Development of a new technique for manufacturing biodegradable magnesium stents

Universidade do Minho Escola de Engenharia Vítor Emanuel Rebelo Lopes Development of a New Technique for Manufacturing Biodegradable Magnesium Stents janeiro de 2022 UMinho | 2022 Vítor Emanuel Rebelo Lopes Development of a New Technique for Manufacturing Biodegradable Magnesium Stents Vítor Emanuel Rebelo Lopes Development of a New Technique for Manufacturing Biodegradable Magnesium Stents Tese de Doutoramento Programa Doutoral em Engenharia Mecânica Trabalho realizado sob a orientação de Professor Doutor Hélder Jesus Fernandes Puga Professor Doutor José Carlos Fernandes Teixeira Janeiro de 2022 ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial CC BY-NC https://creativecommons.org/licenses/by-nc/4.0/ iii ACKNOWLEDGMENTS The support and assistance I received during this work were crucial for the accomplishment of this objective. Over the past few years, with my commitment and resilience, I have been able to learn so much and grow up at a professional and personal level. However, this was only possible with the accompaniment and support of different individuals and institutions. To all of them, I would like to express my deepest gratitude. My supervisor, Professor Hélder Puga, I thank him for believing in my potential and for making this opportunity possible. As a result of his work capacity and nonconformist character, I was always motivated to improve. He was always present and available, which was crucial for the success of this project. To my co-supervisor, Professor José Carlos Teixeira, I am thankful for his wise and friendly advice. His knowledge, care, and friendship contributed greatly to the success of the project. I would like to thank Fundação para a Ciência e a Tecnologia (FCT) for funding this work through the Ph.D. grant SFRH/BD/129223/2017, in the context of the Mechanical Engineering Doctoral Program at University of Minho. I am very grateful to the Department of Mechanical Engineering of the University of Minho for all institutional support and for providing me with the opportunity to serve as an assistant lecturer. A special thanks to all departmental Professors and technical workers for their unceasing support and constant welfare. I would like to extend my sincere thanks to Professor Joaquim Barbosa for his wise pieces of advice, friendship and for supporting me along this academic journey. To all my academic colleagues and friends over the past years, I really appreciated their unique friendship at every moment. To all my laboratory colleagues, I want to express my gratitude for their support and pleasant working hours. Special thanks to my friends Flávia and Filipe for being my true partners on this journey. It was an honor to share these moments with them. Finally, I owe and dedicate this work to my parents, my brothers, my nephews, and my girlfriend. They are the most important people in my life. Agradeço a Deus por tudo! iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. v RESUMO Desenvolvimento de uma nova tecnologia para a fabricação de stents biodegradáveis de magnésio Nos últimos anos, o magnésio e as suas ligas têm sido o principal foco de vários estudos, tanto a nível académico como industrial. As propriedades mecânicas deste material, devido à sua natureza metálica e inerente baixa densidade, proporcionam altos valores de força específica e módulo específico. Contudo, as ligas de magnésio são de grande dificuldade de processamento quando fundidas. Isto deve-se à autoignição de alguns dos seus elementos constituintes abaixo da temperatura de liquidus e à sua elevada reatividade com os elementos circundantes acima desta temperatura. Todos estes fatores tornam a fabricação de estruturas em magnésio de paredes finas um processo extremamente complexo de difícil controlo, sendo normalmente evitado na indústria. Relativamente à fundição de ligas de magnésio, algumas técnicas têm sido desenvolvidas para superar essas dificuldades, havendo, principalmente, um esforço considerável na mitigação das reações existentes. As abordagens estudadas e propostas incluem o uso de atmosferas protetivas, a aplicação de carapaças e revestimentos refratários, bem como a melhoria dos atuais processos de fabrico a que estas ligas são sujeitas. Embora um progresso significativo tenha sido feito para garantir vazamentos bem-sucedidos, a produção de componentes com estruturas de paredes finas continua a ser considerado um processo pouco desenvolvido e confiável. O presente estudo aborda o desenvolvimento de uma nova metodologia de fundição por modelo perdido de paredes finas da liga de magnésio AZ91D-1 wt. % CaO em moldação de gesso, especificamente otimizada para a fabricação de stents . Métodos como manufatura aditiva, revestimento à base de Ítria e fundição com assistência de vácuo são investigados e aplicados para melhorar o processo. Para isso, diferentes estudos experimentais são realizados para estudo e compreensão dos fenómenos que podem ocorrer em cada etapa do processo. A interface moldemetal é devidamente caracterizada e um estudo de fluidez é realizado para perceção da influência de cada variável termodinâmica no comprimento total de enchimento das cavidades. A obtenção de geometrias de stents vazados em liga de magnésio com 0,4 mm e 0,8 mm de espessura de parede demonstra o bom desempenho das técnicas propostas na mitigação das reações molde-metal, permitindo confirmar a eficácia da metodologia desenvolvida. Palavras-chave: Fundição por modelo perdido; Ligas de magnésio; Stents; Manufatura aditiva; Reações molde-metal; Revestimento à base de Ítria; Vácuo. vi ABSTRACT Development of a new technique for manufacturing biodegradable magnesium stents In the last years, magnesium and its alloys have been the focus of many studies in both academic and industrial fields. The mechanical properties of this material, due to its metallic nature and inherent low density, yield high values of specific strength and modulus. However, Mg alloys are known for their relatively difficult processing in casting. This is due to its auto-ignition prior to liquidus temperature in some of its constituent elements and their high reactivity above this temperature. All these factors imply that manufacturing of thin-walled magnesium structures is an extremely complex process and hard to control, which is generally avoided in industry. Regarding the casting of magnesium alloys, some techniques have been developed to surpass this inherent difficulty. Hence, there has been essentially a considerable effort in the reactions mitigation. The studied and proposed approaches include the use of protective atmospheres, the application of refractory shells and coatings, as well as the improvement of current manufacturing processes that these alloys are subjected to. Despite of the significant progress to guarantee successful castings, the production of thin-walled magnesium structures is still considered an unreliable and not well-established process. The present study addresses a thin-walled investment casting methodology of AZ91D-1 wt.% CaO magnesium alloy in plaster molding, specially optimized for stent fabrication. Additive manufacturing, Yttria-based coating, and vacuum-assisted casting are investigated and applied to optimize the process. To do this, different experimental studies are conducted in order to better understand all of the phenomena that may occur during each stage of the process. In this work, the mold-metal interface is extensively explored, and the effect of each thermodynamic variable on the cavity filling length is evaluated on thin-walled samples. Magnesium stents with wall thicknesses of 0.4 mm and 0.8 mm have been obtained using the developed techniques for mitigating mold-metal reactions, demonstrating the effectiveness of the proposed methodology. Keywords: Investment casting; Magnesium alloys; Stents; Additive manufacturing; Mold-metal reactions; Yttria-based coating; Vacuum. vii TABLE OF CONTENTS ACKNOWLEDGMENTS ................................................................................................... III RESUMO ..................................................................................................................... V ABSTRACT ................................................................................................................. VI TABLE OF CONTENTS ................................................................................................... VII LIST OF FIGURES ......................................................................................................... IX LIST OF TABLES ......................................................................................................... XIV LIST OF ABBREVIATIONS ............................................................................................... XVI LIST OF SYMBOLS ..................................................................................................... XVIII CHAPTER 1. INTRODUCTION ........................................................................................ 1 1.1. Overview and Motivation ............................................................................................ 2 1.2. Statement of the Problem .......................................................................................... 7 1.3. Research Objectives (RO) ........................................................................................... 8 1.4. Thesis Structure ......................................................................................................... 9 1.5. Overall Contributions of the Work ............................................................................. 10 CHAPTER 2. BACKGROUND LITERATURE ....................................................................... 12 2.1. Processing of Magnesium Alloys............................................................................... 13 2.2. Magnesium Investment Casting in Ceramic Molding ................................................. 19 2.3. Thin-walled Magnesium Structures ........................................................................... 22 2.4. Stent Manufacturing ................................................................................................. 24 2.5. Summary and Conclusions ...................................................................................... 36 xiv LIST OF TABLES Table 1.1: Main properties of pure magnesium [29]. ................................................................. 5 Table 1.2: Relative power and comparative machinability (According to Luxfer MEL Technologies*). .......................................................................................................... 6 Table 2.1: Advantages and drawbacks of different processing methods for obtaining biomedical devices. ................................................................................................................... 37 Table 3.1: Comparison between FDM and SLA characteristics. ................................................ 43 Table 3.2: Coating application characteristics. ......................................................................... 46 Table 3.3: Acquisition data for wettability analysis (one measurement reading). ....................... 49 Table 3.4: Chemical composition of AZ91D-1 wt.% CaO magnesium alloy (mass%)................... 55 Table 3.5: Key parameters adopted during processing. ........................................................... 55 Table 3.6: Possible mold-metal reactions on Mg-SiO2 interface [48]. .......................................... 58 Table 4.1: Technical data of the experimental setup (according to the manual). ....................... 69 Table 4.2: Optimal safe casting parameters (experimentally optimized). ................................... 70 Table 4.3: The main variables and values used in evaluating fluidity. ....................................... 75 Table 4.4: Calculated values from the experimental fluidity study. ............................................ 78 Table 4.5: Selected factors and their levels. ............................................................................ 86 Table 4.6: Results of Analysis of Variance. ............................................................................... 88 Table 5.1: Filling design optimization using condition YV and 720 ºC pouring temperature. ...... 96 Table 5.2: Optimized painting characteristics. ....................................................................... 100 Table 5.3: EDS analysis of the points indicated in Figure 5.14. .............................................. 107 Table 5.4: EDS analysis of the points indicated in Figure 5.15. .............................................. 108 xv Table 5.5: EDS analysis of the points indicated in Figure 5.23. .............................................. 116 Table 5.6: X-ray µCT scan parameters in Figure 5.25. .......................................................... 118 xvi LIST OF ABBREVIATIONS 2D Two-dimensional 3D Three-dimensional ABE Accumulative Back Extrusion ABS Acrylonitrile Butadiene Styrene ANOVA Analysis of Variance CAD Computer-aided Design CEC Cyclic Extrusion and Compression CEE Cyclic Expansion Extrusion CNC Computerized Numerical Control DoE Design of Experiments DoF Degrees of Freedom DTG Derivative Thermo Gravimetry ECAP Equal Channel Angular Pressing ECO-Mg Environment-Conscious Magnesium EDS Energy-dispersive X-ray Spectroscopy EX Direct Extrusion FDM Fused Deposition Modelling FESEM Field Emission Scanning Electron Microscopy HPDC High Pressure Die Casting HPT High-Pressure Torsion MAP Magnetic Abrasive Polishing xvii Micro-CT (μ-CT) Micro-computed Tomography MIM Metal Injection Molding MSA Mobile Surface Analyzer MTE Microtube Extrusion OM Optical Microscopy OWRK Owens-Wendt-Rabel-Kaelble PLA Polylactic Acid PM Powder Metallurgy PVA Polyvinyl Alcohol RE Rare Earth SEM Scanning Electron Microscopy SLA Stereolithography SLM Selective Laser Melting SMA’s Shape Memory Alloys SPD Severe Plastic Deformation SS Sequential Sum of Squares SS’ Pure Sum of Squares STL Standard Triangle Language TCAP Tubular Channel Angular Pressing TGA Thermal Gravimetric Analysis UST Ultrasound Technique UV Ultraviolet WHO World Health Organization XRD X-ray Diffraction xviii LIST OF SYMBOLS Symbol Unit (SI) Description l m Length of the squared cross-section (spiral) L m Filling length (spiral) and total length (stent) Ø m Diameter p m Spiral pitch R (θ) m Radius of the spiral in position θ R 0 m Initial radius of the spiral center T K Temperature t m Stent wall thickness Δ G o f m Gibbs free energy of formation θ rad Angular parameter along the spiral x, y, z X, Y, Z - Cartesian coordinates 1 There is a scientific publication of 1997 entitled "The origin of the word stent". Although it remains uncertain, it seems likely that the word originated with the dentist Charles T. Stent (1807-1885) [1]. Chapter 1. INTRODUCTION As engineering applies to various knowledge fields, society have become accustomed to significant innovations, constructions, and improvements. Engineering is synonymous with development and will continue to be. It continues to be critically important in every knowledge area and major market sector to optimize processes and products in order to minimize resources employed, namely time, costs, and manpower. Considering its scientific and innovative focus, the main objective of this study is to develop an innovative methodology for casting Mg (magnesium) alloys in plaster molding, as well as its optimization for thin-walled magnesium alloy parts, especially stents. Magnesium Alloy (material), Investment Casting (process), and Stent device (application) are the three major pillars of this research, which is of great interest to both the scientific community and the industry, thereby presenting an added value to the advancement of this theme in the field of engineering. 2 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS 1.1. Overview and Motivation Ischemic heart disease currently ranks as the world’s biggest killer, as can be observed in Figure 1.1. According to recent data collected from the WHO (World Health Organization) and displayed on their website, in the last 21 years, this disease has been the leading cause of death increase, rising by more than 2 million to 8.9 million deaths in 2019, which represents 16 % of all global deaths [2]. The issue is a grave and real one, which has intensified over the years and has not undergone significant evolution regarding the search for new solutions or the enhancement of existing ones. It is necessary to act predictively and not only in a corrective and/or preventative manner. Figure 1.1: Top 5 causes of death [2]. Atherosclerosis is the most common cause of coronary artery disease. The thickness increase and the elasticity loss of the arterial walls can lead to artery blockages, limiting the normal blood flow. A complete blockage of arteries can lead to serious health problems, including angina, myocardial infarction, arrhythmias, and even death caused by heart failure. According to the severity of the injury and the clinical history of the patient [3], percutaneous coronary interventions (also called coronary angioplasty) with stent deployment have constituted one of the main methods to combat this problem and the most appropriate to apply 01000 2000 3000 4000 5000 6000 7000 8000 9000 10000 Ischaemic heart disease Stroke Chronic obstructive pulmonary disease Lower respiratory infections Neonatal conditions NUMBER OF DEATHS (IN MILLIONS) 2019 2000 CHAPTER 1. INTRODUCTION 3 over time [4, 5]. A stent is a metallic expandable knitted medical device introduced into veins and arteries. Through its radial expansion, it unclogs such vessels to increase blood flow to the organs and prevent complications generated therein [6, 7]. In-stent angioplasty, a medical device is mounted in a flask with the orientation of a catheter system. As shown in Figure 1.2, the balloonstent system is introduced as a compressed form into the clogged region, after which it is inflated to compress the plaque against the vessel walls. The balloon is removed after angioplasty, and the stent remains in the artery to support maintaining the vessel walls open, ensuring correct blood circulation. Therefore, the application of the stent has sparked interest by the scientific community in this thematic, and there have been several studies conducted on the design and manufacture of this medical device. Figure 1.2: Stent delivery system (Abbott Laboratories). Nowadays, most metal stents are manufactured in stainless steel and alloys based on nickel and titanium, cobalt and chromium, platinum, and others [4, 8-12]. As a result, these materials provide mechanical properties that enable the stents to perform a structural role within vessel walls without compromising the blood flow [13]. Regarding the manufacturing process, and according to studies reported by Demir and Previtali [10], Hermawan et al. [14], and Moravej and Mantovani [15], laser cutting is the usual method used for fabricating metallic stents, even though some subsequent finishing operations (deburring, chemical finishing, electropolishing, electroplating, and others) are required to improve, essentially, the superficial quality of the stents [10, 16]. Although actual stent fabrication requires sophisticated manufacturing processes and materials offering better properties, the need for metal stents to remain permanently in the same region where they are placed remains a barrier to biocompatibility between stents and the human 4 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS organism, compromising the implant success [17, 18]. A stent’s long-term presence inside the artery can lead to complications, such as restenosis [4, 19], late thrombosis [20, 21], and other complications, for instance, the internal artery lining in response to arterial wall injuries and inflammatory reactions, which raises questions about the evolution of treatments for heart disease [22]. In some cases, a second surgery is required to remove the stent or insert an alternative, resulting in serious health problems, even death. Additionally, the manufacturing process can also be cause-related to some health complications in stent’ applications. The laser cutting process, for example, can induce thermal problems that cause insufficient surface quality to the manufactured stents [6]. Even more, if the stent does not have good visibility under fluoroscopy, the actual stent placement procedure can also pose problems associated with stent placement and interaction [23, 24]. Therefore, addressing stent models improvement presents a clear opportunity for progress in addressing the growth and benefits of engineering advances [14]. Stents made of biodegradable materials, such as magnesium, and through methods that can overcome some of the current manufacturing limitations, can be considered a field that presents an enormous potential for success in combating the deadliest diseases in today’s world [25-27]. On the other hand, the limited information about processing techniques applied to magnesium applications has been an important catalyst for the scientific community to develop this theme, which has sparked an increasing interest in lightweight materials. Magnesium has indeed been the subject of increasing interest both at the academic and industrial levels. This can be justified by the growing knowledge of applying its excellent properties (Table 1.1) and by the compelling need for continuous improvement of engineering systems. In addition to being one of the most abundant elements on the planet, which avoids problems in