynamics study of zinc binding to cysteines in a peptide mimic of the alcohol dehydrogenase structural zinc site. Phys. Chem. Chem. Phys. 2009, 11, 975–983. [CrossRef] [PubMed] 147. Prasad, A.S. Zinc in human health: Effect of zinc on immune cells. Mol. Med. 2008, 14, 353–357. [CrossRef] [PubMed] 148. Bothwell, D.N.; Mair, E.A.; Cable, B.B. Chronic ingestion of a zinc-based penny. Pediatrics 2003, 111, 689–691. [CrossRef] [PubMed] 149. Zhang, S.; Zhang, X.; Zhao, C.; Li, J.; Song, Y.; Xie, C.; Tao, H.; Zhang, Y.; He, Y.; Jiang, Y.; et al. Research on an Mg–Zn alloy as a degradable biomaterial. Acta Biomater. 2010, 6, 626–640. [CrossRef] [PubMed] 150. Li, Y.; Wen, C.; Mushahary, D.; Sravanthi, R.; Harishankar, N.; Pande, G.; Hodgson, P. Mg–Zr–Sr alloys as biodegradable implant materials. Acta Biomater. 2012, 8, 3177–3188. [CrossRef] [PubMed]
Metals 2017, 7, 252
40 of 41
151. Song, G.; StJohn, D. The effect of zirconium grain refinement on the corrosion behaviour of magnesium-rare earth alloy MEZ. J. Light Met. 2002, 2, 1–16. [CrossRef] 152. Jiao, W.; Li, H.F.; Zhao, K.; Bai, H.Y.; Whang, Y.B.; Zheng, Y.F.; Wang, W.H. Development of CaZn based glassy alloys as potential biodegradable bone graft substitute. J. Non-Cryst. Solids 2011, 357, 3830–3840. [CrossRef] 153. Rad, H.R.B.; Idris, M.H.; Kadir, M.R.A.; Farahany, S. Microstructure analysis and corrosion behavior of biodegradable Mg–Ca implant alloys. Mater. Des. 2012, 33, 88–97. [CrossRef] 154. Zhang, E.; Yang, L. Microstructure, mechanical properties and bio-corrosion properties of Mg–Zn–Mn–Ca alloy for biomedical application. Mater. Sci. Eng. A 2008, 497, 111–118. [CrossRef] 155. Kirkland, N.T.; Birbilis, N.; Walker, J.; Woodfield, T.; Dias, G.J.; Staiger, M.P. In-vitro dissolution of magnesium-calcium binary alloys: Clarifying the unique role of calcium additions in bioresorbable magnesium implant alloys. J. Biomed. Mater. Res. Part B Appl. Biomater. 2010, 95, 91–100. [CrossRef] [PubMed] 156. Bornapour, M.; Celikin, M.; Cerruti, M.; Pekguleryuz, M. Magnesium implant alloy with low levels of strontium and calcium: The third element effect and phase selection improve bio-corrosion resistance and mechanical performance. Mater. Sci. Eng. C 2014, 35, 267–282. [CrossRef] [PubMed] 157. Lazar, L.L.; Rokhlin, L. Magnesium Alloys Containing Rare Earth Metals: Structure and Properties; Taylor & Francis: Abingdon, UK, 2003. 158. Kannan, M.B.; Dietzel, W.; Blawert, C.; Atrens, A.; Lyon, P. Stress corrosion cracking of rare-earth containing magnesium alloys ZE41, QE22 and Elektron 21 (EV31A) compared with AZ80. Mater. Sci. Eng. A 2008, 480, 529–539. [CrossRef] 159. Zhang, T.; Meng, G.; Shao, Y.; Cui, Z.; Wang, F. Corrosion of hot extrusion AZ91 magnesium alloy. Part II: Effect of rare earth element neodymium (Nd) on the corrosion behavior of extruded alloy. Corros. Sci. 2011, 53, 2934–2942. [CrossRef] 160. Li, W.; Cao, F.; Zhong, L.; Zheng, L.; Jia, B.; Zhang, Z.; Zhang, J. Influence of rare earth element Ce and La addition on corrosion behavior of AZ91 magnesium alloy. Mater. Corros. 