its exploitation, magnesium has a density value of 1.7 g/cm3, which is 33 % lighter than Al (aluminum) and 75 % lighter than steel, presenting as well good mechanical performance [28]. CHAPTER 1. INTRODUCTION 5 Table 1.1: Main properties of pure magnesium [29]. Property Unit Value Atomic mass 24.31 Atomic number 12 Boiling point ºC 1107 Crystalline structure A3 Density at 20 ºC g/ cm3 1.74 Elastic modulus GPa 40 Elongation % 2-10* Hardness HBW 45-47* Thermal conductivity W/ (m∙K) 418 Melting point ºC 650 Resistivity nΩ∙m 44.5 Specific heat capacity kJ/ (kg∙K) 1.025 Tensile strength MPa 180-220* Thermal expansion coefficient 10-6 1/ K 25.2 Yield point, 0.2 MPa 115-140* *rolling Magnesium has gained its largest application area by creating alloys in combination with other elements. Currently, magnesium alloys have ceased to be future materials and have become present in some specific areas [30]. This fact was recently corroborated by Xu et al [31], which reported that the ongoing interest in studying and exploring magnesium alloys resulted in a growth rate of 491 % of publications in the 21st century. In addition to being one of the lightest metals, these alloys have an excellent strength-to-weight ratio, which helps justify the high investments carried out by the scientific community and the industrial sector. Moreover, the demand for these kinds of alloys is driven by their distinct and attractive engineering properties. As a result of their low density, good castability, and high damping capacity, there has been a continuous demand for their use in a wider range of applications, such as aerospace industry [32], as well as for automotive applications [33, 34]. Thus, magnesium and its alloys (Mg alloys) have tremendous potential to replace materials such as steel, aluminum, and plastic-based materials, increasing the research interest to improve further their properties [35, 36]. Science has continued to investigate emerging technologies for magnesium alloys and increase their use for a significant evolution in these industries [37, 38]. According to Lara-Rodriguez et al. [39] and Kucharczyk et al. [40], the functional properties of Mg foams, such as sound and energy absorption and excellent vibration 12 According to the authoritative scientific literature database Web of Science Core Collection, the number of publications on magnesium alloys has increased from 1344 in 2008 to 3034 in 2020. Chapter 2. BACKGROUND LITERATURE With the coverage of the literature review on biodegradable medical devices, one may conclude that, although the mechanical and medical properties of the material applied are important, finding a focus should be placed on finding the best fabrication process to manufacture those devices. The processing methods greatly impact the final device’s characteristics, especially in stent manufacturing [51]. Additionally, when developing metallic biodegradable applications, the fabrication process determines whether a piece of starting stock can be converted into a functional medical device in a cost-effective manner, which is an exceptionally critical factor [11]. In the case of stent manufacturing, different fabrication processes may be used to improve the mechanical and structural properties and control the degradation rate of the devices, which becomes a disadvantage when using biodegradable materials. Therefore, several techniques for strengthening the properties of magnesium alloys have been studied, and different processes for manufacturing them have been developed according to their end use. In this work, the stent processing is directly correlated to the material, i.e., magnesium alloys. Thus, the topic of this chapter relates to the materials and manufacturing processes used in magnesium processing, and how they relate to the use of stents. Various approaches to stent manufacturing are analyzed, and a new promising method as a potential solution is described. CHAPTER 2. BACKGROUND LITERATURE 13 2.1. Processing of Magnesium Alloys Magnesium alloys have a large margin of progress in obtaining better and higher engineering properties, increasing their interest in constantly growing sectors. Emerging studies have highlighted some methodologies that enhance the mechanical properties of magnesium alloys, focusing on alloying, microstructure control, and others. The addition of elements is one of the most commonly used methods to achieve higher properties, as referred to by Wang et al. [52] and Razzaghi et al. [53], following the studies of Xue et al. [54] and Wang et al. [55]. Particularly regarding the manufacturing of stents and other medical devices, the addition of suitable magnesium elements ensures they can perform their biomechanical role of supporting the vessel walls during the remodeling process, which occurs within 6 to 12 months [11]. Once the stent completes its purpose, the corrosion product of the Mg alloy generated by the electrochemical reaction is absorbed or excreted by the surrounding tissues and by the metabolic system of the human body [12]. Different magnesium alloys can be made by adding different elements, depending on the desired properties. It is common to add aluminum (AZ31, AZ91, LAE422, AM60, etc.) and rare earth elements (WE43, AE21, etc.) to magnesium alloys in order to optimize grain size, improve corrosion resistance, provide mechanical strength through intermetallic formation, and facilitate their processing [56, 57]. As an example, finer grain size alloys promote improved ductility, because they activate deformation mechanisms other than those commonly observed for basal slip and twinning [58]. Growth restriction can also achieve grain refinement during solidification, contributing to a homogeneous distribution of intermetallic constituents and facilitating subsequent thermo-mechanical processing [59, 60]. According to numerous scientific publications, Al, Ca, Mn, Y, and Ce are the most used chemical elements in magnesium. These elements can react with Mg to form intermetallic phases, which can dissolve in the grain matrix or distribute along the grain boundary, thereby influencing their mechanical properties and corrosion behavior. Aluminum is the most used alloying element for Mg alloys, presenting a maximum solubility of 12.7 wt.% in Mg and constituting several Mg-Al-based alloy systems that have been used for industrial applications. Dobrzanski et al. [29] and Agarwal et al. [56] report that only Al additions of 1-5 % can affect the grain size, making them smaller. The as-cast Mg-Al alloys show α-Mg matrix 14 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS and β-phases predominantly in the form of Mg17Al12 phases, which precipitate along the grain boundaries as a result of the Al dissolution in Mg solid solutions. These secondary phases increase with increasing Al contents, exhibiting a ‘net-shaped’ distribution when Al amounts exceed 3 % [29]. The influence of the Mg17Al12 phase on the corrosion resistance is not linear. It can either accelerate the corrosion of the alloy by acting as a cathode concerning the α-Mg matrix [56, 61], or it can reduce corrosion by acting as a corrosive barrier with respect to the β-phase volume fraction, which is higher and distributed along the grain boundary [56, 62]. Moreover, high Al contents can improve the castability of Mg alloys by lowering the liquidus and solidus temperature lines [63]. The addition of calcium (Ca) promotes grain refinement. According to Li et al. [64], the optimal concentration should be less than 1 %. When Mg-Ca alloys have a high Ca content, the Mg2Ca secondary phase which forms at grain boundaries is brittle, reducing ductility and corrosion resistance due to the formation of micro-galvanic cells [56, 64]. Additionally, CaO can be added to Mg alloys to produce ECO-Mg (Environment-Conscious Magnesium) alloys. According to the purpose, these master alloys with a high weight percentage of CaO are prepared in advance and then mixed and diluted to produce ECO-Mg alloys in the CaO range of 0.2 % and 1.5 % [29]. Numerous studies have already shown that adding CaO to Mg alloys has several advantages, including no-SF6 processing, improved melt cleanliness, improved mechanical properties through grain refinement, improved oxidation and ignition resistance, improved recyclability, and cost reduction [65-69]. Likewise, manganese (Mn) is a very common alloying element used in Mg alloys to reduce the grain size. According to Ding et al. [61], Song and Atrens [70], and Walker et al. [71], Mn addition can also improve the mechanical and corrosion properties of extruded Mg alloys, but at controlled levels. Moravej and Mantovani [15] and Wang et al. [55] also report the addition of RE (rare earth) elements such as Y and Ce, as well as other metallic elements including niobium, lithium, zirconium, and calcium. It aims to manipulate and control the microstructure of the magnesium alloy (grain size and intermetallic morphology - Mg17Al12 refinement) in order to increase their corrosion resistance and mechanical properties [72, 73]. The RE elements are a group of 17 elements, typically added to Mg alloys as master alloys [29]. Master alloys, also commonly referred to as hardeners, are pre-alloyed concentrates or mixtures of alloying elements used to add all the major alloying elements in one form to the melt to produce a particular alloy, to modify a melt, or CHAPTER 2. BACKGROUND LITERATURE 15 to alter processing characteristics to achieve desired properties. For instance, Al-Ti-B (aluminumtitanium-boron) master alloys are known to be grain refiners in aluminum alloys. Indeed, some reports have also evaluated its effect on the mechanical properties of AZ (aluminum-zinc) magnesium alloys, concluding that the grain sizes can be reduced effectively [74-76]. The application of an ultrasound technique (UST) is also pointed as a solution for grain refinement, melt cleanliness and degassing, as well as molding filling [77]. Acoustic energy has been the subject of many scientific studies, including studies of magnesium alloys. For instance, higher mechanical properties were achieved with the AZ91D magnesium alloy by microstructure refinement, ultrasound melt treatment [78], and degassing techniques [79]. In fact, one of the most used ultrasound applications is the degassing process, which involves the release of gas. This operation is particularly important when processing light metals and their alloys. These materials, when in the liquid state, present great affinity to the absorption of gas through their surface layer. Indeed, magnesium alloys' melts are characterized by a high hydrogen solubility, making them very suited to pick up gases from the atmosphere. Studies conducted by Lee [80] and Mikucki and Shearouse [81] have found that even a small portion of hydrogen may affect the formation of microporosities, while larger pores may form at the grain boundaries or near the second phase if hydrogen content exceeds the solubility limit at the heading face of liquid-solid [79]. This issue is more common in high alloyed magnesium alloys, such as AZ91, since these materials exhibit a wider solidification range, allowing for more time and space for the nucleation and growth of gas pores – the weak spots of the alloy. Argon degassing itself in magnesium alloys can cause a strong disturbance, increasing the number of melt slags, as these materials are highly reactive, oxidize, and burn. The use of ultrasonic vibrations allows the breaking of the large argon bubbles into smaller ones and disperse them uniformly in the melt, leading to an efficient degassing process that reduces the melt loss, as well as the intense flow turbulence. The high-intensity ultrasonic vibration promotes oscillating pressures that create multiple small cavities in the liquid, which grow fast in response to the alternating pressure and the unidirectional diffusion of the dissolved hydrogen from the melt to the available cavities. This phenomenon, schematically described in Figure 2.1, results in the formation of cavitation bubbles, which then coalesce and float to the surface, driven by the differential density and acoustically induced flows in the melt. This process leads to the release of hydrogen gas produced by the interaction of hydrogen atoms inside bubbles, as well as the reduction of the porosity of the casting [82]. 16 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 2.1: Degassing scheme: (A) dissolution of the argon (Ar) bubbles by the US system; (B) diffusion of hydrogen atoms and inclusions; and (C) flotation of the gas bubbles and release of the hydrogen gas to the atmosphere. According to Liu et al. [79], applying ultrasonic argon degassing allows the ultimate strength to be highest while maintaining the yield strength. Additionally, the ultrasound treatment greatly impacts grain refinement, noticing the size reduction and shape modification of the β-Mg17Al12 phase, the instigation of a spherical shape, as well as improved uniformity and distribution in the α-Mg matrix. Du and Zhang [83] showed that the uniform distribution of the fine β-Mg17Al12 intermetallic phase enhances mechanical properties, while gas pores and microstructural non-uniformity lower them. Stress concentration in bulk materials can be reduced by reducing these defects, while ductility can be improved by preventing early fractures. There are two distinct mechanisms associated with grain refinement efficiency: (i) heterogeneous nucleation at high (above liquidus) temperatures, and (ii) intermetallic compound fragmentation due to the streaming acoustic effect. The co-existence of these mechanisms suggests an enhanced refinement of dendrite cell size, thinning, and dispersion of β-Mg17Al12 intermetallic phase in the α-Mg matrix during the phase of solidification under the effect of ultrasonic vibration. During the first stage of solidification, the ultrasonic treatment improves grain refinement by heterogeneous nucleation as well as the wettability of the β-Mg17Al12 by the metal, further improving secondary phase refinement. During a second step, which corresponds to the formation of the first solid metal, cavitation can produce acoustic streaming due to the collapse of CHAPTER 2. BACKGROUND LITERATURE 17 bubbles in the remaining liquid, resulting in the fragmentation of grain dendrites and intermetallic clusters. As a result of the coexistence of these two mechanisms, a high density of nuclei in the melt is also promoted, resulting in a large number of smaller grains and, thus, in a significant improvement of the mechanical properties [84]. The pouring of the alloy is also improved, avoiding coarse and dendritic structure which harms the mechanical properties of the material [78, 79, 85]. This situation, associated with the increase of elongation, is favorable to the stent’s manufacturing, as it decreases the probability of device collapse during the deployment procedure, providing, however, the possibility of promoting its expansion with lower pressures. In addition to this, the increased mechanical properties of the AZ91 magnesium alloy enable the use of thinner struts and lower thicknesses comparable to stainless-steel products, which would not otherwise be possible due to the inferior mechanical properties of the material as-cast. In addition to alloying and UST techniques, other methods are used to improve the properties of magnesium alloys depending on their application. According to the literature, an indepth investigation has been published to improve their corrosion resistance. Some studies refer that another way to control the biodegradation rate of magnesium stents is by coating the surface with biodegradable polymeric layers [86-90]. The drugs can be released during polymer breakdown or even through the coating diffusion mechanisms, thereby increasing the overall degradation time of the stent. On the other hand, the mechanical behavior and corrosion resistance of magnesium stents are strongly influenced by the geometry and dimensions of the stents, as well as the microstructure of the alloys [91-93]. All of these types of methodologies can be of great assistance in the development and modification of new magnesium alloys with higher physical, chemical, and mechanical properties, making them ideal, among others, for the replacement of the most used metallic structural materials (iron, steel, and aluminum). However, improving materials also requires the demand for new and improved manufacturing processes. Given the decreasing cost of aluminum and magnesium alloys, there has been a higher investment in improving the manufacturing processes of these materials [33, 94, 95]. The purpose of this is to improve the quality and complexity of the obtained castings [96] and improve their mechanical properties directly, for instance, by modifying the gating system, which reduces porosity [97]. Although the industrial process is constantly evolving and researchers are learning more about the advantages of using magnesium alloys, there 18 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS is a need for improvement and the creation of new manufacturing processes, capable of meeting the market's high expectations both in terms of final product quality and quantity. Mahi Sahoo [98] stated that most of the magnesium alloys’ processing was done through HPDC (High Pressure Die Casting) in the previous decade, being directed to large volume part productions. This observation was later corroborated by Luo [99] and has remained until today, as confirmed by Dobrzanski et al. [29]. The possibility of high production volume and low production cost can justify this fact, but magnesium and its alloys can be processed through solid, powder, or liquid phases. A variety of stent manufacturing processes are discussed at the end of this chapter. Nevertheless, despite all the advantages, solid-phase processing has high processing costs, limitations on thickness, and the need for complex handling of fine powders, so liquid phase processing is more common. Casting is a millennial technology that, although still a widely used process to obtain all kinds of casting pieces using the same concepts over the years, has tried to adapt to the demands of industry and technological development. Depending on the type of application and the final component to be obtained, the casting process also comprises different forms of processing. Magnesium alloys can be cast by gravity, by low pressure, by injection, among other methods. According to Carvalho Ferreira [100], the casting process can be classified based on the permanence or perishability of models and moldings. In this thesis, the model, obtained by additive manufacturing, is considered the printed part presenting the geometric form of the final piece to be obtained, molding is referred to the manufactured part containing the cavities in which the metal is poured, and mold-metal is related to the interaction zone between the liquid melt and the molding walls during the pouring phase. In permanent molding casting, there is no model production, and the desired final form is achieved by pouring metal directly into the molding. In this case, the molding is not destroyed and has appropriate mechanical characteristics that are generally made of metal or other materials with high wear resistance. In this process, although many parts can be cast without their characteristics or shape being significantly changed, the drawback is that permanent molds are required for each component to be cast, which is a more time-consuming and more expensive process than ceramic molding. For non-permanent moldings and permanent models, it is necessary to use ceramic or sand moldings that can be removed from their interior before the liquid metal is poured. The molding is destroyed after the metal is poured, i.e., the metal is removed from the molding. Lastly, casting with non-permanent molding and nonpermanent model is characterized by the fact that the model is involved in the construction of the ceramic molding and is also destroyed, usually by sublimation, in order to create the cavities in CHAPTER 2. BACKGROUND LITERATURE 19 which the liquid metal will be poured. It is also known as investment casting, and it is the only type of processing employed throughout this work. 2.2. Magnesium Investment Casting in Ceramic Molding Investment casting is a manufacturing process capable of producing pieces from practically any metal or alloy and economically, compared to other manufacturing processes [43]. However, nowadays it is not normally used to manufacture medical devices, including biodegradable ones, such as stents. The high reactivity of biodegradable alloys, especially magnesium and its alloys, as well as the disadvantages the process can present in terms of surface quality, dimensional accuracy, and filling capacity for thin and complex walls that require complementary finishing operations have made this process unsuitable for the manufacture of small biomedical devices. However, the increasing development of technologies in recent years has altered the landscape of manufacturing processes and the knowledge about them as well. New technologies and equipment have been developed, allowing to redesign and improve many existing processes. As part of the near-net-shape stent's processing, a new approach can be defined that uses investment casting (using non-permanent models and molding) as a potential and practical solution to manufacture magnesium-based castings. Dobrzanski et al. [29] report that besides offering high-quality complicated shapes with a broader range of dimensions and geometries [94], investment casting is also seen as a promising process for the manufacture of thin-walled components that require a good surface finish and a high degree of dimensional tolerance [101]. There are, however, some barriers to overcome in magnesium investment casting using plaster moldings. Due to the significant reactions between magnesium and the ceramic moldings used in the conventional investment casting process, magnesium alloys are not widely applied in these industries. The fact that there are not yet suitable and effective inhibitors to control these reactions makes it impossible to handle magnesium alloys in foundries. The reactions can strongly affect the surface finish of the castings, reducing their final quality. Therefore, it becomes imperative to prevent mold-metal interactions during investment casting. It remains the biggest challenge to overcome due to the high affinity of magnesium for oxygen, specifically for the silicon dioxide (SiO2) present in the plaster composition, which causes a set of strong exothermal reactions [47-49]. The most common reactions, which can occur during the interaction of the liquid (from magnesium 20 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS and its alloys) are represented by Eqs. 2.1-2.4, but other reactions may also take place, depending on the affinity of the liquid magnesium reaction with the constituent elements of the ceramic materials used and the surrounding atmosphere [50]: 2 Mg (l) + O2 (g) → 2 MgO (s) Eq. 2.1 2 Mg (l) + O2 (g) → 2 MgO (s) Eq. 2.2 2 Mg (l) + SiO2 (s) → 2 MgO (s) + Si (s) Eq. 2.3 4 Mg (l) + SiO2 (s) → 2 MgO (s) + Mg2Si (s) Eq. 2.4 The amount and intensity of reactions occurring during the melting and pouring phases of liquid magnesium are also strongly influenced by pressure and temperature conditions. High pressure and temperature significantly increase reaction formation. These reactions can occur due to different situations during the investment casting process, caused by the interaction between the liquid metal with the ceramic molding walls or the surrounding atmosphere. Some approaches have been suggested and implemented in the last decade to minimize the reaction issue. The use of protective atmospheres was reported by Lee et al. [102], while Jafari et al. [47, 103] described the use of shell investment molds as an effective way to reduce these phenomena. Based on silica or zirconia binders, ceramic shell moldings were applied by Zhang and Morin [104] to avoid mold-metal reactions. It was concluded that an atmosphere of SF6 in CO2 is absolutely necessary under an inhibitor gas atmosphere. Furthermore, these techniques are highly dependent on the materials used for the moldings and the alloys used for casting. More recently, it was proposed by Jafari et al. [49] the application of an in situ melting technique for AZ91D magnesium alloy casting, applying a slurry composed of zircon flour and colloidal silica, along with fine alumino-silicate stucco sand. Moreover, a flux of MgCl2, KCl, and CaF2 was used in combination with small-sized granules of the alloy. Although the processing conditions avoided magnesium reactions, the obtained castings revealed defects such as micro-shrinkage porosity, inclusions, and non-fused metal parts. The application of refractory coatings in the investment casting of highly reactive alloys has also been proven successful in overcoming the inherent difficulties of this process, as shown by the use of ZrO2∙SiO2based crucibles with an inner layer of Y2O3 to melt a Ti-48Al alloy [105, 106]. It was concluded that chemical composition and alloy homogeneity are strongly influenced by CHAPTER 2. BACKGROUND LITERATURE 21 processing conditions such as processing melting pressure and superheating time and temperature. A protective flux or protective gas must cover the molten magnesium to prevent it from burning. Flux protects from burning, but flux inclusions trapped in the casting can pose a problem [107]. In terms of protective gases, there are a few available in use today, but SF6 (sulfur hexafluoride) is still the most efficient at protecting molten magnesium. It is non-toxic, noncorrosive, and provides adequate protection against molten magnesium at very low concentrations. Typically, it is mixed with a carrier gas such as dry air and/or CO2. However, even though it is possible to reduce its concentration by combining important processing conditions, such as mold and pouring temperatures, SF6 is a powerful greenhouse gas that negatively impacts the environment [32]. Consequently, it was reported by Ha and Kim [108] that using the HFC-134a gas (1,1,1,2-tetrafluoroethane) provides an effective alternative to SF6. Nevertheless, a gas mixture must be used to prevent the formation of highly corrosive hydrogen fluoride, which can reduce the process’ effectiveness. On the other hand, it has been demonstrated that the addition of a small amount of CaO to Mg alloys, forming Eco-Mg alloys (as mentioned in the previous subchapter), is a non-SF6 process during melting and casting of Mg alloys that significantly increases the ductility and strength of these alloys [66, 67]. Nevertheless, none of these studies reported effective solutions to overtake the reactivity between ceramic-based moldings and magnesium alloys, which is still poorly understood. According to Jafari et al. [47], the progress that has been made does not correspond to a practical and economical alternative to encourage the investment casting foundries to handle magnesium. Thus, searching for a process that effectively mitigates the interfacial reactions and allows obtaining complex and thin-walled magnesium alloy parts constitutes a great stimulus of investigation for the scientific community. To overtake these inherent limitations, it remains necessary to develop an economical, practical, and environmentally friendly method of avoiding mold-metal interaction and suppressing the reactions that result. In this way, the combination of additive manufacturing with investment casting process can be seen as a route for fabricating thin-walled magnesium parts in several industrial applications, such as the manufacture of medical devices. 28 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Casting is only used as a primary process in stents’ manufacturing. As a result, since the as-cast materials have a very coarse microstructure, forming and thermomechanical processes are very important to achieving the desired mechanical properties. There are many processes conducted at temperatures at which the alloy is easily shaped, to improve mechanical properties by means of plastic deformation and hardening. Some of these processes include forging, rolling, and extrusion. After hot working, metallic materials are cold worked and heat-treated to obtain ingots with higher physical and mechanical properties [149]. Ingots are then formed into tubes using forming techniques before being cut by lasers, which enables the stent's design to be defined. Minitube formation The next step concerns the formation of tubular components (semi-products) that will support the fabrication process. It is crucial to select a tube processing that can guarantee excellent mechanical properties in the final stent since this involves microstructural modification of the alloy used. At the industrial level, rolling, drawing, and casting processes are often used to produce seamless tubes. However, whereas these processes are used for large-scale products, the feasibility of components such as stents, that demand high dimensional and geometry quality, does not have the required viability. For fabricating permanent metallic stents, there are basically four steps to follow, including extrusion, machining, and drawing to form tubes [150]: (1) obtaining a solid rod by hot extrusion of the molten feedstock; (2) drilling a deep hole in the rod to obtain the tubular shape; (3) indirect extrusion of the tube; and (4) reduction of the diameter and wall thickness by multiple drawing and intermediate tube annealing. Tube drawing is a typical metal forming procedure that involves stretching a hollow billet under tensile stresses to reduce its cross-sectional area and increase its surface finish and dimensional accuracy. Therefore, this type of procedure requires that the material exhibit high ductility and sufficient tensile strength allowing plastic deformation to occur before fracture. Metallic biodegradable stents require advanced fabrication technologies, which have not received enough attention. For biodegradable Mg-based alloys, which are materials presenting low formability, it is challenging to produce cold-drawn seamless tubes with extremely dimensional rigor due to their poor plasticity at room temperature. Due to its closed hexagonal crystalline structure characteristic and poor room temperature workability, it is common to use hot drawing to form semi-products of this material [151]. However, using a hot mandrel can cause CHAPTER 2. BACKGROUND LITERATURE 29 uncontrollable increases in the wall thickness of the magnesium tube and poor surface quality. Only a few studies have so far been conducted on developing fabrication technology for magnesium microtubes and improving their mechanical properties, as well as on the dimension accuracy of the stent’s semi-finished products [5, 11]. The manufacture of stents requires complete control of all properties and characteristics, and this process does not necessarily satisfy these requirements. As a result, some authors have adopted other processes for the manufacture of minitubes from metallic alloys, including different drawing procedures [152-155] and other tube forming techniques, such as electric discharge machining [156, 157] and SPD (Severe Plastic Deformation) [158-166]. Concerning conventional drawing procedures, some studies have been published in order to reduce tubing size. Wall thickness reductions had been achieved by Furushima and Manabe [152] from hollow billets of the AZ31 magnesium alloy, and tubes with an outside diameter of 2 mm and a thickness of 0.5 mm were obtained through hot drawing without a die. However, it was concluded that in hot tube drawing without a die, uniformity in wall thickness could hardly be kept under control, and the inner surface of drawn tubes would not be perfectly defined. Alternatively, Yoshima and Koiwa [153] fabricated tubes with an external diameter of 3.6 mm and an apparent thickness of 0.61 mm using the AZ31 magnesium alloy at room temperature, applying the fluid as a mandrel. Through multipass cold drawing, Fang et al. [155] were carried out to produce magnesium alloy minitubes using the ZM21 alloy for stent manufacturing. The drawn tubes with an outside diameter of 2.9 mm and wall thickness of 0.217 mm were obtained through five passes of cold drawing with an interpass annealing procedure after the fourth pass. Accordingly, although it is technically challenging, the literature review allows to conclude that in order to obtain minitubes for biodegradable magnesium stents, the cold drawing appears more feasible than hot drawing since cold drawing is better at achieving close dimensional tolerances and a fine surface finish, as well as high mechanical properties [155]. Nevertheless, the values obtained in these studies appear insufficient to meet the stringent needs that medical devices have at the microscale level. Several improvements need to be made, mainly in reducing the thickness of the stent (corresponding to the minitubes thickness) to values close to the 0.1 mm that permanent stents can present, and controlling degradation rates of biodegradable materials, which considerably affect the final device's mechanical properties. SPD processes are thermomechanical processing techniques with great potential for producing bulk samples with extremely small grains, improving the homogeneity of the impurity 30 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS distributions in the alloy microstructures [167]. During SPD processes, the samples' overall shape is not changed, making it possible to ensure that the repetitive severe plastic deformation refines the microstructure through a continual dynamic recrystallization process. SPD processes are therefore applied to all metals, both permanent materials such as stainless steel and aluminum [158, 159] as in biodegradable alloys such as magnesium [160, 168]. Several studies have reported that grain refinement caused by these processing techniques can enhance mechanical properties [169, 170]. Among the most studied, the ECAP (Equal Channel Angular Pressing) [161] and the HPT (High-Pressure Torsion) [162] are the most SPD promising processes. While the ECAP technique allows the production of larger samples, the HPT technique generates high hydrostatic pressures, allowing the processing of materials with low ductility. But other SPD techniques can be used, such as CEC (Cyclic Extrusion and Compression) [171], TCAP (Tubular Channel Angular Pressing) [164], and ABE (Accumulative Back Extrusion) [172]. Several studies have demonstrated the benefits of using the SPD processes in the alloy's properties for biomedical applications. Ge et al. [169] applied ECAP for manufacturing a biodegradable magnesium stent prior to extrusion and laser cutting. The process was used to achieve a significant grain refining in the submicrometric grain-size range and the 0.2 % yield strength increased from 180 to 340 MPa after 150 ºC ECAP processing, keeping a relatively high tensile ductility. On the other hand, Figueiredo et al. [166] described the development of exceptional ductility in pure magnesium after HPT processing at room temperature. The results showed that grain refinement by severe plastic deformation resulted in a higher strain rate sensitivity and a more significant elongation at room temperature. All of these studies have shown that the use of SPD methods at elevated temperatures mixed with other forming processes is considered a very promising approach for improving the mechanical properties of magnesium alloys through grain refinement. According to this, Amani et al. [5] proposed a new combined method of SPD techniques to fabricate magnesium microtubes for biodegradable vascular stents. This method consisted of three forming processes and a drilling step. The WE43 Mg alloy was applied for three SPD techniques: CEE (Cyclic Expansion Extrusion), EX (Direct Extrusion) and MTE (Microtube Extrusion). Three microtubes with outside diameter of 3.3 mm and wall thickness of 0.22 mm were successfully fabricated in three different ways: from the initial material, and through one pass or two passes of CEE technique. In comparison with the original microstructure, the authors were able to significantly reduce the alloy grain size and reduce the cross-section of the samples. Regarding mechanical properties, a significant increase in the maximum strength and elongation was achieved, which was recorded in the final samples. CHAPTER 2. BACKGROUND LITERATURE 31 However, even though SPD processes have proved beneficial in improving mechanical properties, there is no definition or knowledge of how they impact corrosion behavior. Some studies have shown different trends in corrosion behavior using severe plastic deformation methods [173-175]. For instance, Ralston et al. [176] demonstrate in their study that the grain refinement achieved by SPD techniques can lead to a higher corrosion rate of the alloy, affecting the mechanical properties and compromising its use and application. Although the minitubes fabrication for the manufacture of stents from metal alloys may encompass several different processes, considering the aforementioned, there have been no studies and consistent investigations concerning metallic minitubes for medical devices [135]. Cutting and stent design The stent mesh design can be achieved by a number of methods, including photochemical etching, water jet cutting, and electric discharge machining. However, laser cutting is nowadays the most widely used method for metal stents production, with considerable flexibility of shapes and geometries in an extensive production line [177-179]. It is a type of material solid processing that produces small diameter minitubes and thin-walled patterns. Cutting micro stents is a demanding process in the medical device industry. A coronary stent should be between 1 mm and 10 mm in diameter and 0.1 mm in thickness. For this reason, the cutting process must be high in precision. Cutting stents must not have imperfections that can lead to infections in the blood vessels into which they are inserted. The majority of laser-cut stents are either slotted tube designs or modular constructions, and different stent designs are available on the market from different manufacturers [180]. The laser cutting manufacturing technique has many advantages that make it a viable method of obtaining stents. However, it has limitations that compromise the specificity required for biomedical devices, namely at the level of surface quality and thermal issues. These issues directly affect the metal's properties and its corrosion. The heat generated by the laser can cause thermal damage to the application area due to high thermal gradients. This causes deformations, microcracks, burrs, porosities, residual stresses, and dimensional inaccuracy (Figure 2.4), compromising stent properties [181]. Some studies proved that laser cutting can also induce loss of alloying elements from the processed material [182], and that the resulting heat from this manufacturing process is associated with the grain thickening on the stent surface, affecting its mechanical properties [6, 148, 183]. Moreover, the tubing step associated with the conventional laser cutting manufacturing process results in the formation of relatively large 32 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS inclusions in the final material. This poses a series of risks of rupture failure after stent expansion [148]. Figure 2.4: Principal defects of laser cutting: (a) deformations; (b) burrs; and (c) porosities (adapted from [15, 184, 185]). In trying to minimize these problems, additional thermomechanical processes and finishing procedures, such as annealing, chemical etching, electropolishing, and sterilization, among others, are necessarily performed [15, 145]. Furthermore, the stents can be subjected to heat treatments such as vacuum annealing after laser cutting. This will soften the metal, induce grain growth, and relieve residual stresses created during laser cutting, which improves the stent's physical and mechanical properties [15]. In the manufacture of metal stents, the heat treatment procedure is critical because it directly influences the properties of the materials, while other processes determine the geometry, dimensional accuracy, and surface quality. In order to obtain homogeneous grain composition and distribution, as well as fine grain size, annealing requires different temperature and actuation time parameters. These parameters are according to the type of material to be treated. In addition, finishing processes are also helpful for removing any mechanical defects on the stent surface. This is designed to have a roughness value of less than 0.5 µ m [10]. Hence, studies have been conducted to determine the ideal process parameters for laser cutting metallic alloys. These parameters are mainly in terms of dimensional accuracy and surface finish [15, 148], and process parameters [25, 109]. Finishing operations Surface finishing is the final step in manufacturing stents and is very important to enhance product quality, especially in the case of biomedical devices, whose surface characteristics determine the nature of immediate and long-term tissue response [163]. Surface treatments may be electrochemical, chemical, or mechanical, depending on the application. As a result, these processes improve the surface quality of the stent and its corrosion resistance and biocompatibility, (a) (b) (c) CHAPTER 2. BACKGROUND LITERATURE 33 removing contaminants and mechanical imperfections that can occur when laser cutting or other non-conventional method is used. Once again, different finishing processes can be applied depending on the surface characteristics to be improved. As an example, Figure 2.4(c) shows the application of electropolishing after laser cutting. Electropolishing is the conventional finishing procedure used to fabricate stents and is usually used to achieve a smooth surface and remove burrs and defects generated by the heat ablation during laser cutting, etching, or forming steps [15, 163]. It is a controlled and repetitive electrochemical process that removes metal from the surface of complex geometric objects using electrolytic dissolution. As a result, the required surface finishing is ensured, and the stents’ biocompatibility and corrosion resistance are improved. For example, in the manufacture of permanent NiTi (nitinol) stents, electropolishing selectively decreases the amount of nickel on the surface. The passivation is composed of titanium oxide, which improves the biocompatibility of the stent since nickel is harmful to the body due to its association with cancerous and inflammatory reactions. Although electropolishing has been considered a successful method of improving the stents' quality finish, some other techniques can be used after the laser cutting process, such as chemical etching, MAP (Magnetic Abrasive Polishing), ultrasonic cleaning, honing, sterilization, and others. By using these procedures, high-quality biocompatible stents with a bright, shiny surface, defect-free, and improved corrosion resistance are obtained. Chemical etching with nitric acid and ethanol is applied to clean the spatter deposited on the stent surface [178]. MAP can also be applied using flexible