2009, 60, 795–803. [CrossRef] 161. Zhou, X.; Huang, Y.; Wei, Z.; Chen, Q.; Gan, F. Improvement of corrosion resistance of AZ91D magnesium alloy by holmium addition. Corros. Sci. 2006, 48, 4223–4233. [CrossRef] 162. Yao, S.J.; Yi, D.Q.; Yang, S.; Cang, X.H.; Li, W.X. Effect of Sc on Microstructures and Corrosion Properties of AZ91. Mater. Sci. Forum 2007, 546–549, 139–142. [CrossRef] 163. Santamaria, M.; di Quarto, F.; Zanna, S.; Marcus, P. Initial surface film on magnesium metal: A characterization by X-ray photoelectron spectroscopy (XPS) and photocurrent spectroscopy (PCS). Electrochim. Acta 2007, 53, 1314–1324. [CrossRef] 164. Hänzi, A.C.; Gunde, P.; Schinhammer, M.; Uggowitzer, P.J. On the biodegradation performance of an Mg–Y–RE alloy with various surface conditions in simulated body fluid. Acta Biomater. 2009, 5, 162–171. [CrossRef] [PubMed] 165. Best, S.M.; Porter, A.E.; Thian, E.S.; Huang, J. Bioceramics: Past, present and for the future. J. Eur. Ceram. Soc. 2008, 28, 1319–1327. [CrossRef] 166. Hornberger, H.; Virtanen, S.; Boccaccini, A.R. Biomedical coatings on magnesium alloys—A review. Acta Biomater. 2012, 8, 2442–2455. [CrossRef] [PubMed] 167. Song, Y.; Zhang, S.; Li, J.; Zhao, C.; Zhang, X. Electrodeposition of Ca–P coatings on biodegradable Mg alloy: In vitro biomineralization behavior. Acta Biomater. 2010, 6, 1736–1742. [CrossRef] [PubMed] 168. Wang, H.X.; Guan, S.K.; Wang, X.; Ren, C.X.; Wang, L.G. In vitro degradation and mechanical integrity of Mg–Zn–Ca alloy coated with Ca-deficient hydroxyapatite by the pulse electrodeposition process. Acta Biomater. 2010, 6, 1743–1748. [CrossRef] [PubMed] 169. Zhu, B.; Xu, Y.; Sun, J.; Yang, L.; Guo, C.; Liang, J.; Cao, B. Preparation and Characterization of Aminated Hydroxyethyl Cellulose-Induced Biomimetic Hydroxyapatite Coatings on the AZ31 Magnesium Alloy. Metals 2017, 7, 214. [CrossRef] 170. Yang, J.; Cui, F.Z.; Lee, I.; Zhang, Y.; Yin, Q.S.; Xia, H.; Yang, S.X. In vivo biocompatibility and degradation behavior of Mg alloy coated by calcium phosphate in a rabbit model. J. Biomater. Appl. 2012, 27, 153–164. [CrossRef] [PubMed] 171. Chiu, K.Y.; Wong, M.H.; Cheng, F.T.; Man, H.C. Characterization and corrosion studies of fluoride conversion coating on degradable Mg implants. Surf. Coat. Technol. 2007, 202, 590–598. [CrossRef] 172. Witte, F.; Fischer, J.; Nellesen, J.; Vogt, C.; Vogt, J.; Donath, T.; Beckmann, F. In vivo corrosion and corrosion protection of magnesium alloy LAE442. Acta Biomater. 2010, 6, 1792–1799. [CrossRef] [PubMed]
Metals 2017, 7, 252
41 of 41
173. Li, J.N.; Cao, P.; Zhang, X.N.; Zhang, S.X.; He, Y.H. In vitro degradation and cell attachment of a PLGA coated biodegradable Mg–6Zn based alloy. J. Mater. Sci. 2010, 45, 6038–6045. [CrossRef] 174. Albayrak, O.; El-Atwani, O.; Altintas, S. Hydroxyapatite coating on titanium substrate by electrophoretic deposition method: Effects of titanium dioxide inner layer on adhesion strength and hydroxyapatite decomposition. Surf. Coat. Technol. 2008, 202, 2482–2487. [CrossRef] 175. Glocker, D.A.; Ranade, S.V. Medical Coatings and Deposition Technologies; John Wiley & Sons: Hoboken, NJ, USA, 2016. © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