tools, including iron powder and abrasive particles [186]. It is characterized as a super-finishing technique to achieve a nanometric level of surface finish. A controlled magnetic force of extremely small magnitude is used in ferromagnetic abrasive particles, which are a conglomerate of abrasives and iron particles for material removal. It is widely used for ultrafine polishing of non-magnetic stents. Ultrasonic cleaning removes surfaces’ contaminants such as oils, fingerprints, and dust. It is a suitable surface finishing process for stent application since it allows cleaning in difficult access and small diameter areas. During this procedure, the stent is submerged in a solution under high frequency sound waves. Honing is an abrasive machining process performed immediately after laser cutting, producing a precision surface on the metal part to which it is applied. It consists of rubbing an 34 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS abrasive stone in the tested sample along a controlled path and is mainly used to improve the geometric shape and surface texture. In stents manufacturing, the honing process is applied to permanent stents to remove the remaining material between supports and soften the resulting laser rough edges. Finally, sterilization is a process of eliminating or removing all forms of life and other biological agents presented on the surface product. Because it can change the surface properties, this procedure should be the last step in manufacturing a biomedical device [171]. Can be achieved through various ways, including disinfection by ethylene oxide before implantation [187], sterilization by g-radiation exposure [188, 189], and UV (ultraviolet) radiation [190]. In terms of applications, a recent study by Liu et al. [191] focused on the effect of various sterilization methods on the surface characteristics and biocompatibility of pure Mg and Mg alloys. The authors concluded that the Co60 γ ray radiation sterilization may be the most appropriate technology in reaching the best properties for biomedical magnesium alloys. Nonconventional stent processing The biomedical sector requires better, stricter, and smaller devices, so engineers are increasingly challenged to implement better fabrication processes. Several new processes have evolved or have been improved to meet this requirement, leading to realistic alternatives for producing metallic stents that are designed to overcome the conventional laser cutting limitations. Butt welding is a technique that involves microcutting a mesh structure from a solid metal tube or welding preformed wire sections in a composite structure. This process is typically used for permanent stainless steel or cobalt alloy stents, although there has been some interest in applying this process to materials such as magnesium for manufacturing biodegradable stents. On the other hand, some authors favor the production of metallic stents through metallic elementary powders using near-net-shape processes, which reduces the number of manufacturing steps required and, as a result, lowers the total cost of the entire process [192]. For instance, PM (Powder Metallurgy) manufactures shaped components or semi-finished products from metal powder. Here, fine powdered materials are blended inside a mold, compacted into the desired shape, and then heated in a controlled atmosphere, referred to as sintering, in order to facilitate the bonding formation of powder particles in the final part [193]. Alternatively to this conventional PM process, SLM (Selective Laser Melting) or MIM (Metal Injection Molding) can also be used. CHAPTER 2. BACKGROUND LITERATURE 35 SLM is a powder bed fusion technique in which a laser scans the powder layer by layer to generate the desired geometry. SLM's key advantages include shape flexibility and the use of lattice structures [194, 195]. Based on powder bed fusion, this method offers an attractive way to manufacture stents directly from powder, decreasing laser cutting and minitube manufacturing steps into a single process, and reducing production times and enhancing geometrical flexibility [10]. However, additive manufacturing techniques for vascular devices have been limited to a few applications using polymeric materials. There exists an apparent need for a comprehensive analysis of available and novel additive manufacturing techniques for metallic materials with adequate geometrical precision. Wessarges et al. [196] used fine powder (5–20 µ m) with a micro SLM system to produce prototype stents of AISI 316L stainless steel, which were subsequently finished by plasma and chemical polishing. Despite surface cracks after expansion, the prototyped stents showed promising mechanical performance. MIM is a well-established powder metallurgy technology and a viable alternative to other processes [197]. As a near-net-shape process, it has the advantage of allowing significant reductions in production costs, as well as being considered a suitable technique to produce small and complex parts in large quantities [198, 199]. It is an invention derived from the idea of plastic injection molding, in which the metal powder particles are mixed with a binder and injected into a cavity with the desired shape [200-202]. First, powders and binders are mixed to produce feedstock, and then a green compact is formed through injection molding. Afterward, the binders are extracted through de-binding to form the sintered final compact. The MIM process seems to be a very promising manufacturing method for the commercial scale of medical devices made from biodegradable alloys, although its true potential has not yet been sufficiently explored. A significant disadvantage of this method is the presence of pores during the sintering phase, which facilitates corrosion. The corrosion resistance can thus be affected by the heat treatment applied to the material, as stated by Hamidi et al . [199]. However, research involving biodegradable metals for biomedical applications has been in development since the 21st century beginning [203]. In recent years, micro MIM technology has been developed to reach thin walls of 20 µ m and surface roughness values of less than 0.05 µ m [204]. This indicates that, from a geometric point of view, this technology can meet the requirements for stent manufacturing. Moreover, its near-net-shape concept makes it particularly suitable for developing complex geometries with high dimensional accuracy [205-210]. 36 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS PM techniques can thus be considered an attractive alternative manufacturing process in stent fabrication from an economic standpoint. Furthermore, powder metallurgy might be advantageous for metallic biomaterials like titanium, magnesium, or nitinol, which are more difficult to machine with conventional methods [193]. However, the main disadvantage of these processes is that they are restricted to specific geometries and dimensions, which inhibits greater investment in their use for thin-walled medical devices. Finally, the casting process should also be included in this section of unconventional processes for fabricating metallic stents. As discussed earlier in this chapter, casting is primarily used to manufacture stents in the conventional process, specifically in obtaining minitubes. However, it would be desirable and economically advantageous if stents could be cast as a nearnet-shape process. Therefore, and avoiding procedures such as minitube formation and heat treatment, stents could be designed to meet the market’s mechanical, physical, and chemical requirements by investment casting. So, due to the evolution of manufacturing processes and their adaptation for the fabrication of biomedical devices, processing metallic stents through the combination of additive manufacturing with investment casting can be viewed as a promising route for the development of these devices. 2.5. Summary and Conclusions This chapter describes the methods and techniques used for handling and improving magnesium alloys to manufacture stents. The addition of elements and the ultrasonic technique can be applied to improve the physical, chemical, and mechanical properties of magnesium alloys. Regarding processing, it is necessary to develop new methods and improvements for the highpressure die casting processes, usually used for magnesium alloys. In that way, and due to its advantageous dimensional and geometric characteristics, the investment casting process, combined with additive manufacturing, can be seen as a potentially effective solution for the processing of magnesium alloys, mainly in obtaining thin-walled parts, whose knowledge needs further exploration. Considering medical device manufacturing, which involves geometric complexity, different processes can be implemented, whether based on solid, powder, or liquid processes or through conventional or non-conventional methods. Most of these techniques are presented in Table 2.1. CHAPTER 2. BACKGROUND LITERATURE 37 Table 2.1: Advantages and drawbacks of different processing methods for obtaining biomedical devices. Process Type Advantages Disadvantages References Casting Liquid processing Any material can be cast Dimension and geometry complexity Economic Poor dimensional accuracy and surface finish Waste of material [51, 211] Laser cutting Solid processing Shapes and geometries flexibility Large production lines Insufficient surface quality Thermal issues [14-16, 51, 180] Machining Solid processing Variety of materials High mechanical strength and density Good dimensional accuracy Unavoidable defects High cost and maintenance [155, 157, 212, 213] Metal Injection Molding (MIM) Powder processing High production rate Good mechanical properties Good shape complexity with high dimensional accuracy Presence of pores and impurities Part size limitation [197-199, 214] Powder Metallurgy (PM) Powder processing Low cost Variety of materials High production rate Good dimensional precision Development of shape memory alloys (SMAs) Production of pores and residual porosity Size and shape limitation Long heating time [193] Selective Laser Melting (SLM) Powder processing Shape and geometry flexibility Reduced production time Higher energy level necessity Expensive Smaller scan velocities [10, 194196] Severe Plastic Deformation (SPD) Solid processing Effective grain refinement Improved mechanical properties Production of semiproducts Size and shape limitation Acceleration of corrosion rate [158, 166, 168, 170] Surface finishing Solid processing Surface quality Mechanical defects removing Low production rate and time [163] The current manufacturing process of stents, involving laser cutting, encompasses several processing stages, making it expensive, as stated by Hermawan and Mantovani [145]. Additionally, the thermal problems that arise from the laser can compromise the successful application of the devices, mainly if magnesium alloys are used. The sequence of operations required for each process is shown by the descriptive and summary flowchart in Figure 2.5. 44 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Wettability is another factor to consider when choosing a material for three-dimensional models. As the next step involves the application of a coating on the initial models, the printing material should present good wettability, because it is necessary that the coating adheres effectively to the model surface to ensure that all geometric details will be replicated. To assess this property, the surface tension of the liquid/solid contact needs to be determined, specifically, the surface free energy of the solid materials utilized. In order to evaluate the wettability value, it is necessary to use different types of measurement, for example, the contact angle measurements, used in the current study. Owens-Wendt-Rabel-Kaelble (OWRK) is one of the most widely used surface free energy theories based on Young's equation. It divides the interfacial interactions into two categories: polar and dispersive. To calculate the surface free energy with this model, the contact angle of two known liquids must be measured. One of them, with a dominant polar component, and the other as a dispersive liquid, since the interactions occur between similar components (if only dispersive liquid has been used, the potential polar interactions would not have been observed). Most commonly, water and diiodomethane are used. In this research, a mobile measurement system was used to measure the surface free energy of PLA and resin. Using the MSA (Mobile Surface Analyzer - KRUSS) portable equipment, the obtained results are presented and discussed in subchapter 3.3. With a double sessile drop measurement method, the used equipment captures the contact angles of the two drops concerning the polar and dispersive components (water and diiodomethane). After the software has processed the data, the surface free energy values are recorded. A smaller contact angle indicates better wettability of the analyzed material and, consequently, better adhesion to the coating. Figure 3.3 illustrates this process straightforwardly. Figure 3.3: Measurement of the contact angles. CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 45 Following 3D printing, the polymer models are assembled with sprues and channels (feeding parts) through a technique known as tree making in the investment casting process [94, 217]. The printed models must be placed in a “tree” form as possible. Although high detail precision is not required, the feeding parts can also be printed using the same technique. However, FDM is recommended for time and cost savings. There are some aspects to follow in the tree making to achieve successful castings: • For larger trees, single models must be fixed around the main spruce in a spiral shape to facilitate removing the cast pieces; • Spruces should be as short as possible; • The diameter of spruces should guarantee optimal casting feeding; • The PLA or resin connection to the spruce should be made in a hydro-dynamic style, to prevent turbulences while pouring the heavy liquid metal, breaking away thin parts of the plaster molding. 3.2. Coating and Molding Making According to the proposed methodology (Figure 2.6), coating the model and fabricating the plaster molding are the next steps. The application of protective coatings in the investment casting process is one of the mechanisms that are employed to mitigate the reactions that occur during liquid magnesium melting and casting, as described in subchapter 2.2. A simple coating of liquid solutions is applied, followed by a cured period in which the coating dries and gains consistency. Water and investment powders are then applied over the coated sample trees to create the plaster moldings. A variety of variables were tested in this investigation study to analyze the effect of coatings on mold-metal reactions during the investment casting of magnesium alloys. Having selected the best coating to use, other studies were examined, such as the effect of implementing different application methods and the effect of using a different number of coating layers. Based on the experimental evaluation, variables such as immersion time or curing time can be optimized during the procedure. Additionally, in order to avoid losing consistency, it is essential to ensure that the coated model is neither too dry nor too wet. Table 3.2 summarizes all the variables examined in this thesis concerning coating application. 46 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Table 3.2: Coating application characteristics. Coating Application Immersion time Number of layers Curing times Yttria Fused Alumina Sand-based slip Through dipping Through spraying 10 s* 1 layer 3 layers 5 layers 7 layers 3 h* * commonly used in industry Three different compounds were tested regarding the type of coating: Yttria, fused Alumina, and Sand-based ceramic slip. It should be noted that, in the present study, coatings were used only as face-coat slurries and not for shell production. The models were coated, followed by the production of the moldings. The main objective was to create a thin layer of refractory adhering to the plaster in order to prevent any contact between the liquid metal and the plaster from the molding that might affect the metal quality or the dimension accuracy. Although there have been some studies on the use of Yttria (Y2O3) especially for coating crucibles in the casting of reactive alloys [107, 218, 219], it is not widely used in industry due to its high cost and thermal shock brittleness [218]. Yttria coatings have, however, demonstrated superior results in preventing the formation of reactions and reducing bath contamination [107, 219], mainly because of its low free energy formation (ΔGo f = -898,7 kJ mol O2-1 at 2000 K). A considerable negative value of ΔGo f, also called Gibbs free energy, indicates a strong driving force to the compound’s formation and high-compound resistance to thermal decomposition (reverse reaction). Chemical stability generally increases as the melting point increases. Indeed, the high melting point is not so important in magnesium casting because of the relatively low casting temperatures, but chemical stability is critical. Hence, at magnesium melting temperatures, and considering the very negative values of Gibbs free energy, Y2O3 is one of the most stable oxides, making it ideal for use as a refractory coating. In addition, its application in this work for low-volume ratio models takes some weight away from the cost disadvantage. The coating of Yttria used in this research is a liquid solution with 99.99 wt.% purity (from ZYP coatings). Another two solutions were also tested and evaluated for the coating of 3D models. Fused Alumina (Al2O3) is considered an excellent cost-effective refractory product and a suitable ready-touse binder for investment casting. In this work, the VP Disp. W 640 XC8 (EVONIK) solution was used as a coating, which is an Alumina dispersion with high solid content, anionic stabilization, CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 47 and a neutral to basic pH. A first coating layer of Yttria solution was used as a binder before applying the Sand-based slip. A low-granulometry (0.15 - 0.3 mm) sand mixture was used for the refractory material. The models were coated according to two methods, dipping and spray painting, and were left to dry for three hours. Through dipping, the tree assembled samples were immersed in the liquid coating solution for 10 seconds and dried at a constant temperature of 30 ºC in a cubicshaped chamber (800 x 400 x 400 mm3). Applying the spray painting, the samples were coated manually using a commercial paint gun, followed by the same curing conditions. Depending on the number of layers required (Table 3.2), the procedure was repeated for each method. After curing, the assembled models (painted or not) were involved in plaster to produce the moldings. A mixture of water and investment powder (Omega+ from Gold Star) was made according to the conventional molding technique and then left to dry at 30 ºC before thermal processing. Here, a thermal cycle was applied in order to evaporate the ceramic materials (PLA and resin) from the molding interior and give consistency to the plaster, eliminating the water content. It is important to consider several variables when making a molding, such as the type of mixture, the mixing time, the pouring time, the type of water, etc. The ability to control each variable is crucial for time savings and sample repeatability, avoiding poor castings and unnecessary raw material expenses, especially in experimentation in which the same procedure is repeated several times. According to investment manufacturers and casting industry experts, changes in these properties may be the cause of casting defects. For instance, a long pouring time increases the probability of investments’ watermarks, especially if they are not mixed long enough. Having a weak investment will result in the formation of fins and a rough surface, resulting in heavy castings. The water can introduce impurities into the mixture. These properties are dependent on the waterto-powder ratio (w/p) and have a significant effect on casting results, so their control is vital. For this reason, investment manufacturers recommend higher w/p ratios for smaller and more detailed pieces as well as de-ionized water to remove interference compounds. As a result, a 40/100 (mass%) ratio of water (de-ionized) to powder has been employed in this study. Additionally, mixing and pouring were carried out in proper machines (MC - series by Indutherm GmbH), under a vacuum atmosphere to prevent the introduction of air bubbles into the moisture. 48 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS 3.3. Experimental Procedure The application of different additive manufacturing methods is presented in Figure 3.4. In general, it is evident that SLA has a higher quality and precision than FDM. So, for complex geometries and reduced dimensions parts, it can be concluded that SLA must be used. FDM cannot obtain the higher geometric and dimensional precision required by some pieces, even when support material is needed (Figure 3.4a) or an extrusion nozzle cleaning tower is used to prevent material drag (Figure 3.4b). Despite being cheaper, FDM has a poorer surface finish that SLA, apart from thickness limitation. In this work, for example, SLA must be used for stent printing and FDM can be used to print the feeding system. This will be explained further in subchapter 5.1. The castable resin used in the SLA method (JewelCast GRN) is an ideal green material for printing small patterns in plaster investment casting applications. Furthermore, this material provides superior casting quality with minimal ash after burnout and allows the production of finely detailed investment casting parts with a smooth surface finish. Figure 3.4: Additive manufacturing: (a) FDM with support material; (b) FDM using cleaning tower and (c) SLA. Based on the results of the wettability assessment, depicted in Figure 3.5 and presented in Table 3.3, it appears that along with better print quality, the resin also exhibits greater wettability, favoring the subsequent application of the coating. According to the contact angles captured by the MSA equipment for both the polar component (water drop) and the dispersive component (diiodomethane drop), the resin has a greater surface free energy (63.83 mN/m) compared to PLA (42.20 mN/m), using the OWRK calculation method. This means that resin is more wettable than CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 49 PLA and ensures better adhesion to liquids, i.e., to the coating layers. As such, the use of resin must be prioritized over the PLA for better coating adhesion, unless the application does not justify the higher resource expenditure. Figure 3.5: Contact angles measurement for PLA and resin. Table 3.3: Acquisition data for wettability analysis (one measurement reading). PLA Resin Surface free energy 42.20 mN/m 63.83 mN/m Dispersive 12.61 mN/m 45.58 mN/m Polar 29.59 mN/m 18.26 mN/m The coating type to be applied is shown in Figure 3.6 in accordance with the different refractories defined (Yttria, Alumina, and Sand-based slip). The samples were immersed in the liquid solutions for 10 seconds. For the Sand-based slip application, as described in the previous subsection, the sand mixture is applied after a coating layer of liquid Yttria, which is not allowed to dry in order to improve sand adhesion. Alumina (Figure 3.6b) and Sand-based slip (Figure 3.6c) were ineffective. The coatings were unable to protect from the reactions that occurred with the liquid magnesium, which resulted in totally failed castings, characterized by the occurrence of strong exothermic reactions that compromised the process. In both cases, the cavity filling was blocked due to the reactive contact 50 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS between the magnesium and the molding walls. These experiments are not recommended for user safety and equipment preservation, so their use was not repeated in this study. On the other hand, the Yttria coating effectively acted as a barrier to reactions, and a complete cavity filling was achieved. Considering the obtained stent (Figure 3.6a), it may be concluded that the application of Yttria is an effective solution to magnesium casting using plaster molding, although further experiments and an in-depth analysis of the process are still needed. Figure 3.6: Castings after different coating applications: (a) Yttria; (b) Alumina; and (c) Sand-based slip. The application method was tested after determining that Yttria is the appropriate protective coating to use. As displayed in Figure 3.7, both dip and spray methods fulfilled the objective of ensuring a uniform Y2O3 coating for the resin 3D models. In practice, however, the spray method does not appear to apply a layer as adherent to the solid as the dip method. A small degree of coating roughness can be observed when the spray was applied to coat the stent tree. The resulting castings also confirm these observations. The final cast stent showed some degree CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 51 of reaction and high surface roughness in the spray coating, although the cavity filling was not significantly affected. Alternatively, in the casting in which the model tree was dipped in Yttria, the absence of reactions and a sane molten metal is evident, resulting in a better casting. As a result, and based on the comparison of both methods, the dipping method is recommended to guarantee a more uniform coating for magnesium investment casting. Figure 3.7: Application of dip and spray coating methods. Finally, an experimental study was carried out to understand the influence of coating thickness. The procedure of dipping the sample and allowing it to dry was repeated several times, depending on the number of layers to be applied. Then, the plaster moldings were made. More than one coating layer is applied to reduce the Yttria permeability in order to prevent the metal from coming into contact with the molding walls. Figure 3.8. illustrates the obtained results after the thermal cycle. Increasing the number of Y2O3 layers led to worse results. Under the influence of thermal treatment, the outer Yttria layers detach from the inner layers, breaking up and accumulating into the molding cavities. Metal casting is considerably affected by this phenomenon, both in terms of surface roughness and metal contamination by Y2O3 particles released during the casting process. It can be concluded from the results of these castings that the stability of the Yttria coating can be affected by the temperature and/or by the volumetric contraction that occurs during 52 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS the thermal treatment procedure. As the thermal cycle for moldings assumes temperatures of 720 °C, the cracking is higher for thicker coating layers and larger molding cavities. Consequently, the application of coating as a refractory requires a compromise between the thickness of the layer and the temperatures involved. Figure 3.8: Application of different coating thicknesses. In the following subchapter, a more detailed analysis will be conducted to the mechanisms of the reactions occurring at the metal-liquid interface between the liquid magnesium and the molding walls, as well as the effectiveness of using Yttria for casting magnesium parts. 3.4. Case Study 1 – Effect of model’s coating on mold-metal reactions 3.4.1. Experimental procedure The experimental procedure was carried out through a vacuum-assisted investment casting process, which was based on the five steps belonging to the proposed methodology (Figure 2.6): (i) production of CAD models; (ii) manufacture and assembly of polymeric models; (iii) coating of models; (iv) preparation of ceramic moldings (plaster); and (v) casting of the magnesium alloy. The initial models were made by additive manufacturing, using the FDM process. Using the tree-making technique, models were printed in PLA and assembled with sprues and channels (also made by the same method). A disk specimen with a thickness of 2 mm was selected to ensure adequate cavity filling and evaluate the mold-metal interface. For reproducibility purposes, each tree consists of two specimens. Figure 3.9 provides the CAD corresponding to the tree assembly of the produced models (Figure 3.9a) and the cross-section of the sectioned part (Figure 3.9b) for analysis. The extension of the distribution gate in the tree assembly is justified in order to CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 53 prevent turbulent flow within the cavity. Furthermore, this extension promotes the retention of inclusions and slag that may occur as a result of the interaction of the liquid magnesium with the surrounding atmosphere during the melting phase [97]. Figure 3.9: Models geometry (dimensions in mm): (a) tree assembly and (b) analyzed sectioned part. Six castings were performed, consisting of twelve specimens for analysis, and using two different conditions (samples with models' coating and samples without coating). So, half of the samples were coated with three Yttria layers (comprising a thickness of approximately 0.3 mm each) in order to analyze the interaction between the metal and the plaster during the casting process. The PLA-assembled models were immersed in liquid Y2O3 for 10 seconds and then dried in a cubic chamber (800 x 400 x 400 mm3) at a controlled temperature of 30 °C, accelerating the cooling process and removing moisture. This procedure was repeated twice in a one-hour interval, resulting in three layers of refractory protection for each assembled model. As described in the previous chapter, the ceramic moldings for both models with and without coating were made from plaster. A mixture of water and investment powder (Omega + from Gold Star) was used in a ratio of 40/100 (mass%) following the conventional molding technique and following the instructions of the plaster manufacturer (Ultra-Vest, Ranson & Randolph, Maumee, OH, USA). A thermal cycle was then used to remove the water content and polymeric materials, giving consistency to the plaster [220]. It involved first heating of the moldings at 300 °C 60 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 3.14: Schematic mold-metal interface: (a) no use of models coating and (b) use of Y2O3 to coat the models: 1before pouring; 2after pouring. On the other hand, a well-defined surface free of large oxidations was denoted in the obtained samples using Y2O3 to coat the models, as displayed in Figure 3.16a. Microstructural analysis shows a sane matrix, free of large oxidations and reaction products, which indicates a more uniform distribution of the α-Mg and β-Mg17Al12 phases. The EDS analysis of the point Z3 (Figure 3.16b) suggests the appearance of an Mn-Al intermetallic into the matrix, given the Mg content of 34.40 wt.% detected from the surrounding matrix and the very fine size of the particles. The absence of silica in this analysis confirms that an apparent inhibition of reactions was achieved by preventing direct contact between magnesium and plaster. Furthermore, the absence of Yttrium in the analysis indicates no dilution of the Yttria layers into the castings. The presence of a small percentage of oxygen indicates that a slight degree of oxidation has occurred in the first reaction stage, during the melting process. A possible explanation for this is the presence of oxygen in the chamber atmosphere due to the handling procedures since it is required to initiate Mg reactions. CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 61 Figure 3.15: SEM analysis of the samples obtained without coating: (a) backscattered FESEM images in the periphery; (b) EDS analysis of point Z2 and (c) EDS analysis of point Z1. Figure 3.16: (a) backscattered FESEM images of the samples obtained using Y2O3 to coat the models and (b) EDS analysis of point Z3. 62 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Based on the analysis performed in this case study, Y2O3 layer application in the coating of the model is an effective inhibitor of mold-metal interaction in investment casting, preventing serious reactions triggered by this interaction. Therefore, further research could have a positive impact on the magnesium alloy casting industry. 3.5. Summary and Conclusions This chapter addresses the initial steps regarding the methodology proposed in Figure 2.6 (1-4 steps). Considering the conditions and objectives outlined, optimization approaches were explored for each stage of the conventional process. Concerning the initial three-dimensional models, used to create the molding cavities into which the magnesium is poured, it was concluded that the SLA technology offers the best printing results, in terms of detail precision and surface roughness. Furthermore, additive manufacturing of models allows for a significant reduction in time and costs compared to conventional manufacturing processes. Aside from being capable of excellent geometric replication, the resin used in this process had also better wettability values than PLA, which indicates better adhesion to the coating. The three-dimensional models were coated in several different ways in order to study the mitigation of mold-metal reactions. The use of Alumina and Sand-based slip as refractories led to failed castings as a result of strong exothermic reactions. On the other hand, the use of Yttria proved to be very effective in preventing these phenomena. However, the stability of the Y2O3 layers is compromised at elevated temperatures, resulting in cracking of the outer layers. Therefore, Yttria coatings must be applied by dipping with a few layers and a compromise between the coating's permeability and stability is needed. A case study was performed to investigate the effectiveness of Yttria coating. Based on the experiments and analyzes carried out, the following conclusions can be drawn: • Using Yttria (Y2O3) as a refractory coating on the initial models effectively prevented the mold-metal interaction that occurred in the investment casting of AZ91D-1 wt.% CaO magnesium alloy. The high precision casting process and the dimensional complexity of the samples, however, require further improvement of the coating application technique. CHAPTER 3. MODEL-MOLD-METAL INTERFACE CHARACTERIZATION 63 • Strong reaction products were detected in samples in which Y2O3 was not used to coat the models. An extensive oxidation layer was detected along with the mold-metal interface. The formation of specific particles along the entire matrix (new phase) can be attributed to the direct contact between the liquid metal and the plaster molding. Significant porosity defects were also originated due to the gas released by the reactions that occurred during the casting process. • A cleaner surface and a more uniform microstructure were evidenced in the experiments performed in which the models were coated with Y2O3. These experiments revealed no significant oxidation. The absence of Si and Y in the EDS analysis indicates no interaction between the liquid metal and the plaster from the molding, and therefore no dilution of the coating into the castings. • The detection of low oxygen content in the EDS analysis of samples in which Y2O3 coating was applied indicates slight oxidation during the melting process, possibly attributable to residual oxygen in the chamber atmosphere. However, continuous improvement of these procedures can be a promising way to invest in casting as a near-net-shape process to perform magnesium castings. 64 Foundry is believed to be the industry genesis. It is perhaps the oldest known process, as there are records of objects in molten copper made more than 6000 thousand years ago. Chapter 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING A successful investment casting part requires identifying the most influential processing parameters. Variable values of these parameters should satisfy the desired final properties, making them relevant to the quality of the final product. A successful casting process cannot be achieved without knowing the influence of the casting parameters and variables related to the temperatures and pressures involved in the process. These variables assume even greater importance in the investment casting of magnesium alloys, as they are mainly responsible for the thermodynamic conditions that can trigger the occurrence of reactions. On the other hand, the use of a protective atmosphere and the timing handling of the procedures during the process are other essential characteristics to be considered. This chapter examines and discusses the melting and pouring phases that are critical for magnesium investment casting. A thermal cycle optimization is also applied to the plaster moldings. This corresponds to steps 4 and 5 in the methodology proposed in this study (Figure 2.6). In addition, another case study is conducted to analyze and validate the performed optimizations to determine the importance of each casting variable in the fluidity and the filling length of the AZ91D-1 wt.% CaO magnesium alloy. CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 65 4.1. Molding Thermal Cycle Optimization The investment casting process employs a thermal cycle after the manufacture of the plaster moldings to eliminate the water content and ensure consistency of the plaster, as described in case study 1 of Chapter 3. During the process, the polymeric material (fabricated by Additive Manufacturing) melts and is eliminated from the interior of the moldings, resulting in empty cavities of the same geometry that can be filled with metal. However, it was determined that the Yttria layers' stability is compromised at high temperatures, especially when using a thermal treatment that reaches temperatures above 700 ºC, as applied before. In order to determine the optimal temperature conditions, DTG (Derivative Thermo Gravimetry) and TGA (Thermal Gravimetric Analysis) were performed on the printing and molding materials to obtain a better understanding of the thermal impact. According to the results of the graphs presented in Figure 4.1, it is possible to conclude that the printed materials, resin, and PLA, started to evaporate only at temperatures around 400 ºC, as shown in Figures 4.1a and 4.1b, respectively. Based on the weight loss percentage and the DTG analysis, it appears that these materials were almost completely burned. Further, the plaster analysis (Figure 4.1c) suggests that the most significant loss in weight occurs until 450 ºC, due to the water's elimination, and is only 0.3 % up to that temperature. This thermal study made it possible to determine the maximum temperature that could be applied to the plaster moldings without impairing the stability of the Yttria coating. Therefore, an optimized thermal cycle was adopted as part of the proposed methodology for curing the plaster moldings, as shown in Figure 4.2. The moldings were placed into the furnace for an initial isothermal period of 30 ºC. To avoid large thermal gradients that would compromise the stability of the plaster, the heating rate for the following isothermal slopes was lowered. In light of previous analyzes, the maximum temperature of this new cycle was set at 420 ºC. Thus, after 13 h, the elimination of the polymeric materials and the plaster water content was guaranteed, safeguarding the Yttria coating’s stability. The thermal cycle ended at the temperature that corresponded to the temperature of the molding during the casting process, which is normally between 350 ºC and 400 ºC. 66 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 4.1: Weight loss and DTG: (a) resin; (b) PLA; and (c) plaster. Figure 4.2: Thermal processing for curing the plaster moldings. CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 67 An experimental trial was carried out to better understand the influence of the thermal cycle optimization applied to the plaster moldings on the Yttria coating stability. For this, simplified geometry samples were used comprising different section thicknesses and coated with varying numbers of layers of Y2O3. The moldings were cut according to Figure 4.3 after applying the thermal cycle for visual analysis. As shown in Figure 3.8, Yttria presents a significant degree of thermal instability when 720 ºC is reached in the normal cycle, worsening as more layers are added. On the other hand, there was a significant difference in results when using the optimized thermal cycle. Despite the presence of some cracking in the Y2O3 layers, particularly when multiple layers are applied, the coating stability in this case improved. Using the optimized thermal cycle, where the maximum temperature reached is around 420 °C, Yttria adheres better to the molding walls and no parts are detaching from the outer coating layers into the cavities. Comparing both cases, different findings were also observed regarding coating thickness. By the use of a general thermal cycle, higher Yttria instability was observed for thicker coating layers, regardless of the number of layers applied. In contrast, when applying the optimized thermal cycle, it was not possible to establish a pattern of influence between the cavity thickness and the stability of the Y203 coating. Figure 4.3: Application of both normal and optimized thermal cycles in plaster molding. In this regard, and because the thermal cycle optimization applied to the plaster moldings arose as a result of the thermal instability of Yttria layers as a refractory between the liquid magnesium and the plaster from the molding walls, the optimized thermal cycle (Figure 4.2) can be included in the proposed methodology of this work when the use of Y2O3 coating is necessary. 68 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS 4.2. Casting Parameters After thermal treatment, the plaster moldings are put inside the furnace at the adequate pouring temperature. Additionally, a graphite crucible is placed and can contain up to 12 grams of the alloy to be used. Figure 4.4 shows the experimental setup of the induction furnace used in this study (already illustrated for case study 1 in Figure 3.10), with its main characteristics shown in Table 4.1. Figure 4.4: Induction furnace equipped with 112 g of AZ91D-1 wt.% CaO, 2crucible, 3thermocouple, 4induction coil, 5plaster molding, 6argon/vacuum atmosphere, and 7external structure. The magnesium alloy should be placed into the crucible in a single part in order to minimize the contact surface area of magnesium that can be susceptible to reactions. A melting charge is heated by induction in the used furnace (MC15+ from Indutherm). Inductive heating has the advantage of rapidly reaching melting temperatures due to direct heat transfer to the metal through a crucible with a graphite insert. The metal is thoroughly mixed by a magnetic field during the melting phase, ensuring homogeneous mixing. The magnesium alloy used in this study is AZ91D– 1 wt.% CaO, as its composition is shown in Table 3.4. It has already been mentioned in Chapter 2 that the CaO addition turns this alloy into an Eco-Mg alloy, which is formed to avoid the use of SF6 as a protective atmosphere. Thus, only argon is used as an inert element, overpressurized at 2 bar into the furnace when the metal is poured into the molding cavities to help the filling of the cavities. CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 69 Thereafter, and following a short period of metal solidification and cooling, the plaster molding is lowered into the water for demolding, and the resulting magnesium part is cleaned. Table 4.1: Technical data of the experimental setup (according to the manual). Parameters Range Crucible volume 10 cm3 (with graphite inlay) Crucible temperature max. 2000 ºC Melting performance 3.5 kW Main supply 230 V/ 16 A, 50 or 60 Hz Cooling water supply 2.5 – 5 bar/min. 1.5 liter/minute Cooling water recoil Pressureless Cooling water entry temperature 15 – 25 ºC/59 – 77 ºF Ambient temperature 10 – 35 ºC/50 – 95 ºF Relative atmospheric humidity 20 – 80 % Weight ca. 27 kg Dimensions (width x depth x height) 400 x 400 x 450 mm In subchapter 2.3, it is mentioned that the fluidity of liquid magnesium is greatly influenced by its pressure and temperature properties. These properties must be monitored to ensure successful magnesium castings, particularly for thin-walled applications. Several reference values were experimentally optimized in this study, as presented in Table 4.2, ensuring efficient and safe magnesium castings. However, it is important to note that the conditions are always dependent on the intended objectives for each casting, mainly pouring and molding temperatures. According to the literature review, pouring temperature has a greater effect on liquid metal fluidity and reaction propensity than molding temperature. Based on the defined safety values for both temperatures, there must be a compromise depending on the desired final magnesium part. Increasing the temperature results in greater fluidity and consequently greater cavity filling, but it also promotes the reaction of the liquid metals both with the surrounding atmosphere and with the molding walls. This will be discussed further in Chapter 5 for the manufacturing of stents in order to provide a better understanding of the effect of thermodynamic variables, namely pouring and molding temperatures. 76 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS 𝑅(𝜃)=𝑅0+ 𝑝 2𝜋 ∙ 𝜃 Eq. 4.3 in which 𝑅0 corresponds to the initial radius of the spiral center, set to 4 mm, and 𝑝 represents the spiral pitch, set to 5 mm. Thus, the filling length can be calculated through the integration of the arc length of the spiral path with respect to θ as follows: 𝐿(𝜃)=∫√𝑑𝑥2 𝑑𝜃 +𝑑𝑦2 𝑑𝜃 𝜃 0 𝑑𝜃 Eq. 4.4 As integration of Eqs. 4.1-4.3 into Eq. 4.4, and solving the corresponding integral, the final filling length expression is given by: 𝐿= 𝑝 2𝜋ln(𝑝 2𝜋√(𝑝 2𝜋∙𝜃+𝑅0)2+(𝑝 2𝜋)2+𝑝 2𝜋(𝑝 2𝜋𝜃+𝑅0)) 2+(𝑝 2𝜋𝜃+𝑅0)√(𝑝 2𝜋𝜃+𝑅0)2 (𝑝 2𝜋)2+1 2 Eq. 4.5 The initial models were made using additive manufacturing following the proposed methodology, based on FDM technology for each side length. After printing, only the models related to the Y and YV conditions were coated with two Yttria layers (according to the optimized procedures in Chapter 3), followed by the plaster molding fabrication. The Y2O3 coating application was reduced to two layers in order to reduce the cracking content. Since Yttria was used to coat most of the models, the optimized thermal cycle (Figure 4.2) was applied for plaster molding treatment. Regarding the casting parameters described in Table 4.2, the pouring temperature varies at the safe extremes, as indicated in Table 4.3. The same experimental setup, illustrated in Figure 4.4, was employed for melting 10 g of AZ91D-1 wt.% CaO (Table 3.4) in a SiC crucible for each casting. A total of 72 spiral castings were poured. For reproducibility purposes, tree castings were performed for each condition. After pouring and demolding procedures, the measurement of the spiral angle was made. 4.3.2. Results and Discussion Figure 4.9 displays the samples obtained for all performed castings according to variables: casting condition, pouring temperature, and edge length of the squared cross-section. For each casting, the spiral angle, θ, was measured using image analysis software ( ImageJ ), as described step by step in Figure 4.10. Considering a maximum angle of 952.5º for a total spiral length of 177 mm, which corresponds to 100 % filling, the final calculated results are presented in Table 4.4. The 0.5N660 sample, for instance, corresponds to the casting in which the model has a side CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 77 length of 0.5 mm, poured at 660 ºC without the use of a model's coating and vacuum during the melting process (condition N). The process’s reliability and the achievement of successful castings depend a lot on strict compliance with all the involved variables to reduce the effects of the user's handling sensitivity. Figure 4.9: Experimental casting samples resulting from fluidity studies. Figure 4.10: Steps for spirals’ angle measurement. 78 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Table 4.4: Calculated values from the experimental fluidity study. Sample Measured angle - Trial 1 (°) Measured angle - Trial 2 (°) Measured angle - Trial 3 (°) Measured angle - average (°) Filling length (mm) Filling length percentage (%) 0.5N660 281.7 277.1 291.5 283.4 29.8 16.8 0.5N720 292.5 293.3 302.0 295.9 31.6 17.8 0.5Y660 277.3 269.2 293.2 279.9 29.3 16.6 0.5Y720 337.2 332.1 311.1 326.8 36.1 20.4 0.5YV660 403.5 417.2 391.9 404.2 48.4 27.3 0.5YV720 526.4 489.4 509.7 508.5 67.2 38.0 0.75N660 354.0 346.7 362.4 354.4 40.3 22.8 0.75N720 660.7 658.9 649.2 656.3 98.5 55.6 0.75Y660 184.0 169.1 183.6 178.9 16.6 9.4 0.75Y720 531.0 502.5 497.9 510.5 67.6 38.2 0.75YV660 484.0 483.7 460.3 476.0 61.1 34.5 0.75YV720 528.8 532.5 544.5 535.3 72.5 41.0 1N660 769.4 778.2 752.2 766.6 125.3 70.8 1N720 945.0 952.5 941.2 946.2 175.2 99.0 1Y660 285.5 296.5 320.3 300.8 32.2 18.2 1Y720 558.4 565.7 531.4 551.8 75.9 42.8 1YV660 476.8 457.0 433.1 455.6 57.3 32.4 1YV720 767.1 790.5 778.4 778.7 128.4 72.5 1.5N660 952.5 952.5 952.5 952.5 177.0 100.0 1.5N720 952.5 952.5 952.5 952.5 177.0 100.0 1.5Y660 691.7 683.6 714.3 696.5 107.9 61.0 1.5Y720 577.2 597.1 614.1 596.2 85.1 48.1 1.5YV660 942.7 927.5 916.3 928.8 170.0 96.0 1.5YV720 728.8 726.0 710.3 721.7 114.0 64.4 From the visual assessment of castings (Figure 4.9), some trends were noticeable, confirmed by the calculated values presented in Table 4.4. Overall, and as expected, it was evident that a longer filling length was achieved with an increase in the size of the squared cross-section and the pouring temperature, apart from a few exceptions. The obvious difference between the resulting metal reactions between the different casting conditions was also evident from visual observation, which corroborated the obtained results, discussed earlier in this chapter. The cast samples produced from condition N, in which protective coatings and vacuum atmospheres were not employed, resulted in clean metal parts with good dimensional accuracy and finish surface. When only a protective coating was applied to the initial models (condition Y), although it effectively prevented the liquid magnesium reactions, the dimensional accuracy and finish surface were affected, resulting in a worsened final quality of the castings. As discussed in case study 1 of Chapter 3, this phenomenon is caused by the deterioration of Yttria layers due to thermal instability. CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 79 Although the optimization of the thermal cycle (Figure 4.2) applied to the plaster moldings improves the stability of the Y2O3 layers and permits the mitigation of mold-metal reactions, there is still a slight degree of cracking, visible through the visual assessment of the castings' surface roughness. Moreover, the cracking of the outer layers of Yttria acts as a barrier to the metal's fluidity, thus reducing its filling length. As a final observation, contrary to the previous conditions, the obtained castings revealed the occurrence of reactions in the samples when the vacuum was applied during magnesium melting, even with Yttria as a refractory coating (condition YV). In these samples, the molten metal exhibited a black color in the superficial and middle zones of the spiral, due to the liquid remaining for a long time, leading to a longer reaction time with the plaster molding walls. Even though the Y2O3 coating is essential to the castings' success, it prevents strong exothermic reactions that are destructive to the process. Furthermore, these results suggest that vacuum assistance is an important factor in enhancing liquid magnesium reactions with the surrounding elements. Despite this, there was a higher percentage of cavity filling for shorter side length samples in the castings for condition YV, indicating that the vacuum resource effectively improves cavity filling for thin-walled samples. The values presented in Table 4.4 were treated and graphically illustrated in the following figures for better perception and quantification of the obtained results. According to the increasing percentage of filling length for all the samples in Figure 4.11, it is easy to see that the samples with an edge of 1 mm and 1.5 mm long exhibited the highest percent of cavity filling as compared to the samples with a side length of 0.5 mm and 0.75 mm. Furthermore, only two samples were obtained with complete cavity filling: 1.5N660 and 1.5N720. This fact is particularly noteworthy since it indicates the total effectiveness of the process in preventing reactions without requiring the use of protective coatings and vacuum protection. This is in opposition to current scientific knowledge regarding the effect of vacuum in reducing the Mg reactions. Figure 4.11 also demonstrates that the models with a lower percentage of cavity filling corresponded to the samples for condition Y. This highlights the Yttria obstruction of the magnesium pouring into the molding cavities. With respect to the obtained standard deviations, corresponding to three castings per sample, the registered values suggest no significant variation between samples, ensuring good casting repeatability. Despite that, it is evident that higher standard deviations are achieved in conditions Y and YV, which involve more uncontrollable variables due to the use of Yttria and vacuum. 80 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 4.11: Filling length percentage for all samples obtained. Figure 4.12 presents the results as a function of the edge of the squared cross-section and the casting conditions for the samples. It seems evident that the highest percentage of filling length was obtained in the spirals with the highest thickness and temperature for castings following the condition N. Indeed, this was the only case where the cavity filling followed an increasing orientation with increasing thickness and temperature. The obtained results were distinct in condition Y, and the lowest cavity filling percentages were achieved for almost all samples except for 0.75Y720 and 1Y720. In addition, for this condition, the negative influence of the Y2O3 coating on the flow of liquid magnesium into the cavities was observed, being particularly evident for the 1.5Y720 sample. There are two important pieces of evidence to highlight regarding the condition YV. In the first place, the vacuum effect appeared to significantly impact the cavity filling of thin-walled magnesium geometries, which allowed for an optimal filling rate for samples with a thickness of 0.5 mm at both temperatures studied. Secondly, the obtained value of 64.40 % filling for the 1.5YV720 sample, which is lower than the same condition at 660 ºC, suggests a detrimental effect when the vacuum is used to assist higher thickness and temperature samples. Overall, considering the same casting condition, only three samples (marked in Figure 4.12 by the blue arrow) did not verify the tendency to rise the cavity filling with the increasing square cross-section. This occurred for conditions Y and YV, again related to the influence of Yttria coating and vacuum assistance. 0% 20% 40% 60% 80% 100% Filling length CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 81 Figure 4.12: Filling length comparison between the different sizes of the squared cross-sections. Additional statements can be traced from the obtained results when observing the cavity filling percentage concerning the temperatures used. In Chapter 2, it was mentioned that temperature is an essential factor affecting the fluidity of the liquid metal, inducing greater fluidity at higher temperatures. According to Figure 4.13, this fact was proven for most samples in this case study, although there was one point of inversion of the trend. With larger cross-sectional area spirals, the increase in temperature appears to have a negative effect on the filling length, which is particularly evident for condition YV. It can be explained by the action of the vacuum pressure, which seemed to enhance the mold-metal reactions, blocking the flow and decreasing the fluidity of the liquid magnesium. In these cases, the greater the sample thickness, the longer the moldmetal contact and the greater the rate of reaction. Figure 4.13: Filling length comparison between both temperatures. 0% 20% 40% 60% 80% 100% N-660 N-720 Y-660 Y-720 YV-660 YV-720 Filling length Casting condition 0.5 0.75 11.5 0% 20% 40% 60% 80% 100% 0.5 N 0.5 Y 0.5 YV 0.75 N 0.75 Y 0.75 YV 1 N 1 Y 1 YV 1.5 N 1.5 Y 1.5 YV Filling length Casting condition 660ºC 720ºC 82 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 4.14 illustrates the compilation of results for each casting condition. Here, it is also noticeable that higher filling length is achieved for thicker samples using condition N, and for thinner samples using condition YV. Additionally, the temperature increase only affects the cavity filling when Yttria and vacuum are used for thicker samples. Otherwise, there is a visible increase in cavity filling. As a final note, it must be noted once again the lower filling length percentages obtained in condition Y, which seems to be little affected by temperature variation. Figure 4.14: Filling length comparison between casting conditions: (a) condition N; (b) condition Y, and (c) condition YV. To further analyze the different casting conditions and their influence on the filling length, an µ -CT (X-ray Micro-computed Tomography) was conducted in order to validate the obtained results. The Bruker SkyScan 1275 (Bruker, Kontich, Belgium) used is an automated desktop laboratory system with an X-ray beam with peak energy from 20 kV to 100 kV, 10 W of the maximum power, small X-ray spot size (>55 µ m), along with multiple filter options (aluminum and copper filters). The X-ray detector of this µ -CT scanner is a 3 MP (1944 x 1536 pixels) active pixel CMOS flat-panel [225]. This equipment is composed of an X-ray source, a motor-controlled rotating CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 83 stage, and a detector. The specimen is mounted between the X-ray source and the detector panel, in the rotating stage. The radiation emitted by the X-ray source passes across the specimen and is projected on a digital detector, measuring the attenuation of the X-rays and producing a radiograph (known as 2D scanned or projection image). During acquisition, the object rotates over 180 or 360 with a fixed rotation step defined by the operator. At each angular position, a 2D scanned image is acquired. The sequentially acquired projection images are then processed using computer software, producing a series of reconstructed images (2D slices) that allow observing the object’s internal structure. The reconstructed µ -CT images can be used for volume rendering of tomographic data, creating 3D models using numerous available software tools (for example, CTAn, CTvox, CTVol, VGStudio) [226]. Image optimization comprises adjusting several µ -CT parameters such as the magnification, the incident X-ray intensity, the filter type (e.g., no filter, copper, or aluminum), the rotation step and the acquisition time, and the threshold. Through the 2D acquired projection images of the spirals with 1.5 mm of the squared cross-section regarding all casting conditions, as shown in Figure 4.15, both samples’ good casting quality is notorious for condition N (Figures 4.15a and 4.15b), as mentioned early. Additionally, the appearance of a 'cracking husk' is noticed around the spiral specimens for the condition Y samples (Figures 4.15b and 4.15c) and condition YV samples (Figures 4.15e and 4.15f). It can be explained by the Y2O3 layers cracking, which, when detached from the molding walls, allow the liquid magnesium to penetrate through its interstices. This creates an outer Yttria shell that aggregates to the metal and is not naturally removed during demolding. As already mentioned in this work, this phenomenon can block the metal and affect the cavity filling, which may explain the results obtained in this case study. In addition, it appears that the formation of this Yttria "cracking husk" seems to be more evident in samples poured at 720 ºC due to the greater fluidity of the liquid metal, and the latter samples for condition YV due to the vacuum effect. The 3D models of the Yttria ‘cracking husks’ formed around metallic spirals can be created through a 3D analysis software (Avizo) by µ -CT images reconstruction and observed separately due to the density difference between materials. In Figure 4.16, it is illustrated a higher magnification of the Yttria 3D models for 1.5Y660 and 1.5YV720 samples, demonstrating the existence of cracks in the Y2O3 coating. Comparing both samples, it is observed that the amount of aggregation to the metal is greater for condition YV720, demonstrating a direct correlation between the increase in pouring temperature and the use of vacuum with the greater penetration of the liquid metal into the coating interstices. A further consequence of the degradation of the Y2O3 coating 84 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS layer is the increased diffusion of elements at the mold-metal interface, which results in forming a reaction layer in the peripheral zone of the sample, as shown in Figure 4.17 (analysis SEM 1.5YV720 condition). The longer the cooling time, the greater the diffusion rate of elements and, consequently, the greater the formation of reaction layers. In conjunction with the vacuum effect, this phenomenon is responsible for the higher rate of surface reactions occurring in samples of condition YV. Figure 4.15: 2D µ -CT images of the samples with 1.5 mm of thickness: (a) N660; (b) N720; (c) Y660; (d) Y720; (e) YV660, and (f) YV720. CHAPTER 4. THERMODYNAMIC ANALYSIS IN MAGNESIUM INVESTMENT CASTING 85 Figure 4.16: 3D µ -CT models of the Yttria ‘cracking husks’ for samples: (a) 1.5Y660 and (b) 1.5YV720. Figure 4.17: SEM analysis of the mold-metal interface of sample 1.5YV720: (a) backscattered FESEM image, and EDS elemental mapping of the element (b) Y, (c) Mg, and (d) O. 4.4. Influence of Casting Variables DoE (Design of Experiments) is a relevant methodology applied to quantify effects, understand interactions between variables, model relationships, and measure experimental errors to obtain unambiguous answers to specific research questions, at minimum cost [227]. It has been 92 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 4.19: Evaluation of the filling length comparing all the studied variables. 0% 20% 40% 60% 80% 100% 0,25 0,50 0,75 1,00 1,25 1,50 1,75 Filling length Side length (mm) N-660ºC Y-660ºC YV-660ºC N-720ºC Y-720ºC YV-720ºC 93 No references are found in the literature to obtaining stents by investment casting. Even the available data about using this process to produce thin-walled magnesium components is scant. Chapter 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING As discussed in Chapter 2, the manufacture of stents, by laser cutting, is an expensive and complex process due to the particularities of geometrical shapes and the micro dimensions involved. In addition, this process is not feasible for magnesium alloys because of their high reactivity. Therefore, the objectivity of this doctoral thesis gains greater interest and significance, both in the scientific and industrial context. This chapter discusses and characterizes the procedure for obtaining magnesium stents through investment casting in plaster molding. Based on all the optimizations made in the previous chapters of the various steps that comprise the methodology outlined in Chapter 2 (Figure 2.6), magnesium stents with a minimum thickness of up to 0.4 mm and a complex geometry were cast. Both casting conditions approved in the previous chapter, YV (Yttria and vacuum) and N (No Yttria and no vacuum), were applied and characterized for stent manufacturing. In the conclusion of this chapter, it will be possible to evaluate all the phenomena that occurred and determine what are the optimal casting conditions to guarantee the successful casting of the AZ91D-1 wt.% CaO magnesium alloy in plaster molding to obtain any magnesium part. 94 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS 5.1. Model’s Production: CAD and Printing Following the proposed methodology described in Figure 2.6 (subchapter 2.5), and taking all previous processing optimizations into account, CAD modeling is the first step in stent production. Figure 5.1 illustrates the CAD stent model adopted in this study [232]. Throughout this chapter, and using a particular unit cell (Figure 5.1a) two different dimensions of the model have been used for different studies, as shown and described in Figure 5.1b. The thin-walled magnesium casting is evaluated using Model 1, which has a thickness of 0.4 mm. On the other hand, a detailed analysis of the microstructure has been carried out on Model 2, which has twice the thickness of the walls of Model 1. Figure 5.1: CAD stent geometry (mm): (a) design unit cell, and (b) models for testing. For prototyping the stents' 3D models, SLA printing was used. The respective filling system was printed by FDM printing. Due to specific application requirements, both stent models were produced on the FabPro 1000 printer using the castable resin (JewelCast GRN) already described in Chapter 3. The optimal 3D printing of stents requires the use of support material and an angle of 20º for each stent system, as shown in Figure 5.2a. This printing optimized process results in an overall printing time of 2 hours and 4 minutes for Model 1 stent printing (Figure 5.2b). CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 95 Figure 5.2: Additive manufacturing of stents Model 1: (a) construction of support material, and (b) printing maximization. Regarding the stent design, the filling system was specifically investigated. As the investment casting process is widely used for large productions of small pieces, as is often the case in jewellery manufacturing, it is essential to maximize the filling tree of the initial models to ensure good metal flow and achieve the maximum number of parts for every casting. To define the best filling method and determine the possibility of maximizing the stent's production, Table 5.1 shows the different approaches tested in this experimental research on the filling tree and the characteristics of each casting and the obtained results. The experimental setup and casting parameters of the previous case study 2 in Chapter 4 were applied to this process (Table 4.2). In addition to the use of Yttria to coat the samples and the vacuum assistance during melting, the pouring temperature of 720 °C was determined to increase magnesium fluidity into the molding cavities of the stent Model 1. 96 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Table 5.1: Filling design optimization using condition YV and 720 ºC pouring temperature. Tree assembly design Production Type flow Casting result 1 stent Direct filling Complete filling length. Medium surface quality. Occurrence of reaction gradient. 1 stent Indirect filling Incomplete filling length. Poor surface quality. Metal stent without reaction. 2 stents Direct filling Complete filling length with some defects. Poor surface quality. Occurrence of reaction gradient. 2 stents Indirect filling Complete filling length with some defects. Medium surface quality. Metal stents without reaction. 5 mm CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 97 The results of the performed castings allow to conclude that no ideal condition existed for the design of the filling tree. As shown in Table 5.1, relating to the Model 1 stent castings, a more turbulent metal flow (direct filling) enhances the rate of reactions even when the cavity has been completely filled in both tree designs (1 and 2 stents). These castings revealed a black color gradient around the metal that was more evident in the zones closest to the filling basin corresponding to the zones in which the metal solidification rate was lower and, consequently, the metal stayed in contact with the molding walls longer. This phenomenon is consistent with the negative impact of vacuum highlighted in case study 2 of the previous chapter. In contrast, a uniform flow, promises greater magnesium purity from the bottom to the top (indirect filling), although complete stent cavity filling was not achieved. The cracking of the Yttria coating affected the surface finish of the metal, as expected, although it was less noticeable in the case of indirect filling for the casting of two stents. Although the direct filling of liquid magnesium enhances the occurrence of reactions due to the vacuum influence, it was only in this condition that the cavity of the 0.4 mm thick stents was completely filled. Thus, the use of vacuum proved to be a fundamental condition for the effectiveness of the thin-walled filling length process, in conjunction with the metal flow by direct filling. In order to optimize the process, a thermal analysis was performed on the plaster moldings during the casting process (after the thermal cycle) to better comprehend the influence that the filling tree design has on the casting of the magnesium alloy. As illustrated in Figure 5.3a, three thermocouples were placed at distinct positions in the plaster molding (periphery, middle, and center). The temperatures in each position were then recorded at the defined molding temperature (400 °C) during the cooling phase. The temperature readings in Figure 5.3b indicate that there was a significant temperature loss in the most peripheral zone of the plaster molding during the time of molding cooling. This temperature loss was more significant from periphery to the center. During the cooling time defined in Table 4.2, the peripheral zone decreases its temperature by about 100 ºC to the surrounding atmosphere. This decrease begins almost immediately after the reading of the values. On the other hand, the heat transfer in the molding center was very low, which means that the central zone remained practically at the same temperature until demolding. 98 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 5.3: Molding cooling (after pouring): (a) placement of thermocouples, and (b) recording of temperatures. The results allow to conclude that the solidification of the metal occurs faster in the peripheral areas of the molding, which can explain the cavity filling defects observed during the testing of a tree consisting of two stents when the thin-walled cavity filling is required close to the molding periphery. Therefore, to ensure greater metal fluidity and, consequently, a longer cavity filling length of thin-walled stents, the unitary tree system should be applied to each stent within each molding. Furthermore, a higher metallostatic pressure helps fill the cavity in this condition. Figures 5.4 and 5.5 illustrate the thermal gradient suffered by the plaster molding for both direct (1 stent) and indirect (2 stents) casting models at two distinct times: at the exact moment when magnesium is poured, and 3 minutes afterward. During this last condition, the temperature loss verified in the peripheral zones of the plaster molding was quite visible. This loss can significantly affect the stent cavity filling, especially in the areas farther from the molding center. CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 99 Figure 5.4: Schematic representation of molding temperature distribution during pouring. Figure 5.5: Schematic representation of molding temperature distribution 3 minutes after pouring. 5.2. Coating, Molding, and Casting Procedures After modeling and printing the initial models and respective filling systems (Figure 5.6a), the steps of coating, fabrication of plaster moldings, melting, and casting are discussed below for both stent models (see Figure 5.1b), using both casting conditions defined in Chapter 4 (condition YV and condition N). The casting condition YV was applied to the casting of stents almost like it was in the previous chapter for spirals (case study 2). After assembling the trees (Figure 5.6b), the coating of the models assumes a specific and essential step that should be more careful and homogeneous 100 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS as possible (Figure 5.6c). The stent models were immersed in liquid Y2O3 (99.99 wt.% purity - ZYP coating) and dried at a controlled temperature of 30 ºC in a cubic chamber (800 x 400 x 400 mm3). The optimized painting characteristics for each model, shown in Table 5.2, were determined experimentally. Different coating layers and times between coatings were applied to each model as a result of different contact areas. Table 5.2: Optimized painting characteristics. Model Coating layers Immersion time Time between coatings Curing time after coatings 1 3 10 s 30 min 24 h 2 2 10 s 60 min 24 h Figure 5.6: Model 1 processing: (a) additive manufacturing; (b) tree assembly; and (c) Yttria coating. As the protective coating of the initial models is not used for condition N castings, the Yttria coating step is not incorporated into the process methodology. After curing the model in condition YV and the assembly of the tree in condition N, the tree was put in plaster to produce the ceramic moldings according to the procedures outlined in the previous case studies. For that, a mixture of water and investment powder (Omega + from Gold Star) in a 40/100 (mass%) ratio was made according to the same conventional molding technique, and the moldings were dried at 30 ºC before the thermal processing cycle to give consistency to CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 101 the plaster, eliminating the water content. Because of the thermal instability of the Yttria coating, the thermal cycle optimized in Chapter 4, and illustrated in Figure 5.7, was applied to the casting condition YV (see subchapter 4.1), using a molding temperature of 400 °C. The casting process was performed for both casting conditions by applying the same experimental setup as in the previous case studies and using an induction melting furnace assisted by vacuum (condition YV) or without vacuum assistance (condition N), as illustrated in Figure 5.8. 10 g of AZ91D-1 wt.% CaO was cut, dried, and melted in a SiC crucible for each experiment. Table 3.4 shows the composition of the commercial alloy used in this study. As discussed in Chapter 2, this alloy was used to promote a no-SF6 casting, which permits performing all tests under an argonprotected atmosphere and minimizes reactions during the melting stage. For the casting of condition YV, the metal was melted at 700 ºC under a 1.0 bar vacuum atmosphere and kept isothermal inside the crucible for 1 minute for homogenization. In condition N, magnesium was melted at the same temperature in an air atmosphere. After this period, the melt was poured for both casting conditions at the same temperature into the preheated plaster molding cavities (400 ºC) with an overpressure of 2.0 bar, which facilitates the metal pouring. Following a cooling period of 10 minutes, the stents were removed and cleaned after immersing the moldings in water for de-molding. Figure 5.7: Thermal cycle optimized and applied to the plaster moldings when Yttria coating is used in stent manufacturing. 0 50 100 150 200 250 300 350 400 450 0 1 4 7 10 13 16 19 Temperature [°C] Time[h] 108 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Figure 5.15: Backscattered FESEM images of the Model 2 coated unit cell. Table 5.4: EDS analysis of the points indicated in Figure 5.15. EDS analysis confirmed the findings already described for coated and uncoated models. In addition to the well-defined α-Mg (‘a’ and ‘i’ points) and β-Mg17Al12 (‘b’ and ‘j-k’ points), the presence of an Al-Mn phase was found for both models, corresponding to the point ‘e’ on the uncoated model (Figure 5.14) and the point ‘l’ on the coated model (Figure 5.15). The presence of elements that do notcompose the studied alloy, such as Si and S, provides evidence that reactions occur in the uncoated model involving the plaster from the ceramic molding. In contrast, these elements are virtually undetectable in the coated model. As a result of the interaction between the metal and the plaster, the observation of 19.4 % Si at point 'c' suggests the formation of the Mg2Si phase in the uncoated model, which is expected. Element/ Point Mg Al Mn Ca Y Zn O Si S Weight % i 90.9 6.0 0.5 2.6 j 63.6 26.9 2.6 3.3 3.6 k 63.1 23.9 8.5 0.2 4.3 l 39.6 28.7 26.9 0.5 0.0 3.8 0.4 m 26.2 31.0 0.8 1.5 32.8 7.7 n 90.0 6.8 0.5 0.6 2.1 o 89.3 7.1 0.7 0.8 2.1 p 91.9 4.6 3.5 CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 109 An intense degree of reactions occurred in the uncoated model periphery, as evidenced by the 'd' point values of 38.5 % O, 17.5 % S, and 4.3 % Si. Point ‘k’ suggests a shape change from strips to networks of the intermetallic Mg17Al12 due to the greater presence of Ca, as evidenced by the results of the EDS analysis and in accordance with research work reported Wang et al. [55]. This mechanism suggests contributing to the refinement of the intermetallic phase, being more noticeable in the coated model. Additionally, the EDS analysis performed on coated model cast samples revealed that a Y trace was present within the alloy's microstructure ('m' point), near its periphery. As previously reported, this is directly related to the coating cohesion on the walls of plaster molding. If the same printing model is immersed in Yttria many times to thicken the protective layer, as was done with Model 2, small flaking may occur in the innermost layer of the molding walls due to the thermal cycle process. Due to this phenomenon, small Y2O3 particles detach from the molding wall, migrating towards the liquid metal close to the periphery due to their very high stability. A possible cause of the oxidation rate that can occur during the entire process is the reaction between the liquid metal and the surrounding environment, as well as the contact between the liquid metal and the molding. Despite the controversy surrounding the quantification of oxygen (O) via EDS, in this study, the oxygen percentage values served only as a reference for the comparison of the different casting conditions, allowing to understand the relationship between oxidation rate and oxygen values. The higher oxygen values recorded in the uncoated model suggest that this kind of cast sample has a higher level of oxidation. More specifically, higher oxygen values were recorded in the ‘f-g-h’ points (uncoated model – Figure 5.14), varying from 2.3 % to 6.7 %, compared with ‘m-n-o’ points (coated model – Figure 5.15), which varied from 2.1 % to 3.5 %. The increase in oxygen in the uncoated microstructure is directly related to the decrease in Mg at the same points. In other words, and according to [234], the reaction between Mg and O seems to be a two-way diffusion reaction and thus, MgxOy phases can be formed. The O percentages were also measured in other samples to guarantee the casting’s reproducibility, both in the innermost zone and in the peripheral zone of the microstructure, as shown in Figure 5.16. The recorded values demonstrate and confirm the tendency for higher oxidation in uncoated models, mainly in peripheral zones. As can be seen in Figure 5.17, these facts were confirmed by the hardness analysis performed on both samples. On the coated sample periphery, higher hardness values were measured, justified by the rapid solidification of this zone. Conversely, no 110 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS significant differences were observed in hardness values between coated and uncoated samples in the center, indicating that the coating affects the mechanical properties of the metal. Figure 5.16: Values of oxygen concentrations in all castings. Figure 5.17: Hardness analysis of the Model 2 unit cell. An XRD analysis was also performed on both cast samples of Model 2 (Figure 5.18). The obtained results were consistent with the previous analyzes, which showed that reaction products were generated only in the uncoated models at the mold-metal interface, such as MgO, SiO2, and Mg2Si. In the coated model, besides α-Mg and β-Mg17Al12 phases, Y element was also registered, which proves the presence of Yttria traces particles in the microstructure. The CaO compound is a constituent present on the alloy and was only detected in the coated sample, which means it was decomposed and participated in the reactions that occurred in the uncoated sample. 20 30 40 50 0100 200 300 400 Hardness [HV] Distance to the center [ µ m] Coated Uncoated 0 2 4 6 8 10 Sample 1 Sample 2 Sample 3 O [%] Coated center periphery 0 2 4 6 8 10 Sample 1 Sample 2 Sample 3 O [%] Uncoated center periphery CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 111 Figure 5.18: XRD analysis of the Model 2 unit cell. 5.3.2. No Yttria and no Vacuum The castings for condition N were made without the vacuum atmosphere for melting and without the Yttria for coating the initial 3D models, albeit the same experimental setup was employed, as illustrated in Figure 5.8. The processing methodology for this condition was therefore limited to optimizations carried out along the various steps that can be applied to these castings, such as the printing of models and the optimization of casting variables, such as pouring and molding temperatures. In the case of study 2 reported in Chapter 4, concerning the AZ91D-1 wt.% CaO magnesium alloy fluidity, no reactions between metal and the surrounding elements were observed for condition N. Therefore, an experimental study was conducted to investigate the effects of temperature on the process. For this purpose, four castings were performed at pouring temperatures of 600 ºC and 800 ºC and molding temperatures of 200 ºC and 600 ºC in order to obtain Model 1 stents. Figure 5.19 shows the resulting castings. Therefore, it can be concluded that temperature was a significant variable for this condition, both regarding cavity filling and the occurrence of reactions. The complete stent cavity filling was not achieved for any casting, which was almost non-existent at low casting temperatures for both molding temperatures; however, increasing the latter can slightly fill thin-walled ribs. In contrast, the casting temperature of 800 ºC was critical, and there was a severe occurrence of magnesium reactions aggravated by an increase in the molding temperature. The 0 100 200 300 400 500 600 700 800 900 30 32 34 36 38 40 42 44 46 48 50 Counts 2θ Mg Mg Mg17Al12 Y2O3 CaO Mg Mg17Al12 Mg17Al12 Mg Mg Mg Mg17Al12 Mg MgO Mg17Al12 Mg2Si SiO2 MgO Mg 112 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS occurrence of this phenomenon allowed to conclude that the existence of magnesium reactions in condition YV was not caused directly by the use of a vacuum atmosphere. This suggests that the vacuum atmosphere enhances the reactivity of magnesium and even its vaporization. Furthermore, mold-metal reactions also appear to be influenced by the behavior of the liquid metal at high temperatures. The effect of vacuum to assist the pouring enhances the mold-metal contact and, consequently, reactions between liquid magnesium and the plaster. Thus, since vacuum assistance is not used, a more extensive range of temperatures can ensure safe magnesium casting in plaster molding under condition N. Figure 5.19: Study of the influence of the molding and pouring temperatures for condition N. As Yttria was not used as a protective coating, the optimal thermal cycle for plaster moldings was not applied to this casting condition (condition N). Therefore, since a vacuum reduces the magnesium fluidity, this decrease can be compensated for by a small increase in the temperature involved. For that purpose, it was necessary to know the limit temperatures for the occurrence of reactions, thereby allowing to define the optimal casting conditions for condition N. CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 113 Applying the molding limit temperature of 450 °C, another experimental study was carried out to determine the optimal casting temperature for the casting of thin-walled stents through the condition N. The analysis of stent castings for different temperatures, shown in Figure 5.20, led to the conclusion that the mitigation of reactions was only possible for pouring temperature less than 720 °C. At temperatures higher than 720 ºC, the necessary thermodynamic conditions for the occurrence of magnesium reactions appeared to be met, just as in condition YV at lower temperatures. Moreover, the complete cavity filling could only be achieved at temperatures higher than 720 °C, although it seems to be the limit temperature that inhibits reactions. The casting parameters should therefore be defined based on the characteristics of the final product considering the linear relationship between the filling length and the occurrence of reactions. In the case of Model 1, a complete casting without reactions was achieved by defining molding temperature at 450 °C and pouring at 720 °C. These results are consistent with those obtained in case study 2 (Chapter 4), in which melts occur in condition N without any magnesium reactions occurring at a pouring temperature of 720 °C. Figure 5.20: Influence of pouring temperature for condition N. By applying the same processing method and all the optimizations performed during this work, both stent models were successfully obtained for condition N, as shown in Figure 5.21. Consistent with the results from the previous chapter, and contrary to what was expected for these casting conditions, the stents demonstrated good geometric rigor, and good surface finish, as well as high metal sanity. In addition to the complete filling length, the dimensional and geometric accuracy were quite faithful to those of the initial models, revealing only some imperfections due 114 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS to the associated geometric complexity. It was also reflected in the early stages of 3D printing and was agreed upon by all analyzed conditions. A compromise must be struck between the used temperatures to ensure reaction-free castings and sufficient fluidity for complete stent cavity filling. In the case of stent Model 2, good results were achieved using a molding temperature of 400 ºC and a pouring temperature of 700 ºC, which were a little lower than those for stent Model 1. This can be explained by the larger dimensions of stent Model 2. Figure 5.21: Stent obtained for both models through casting condition N: (a) printed Model 1; (b) printed Model 2, and (c) cast stents for both models. The most surprising aspect of casting condition N was the absence of magnesium reactions throughout the entire process, despite the absence of a vacuum atmosphere that greatly facilitates cavity filling, especially in cavities of 0.4 mm thickness as in the case of stent Model 1. During the melting and pouring phases (with argon at 2 bar overpressure), liquid magnesium appeared to remain chemically stable, with no reactions occurring that could affect the alloy composition and the metal casting. The magnesium did not show significant reactions during the CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 115 cavity filling and solidification of the metal within the plaster molding, suggesting that direct moldmetal contact is not detrimental to the casting and sanity of the molten metal. Likewise, condition N was tested for an indirect pouring metal design in order to obtain two stents, which had been more successful in condition YV, and both conditions were compared (Figure 5.22). According to the obtained castings, conditions N allowed achieving better metal sanity provided by reactions, although condition YV did not permit achieving a complete stent cavity filling. Therefore, although vacuum enhanced the fluidity of the magnesium, thereby promoting greater cavity filling, it also enhanced reactions that affected the sanity of the melt surface. Figure 5.22: Comparison of conditions YV and N for indirect filling stent casting. The microstructural analysis of the stent metal for condition N is presented in Figure 5.23 and Table 5.5. Observing the microstructures for both indicated areas, a good surface finish and higher dimensional rigor were notorious, showing a clear definition of the predominant matrix (αMg) and intermetallic (β-Mg17Al12) phases. The rapid metal solidification in these castings explains the small growth of the intermetallic phase, although it was evenly distributed throughout the microstructure. The EDS analysis provided additional support to the obtained results. In one sense, the absence of the elements Si and S in the samples appeared to indicate the absence of severe reaction products in the castings, as shown in the samples of condition YV presented in the previous subchapter. On the other hand, in addition to the known matrix (α-Mg) and intermetallic 116 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS (β-Mg17Al12) phases, it was observed the formation only of phases involving Al and other predominant elements from the studied alloy, forming for instance Al4Mn and AL2Ca phases. Figure 5.23: SEM analysis of the Model 1 unit cell for condition N. Table 5.5: EDS analysis of the points indicated in Figure 5.23. Element/ Point Mg Al Mn Ca Zn O Weight % a 89.9 6.2 0.8 3.1 b 43.4 36.7 15.7 0.8 3.4 c 41.5 37.4 12.4 4.7 4.0 d 87.8 8.9 0.9 2.3 e 36.0 43.1 16.9 0.6 3.5 f 41.0 26.9 23.1 1.3 0.2 7.4 CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 117 5.4. Processing Methodology Overview Experimental casting studies have been carried out during this study to optimize the proposed methodology for thin-walled magnesium investment casting, taking into account all the defined steps (see subchapter 2.5). Based on the intended purpose and the final casting characteristics, the investment casting process of AZ91D-1 wt.% CaO magnesium alloy in plaster molding covered a variety of approaches. Specific to the manufacture of stents, different process conditions were investigated and optimized in order to obtain castings with high metal sanity. Figure 5.24 illustrates the Model 1 stent casting fabricated under the conditions YV, Y, and N, and using the optimal casting variables for each one. In comparing the cast stents, and although both conditions YV and Y resulted in successful thin-walled cavity filling, the casting following condition N produced a 0.4 mm magnesium stent that presented a better surface finish and metallic sanity, which followed the discussion line of the previous subchapters. Using vacuum for condition YV produced a slight level of superficial magnesium reactions in the filling system zone, whereas Y2O3 layers slightly affected the dimensional accuracy of the molding cavities in condition Y. Figure 5.24: Comparison of Model 1 stent casting applying different casting conditions. For a better understanding of the cavity filling quality when condition N was applied, an Xray µCT analysis was performed for both cast stent models, as shown in Figure 5.25. The analysis was performed according to the software used in case study 2 of Chapter 4 (Avizo) and using the 124 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS Considering the (ii) magnesium alloy used, Figure 5.30 represents an experimental attempt to apply the same processing methodology to the commercial AZ91D magnesium alloy. The casting was performed under the casting condition N (no vacuum and no Yttria coating), which inhibited the occurrence of reactions in the handling of the AZ91D-1 wt.% CaO magnesium alloy during this work. This casting, however, was seriously compromised by severe reactions and high vaporization of magnesium, both during the melting phase (Figure 5.30a) as well as after magnesium pouring, as the result of reactions at the mold-metal interface. In this case, no protective oxide surface layer was formed. Instead, magnesium was continuously oxidized with elements from the furnace atmosphere, later leading to a reactive casting that resulted in oxide nodules with cauliflower structure (Figure 5.30b), as reported by Tan et al. [236]. As a result of this evidence, it can be concluded that the magnesium alloy used is also an important factor for the process's effectiveness. Figure 5.30: Stent investment casting using the commercial AZ91D magnesium alloy for condition YV: (a) during melting, and (b) after pouring. 5.5. Summary and Conclusions This chapter applied the optimized methodology for the investment casting of AZ91D-1 wt.% CaO magnesium alloy in plaster molding to obtain stents, applying the casting conditions defined in the previous chapter. Two stent models with rib thicknesses of 0.4 mm and 0.8 mm CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 125 were successfully cast, demonstrating complete cavity filling, good surface finish, and high dimensional and geometrical accuracy. In order to characterize the surface and microstructure of the castings, OM and SEM were used to perform EDS and XRD analyzes. Different behaviors were reported by comparing different casting conditions, and different explanations were presented to justify the observed phenomena and the obtained results. Based on the experiments and analyzes carried out, the following conclusions can be drawn: • The optimization of the filling design in investment casting using plaster moldings is detrimental to the manufacture of stents. The analysis of the molding cooling behavior showed significant temperature losses at the periphery a few minutes following the thermal cycle, which negatively influenced the flow of the liquid metal in the molding cavities in these areas, affecting especially the filling length of thin-walled parts. Thus, a single and direct filling method results in a more successful stent casting. • The use of vacuum assistance to improve the fluidity of the alloy, in combination with the application of the casting condition YV, led to the casting of stents with a higher rate of cavity filling and an almost complete absence of reactions during the process, as well as high geometric rigor compared to the initial 3D models fabricated by additive manufacturing. Coated samples with Y2O3 layers revealed better metal purity in addition to a more uniform and refined microstructure. In the case of uncoated sample castings, several defects were observed affecting fluidity, dimensional and surface quality, as well as microstructure sanity. In Chapter 3, it was shown that severe reactions and oxidation occurred at the mold-metal interface in which no Yttria coating was applied (condition N). This resulted in the formation of reaction products that contaminated the microstructure of the magnesium alloy. According to these results, Y2O3 coatings should always be used in vacuum-environment castings. • It should be noted that the registered values for oxygen concentrations (O) were only a complementary part of the analysis. Nevertheless, it shows a more pronounced degree of oxidation in the peripheral zones for both models analyzed, especially in those models in which no coating was applied. • The casting of stents can be successfully achieved in any of the three conditions tested (Y, YV, and N). However, condition N guarantees the highest quality castings for both stent models. According to the stent model, successful Mg casting is dependent on the balance 126 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS between pouring and molding temperatures, which can vary according to defined safe values, as shown in Figure 5.31. • In Chapter 4, it was concluded that condition Y was ineffective in this type of application. The obtained results confirmed that Yttria should only be used for castings in which a vacuum atmosphere is required, such as for the manufacturing of thin-walled magnesium parts. If the samples are thicker, the Y2O3 layers can create resistance to the metal flow and affect the castings’ final surface roughness. Furthermore, the Yttria degradation phenomenon allows the liquid magnesium to penetrate through the coating interstices, creating a "cracking husk" around the metallic sample and increasing the diffusion of elements in the mold-metal interface. As a result, the surface samples react faster, particularly when the vacuum is applied. • Although the proposed methodology provides effective thin-walled magnesium castings, the final samples revealed some defects. According to the analyzes carried out, these imperfections are a consequence of the casting process and the original 3D samples, which were created through additive manufacturing. • The obtained results for conditions N and YV are based on two important factors: (i) the casting parameters implemented and (ii) the magnesium alloy used. Pressure and temperature were the key factors determining the greater or lesser occurrence of magnesium reactions, both during the melting and pouring phases. On the other hand, AZ91D-1 wt.% CaO Mg-Eco alloy has shown greater chemical and thermal stability than the commercial AZ91D magnesium alloy, which is critical to successfully applying the investment casting process in plaster molding. The use of a vacuum atmosphere (condition YV) for the casting of AZ91D-1 wt.% CaO magnesium alloy seems to inhibit the liquid metal from forming and preserving the protective oxide layer of MgO along the surface, which happens in castings under the condition N. For this condition, the oxide layer ensures the chemical stability of the molten metal, increases its surface tension, and prevents further magnesium reactions. CHAPTER 5. PROCESS VALIDATION: MAGNESIUM STENTS MANUFACTURING 127 Figure 5.31: Optimal range of temperatures for stent manufacturing under the casting condition N. 128 The present thesis reflects extensive experimental work. In total, 4.5 kilograms of magnesium alloy were consumed. Since it was indicated that 10 grams of magnesium were required for each thin-walled casting, it means performing 450 castings! Chapter 6. CONCLUDING REMARKS 6.1. Final Conclusions An innovative methodology for manufacturing biodegradable magnesium stents has been presented in this thesis. The method consisted of the combination of additive manufacturing with investment casting of AZ91D-1 wt.% CaO magnesium alloy using plaster molding. Several aspects have been discussed throughout this research, including the printing of the 3D model, the application of the coating, the manufacture of the molding, and the casting processing. Novel methods and approaches have been proposed and considered throughout the document as well. The main conclusions are emphasized in this section, seeking to respond to the Research Objectives (RO’s) initially outlined in Chapter 1. Chapter 2 reviews the literature on the most pertinent issues related to the handling of magnesium alloys and the fabrication of stents. This chapter describes different magnesium alloy processing approaches, emphasising the addition of elements and the liquid metal treatment through the application of an ultrasound technique. Such approaches can significantly contribute to the development of new magnesium alloys with enhanced physical, chemical, and mechanical properties. These alloys can be utilized to replace the most commonly used metallic structural materials, such as iron, steel, and aluminum. Regarding the investment casting of magnesium alloy in ceramic molding, it is known that a number of methods have been developed and applied to mitigate the effects of reactions that occur during melting and pouring. Besides ceramic shell molding, other methods have also been reported to minimize the reaction issue, such as refractory coatings and protective fluxes and gases. However, none of these studies provided effective solutions to overcome the reactivity in the process, which is still poorly understood. Concerning stent manufacturing, various procedures are detailed, including different materials and conventional and non-conventional processes. Considering the main limitations associated with the use of permanent stents, the need for biodegradable stents manufactured with biodegradable CHAPTER 6. CONCLUDING REMARKS 129 materials is adequately explored, although some important properties must be taken into account in order to ensure the reliable functional performance of the biomedical devices. Regarding the stent fabrication, laser cutting is the primary method of manufacturing stents. However, it is a process that involves several steps and introduces thermal issues that may later compromise the effectiveness of these devices, especially those made from magnesium alloys. In addition to some advantageous characteristics of most processes presented in Table 2.1, a clear opportunity for progress in the evolution of stent models and their fabrication was found. This is why a hybrid methodology combining additive manufacturing and investment casting in plaster molding is proposed in this work for magnesium casting, allowing time and cost savings. Different methods for printing thin-walled parts are discussed in Chapter 3. It has been demonstrated that SLA printing offers better results than FDM printing, namely in terms of surface roughness and detail accuracy. In regard to printing materials, the resin used in SLA has a better surface wettability than PLA used in FDM, which leads to better coating adhesion. FDM technology, however, provides time and cost savings. Therefore, this study presents an optimization technique combining both printing processes, using SLA for the printing of stent models and FDM for the printing of the filling gating system. The mold-metal interface between liquid metal and plaster moldings is also characterized in this chapter through an experimental case study. Different protective coatings for magnesium investment casting in plaster molding are studied and analyzed. The Yttria coating was found to be more effective than Alumina and Sand-based slip coatings in preventing mold-metal reactions. As a result, it should be applied by the dipping method in a few layers due to its unstable stability at higher temperatures. Based on the obtained castings, the effectiveness of Y2O3 is investigated and confirmed in the case study. The samples using Yttria as a refractory coating evidenced a cleaner surface and a more uniform microstructure. In contrast, extreme reaction products and porosity defects have been detected in castings in which no coating has been applied to the 3D printed models. Considering the Yttria instability at elevated temperatures, a thermal cycle optimization is proposed in Chapter 4 for thin-walled magnesium investment castings. Based on TGA and DTG analyzes, a new thermal cycle is recommended that involves a maximum temperature of 450 ºC in order to ensure the proper burning of polymeric materials and the stability of the plaster solution. Furthermore, the optimal casting parameters for the AZ91D-1 wt.% magnesium alloy are experimentally optimized in this work through a fluidity case study. The use of a vacuum atmosphere during melting and pouring (condition YV) is shown to be vital to the occurrence of 130 DEVELOPMENT OF A NEW TECHNIQUE FOR MANUFACTURING BIODEGRADABLE MAGNESIUM STENTS magnesium reactions. Even though it contributes positively to cavity filling, especially for thin-walled parts, it also appears to significantly enhance the reaction between liquid magnesium and the surrounding elements. In general, vacuum is necessary for successful cavity filling of thin-walled castings, although Yttria coating plays a crucial role in mitigating this reaction rate. The casting of magnesium without Y2O3 coating and vacuum (condition N) is recommended for thicker parts. The result is a lower level of surface contamination and a better surface finish, although their fluidity may be affected. In general, a compromise between the molding and pouring temperatures is recommended for successful magnesium casting, since higher temperatures may potentiate more reactions, but will also guarantee a greater cavity filling length. The application of Yttria coating without vacuum assistance (condition Y) reveals to be ineffective for magnesium investment casting because the cracking into the molding cavities acts as a barrier to the cavity filling. The process of manufacturing magnesium stents through investment casting in plaster molding is described and characterized in Chapter 5. In this work, magnesium stents with thicknesses of 0.4 mm and 0.8 mm are successfully fabricated using the proposed methodology, with a good surface finish, high degree of geometric and dimensional rigor, as well as being free of any reaction content. For that, a single and direct filling casting must be adopted. Despite the fact that complete cavity filling can be achieved for any tested casting condition, the highest quality final casting is achieved when Yttria coating and vacuum are not used during the processing (condition N). In this condition, oxygen does not react with the surrounding elements. The residual oxygen present in the furnace produces MgO, which acts as a protective oxide on the melt surface, thereby increasing the surface tension of the metal and preventing mold-metal contact. As demonstrated in Chapter 4, the use of vacuum presupposes the application of a Yttria coating, which improves the fluidity of the alloy and the cavity filling, although a surface reaction layer can be created in the process. This means that the casting condition to be used depends on the final application, as using vacuum guarantees better thin-wall cavity filling, while the absence of vacuum inhibits magnesium reactions. The same compromise exists for pouring and molding temperatures since higher temperatures ensure better cavity filling and lower temperatures mitigate reactions. To sum up, it was demonstrated and proved that the proposed methodology is effective for the thin-walled investment casting of AZ91D-1 wt.% CaO magnesium alloy in plaster molding, optimized for stent manufacturing. This research work offers an opportunity to fill some gaps in the literature on the processing of magnesium alloys, which are important for the necessary expansion of these alloys into the foundry sector. This methodology also allows a more realistic assessment CHAPTER 6. CONCLUDING REMARKS 131 of the potential of investment casting in plaster molding, which has received relatively scant attention in the industrial field. 6.2. Future Work This work examined and discussed the main issue associated with magnesium investment casting in plaster molding, namely for thin-walled applications. The obtained experimental results constitute advances in the state-of-the-art, however further developments are needed to complement the existing facts and address some of their limitations, extending their applicability. In that sense, some suggestions for future research activities in this field are presented below. The mold-metal interaction evaluated in this work can be extended to other experimental cases, including thicker samples. Thus, higher solidification times would be expected and its influence on the occurrence of reactions after the pouring could be more deeply exploited. This work has demonstrated the process's effectiveness; however, further mechanical and corrosion characterization is required in order to validate the proposed methodology and understand its range of applications. There is a lack of information in the literature regarding investment casting magnesium alloys in plaster molding. Therefore, the mechanical properties and degradation rate of magnesium alloy samples obtained in this thesis are of particular significance. Regarding the previous point, other strategies can be employed to improve the properties of the cast samples. Applying the proposed methodology of this work, methods such as liquid metal treatment by ultrasonic technique and chemical refinement by addition of elements could be tested to analyze their influence on the final cast parts. In this study, it was observed that thermodynamic characteristics play an important role during the processing of magnesium alloys. Even though the cavity filling simulation is a challenging task given the large number of factors involved, a numerical study of the processing performed in this research would provide an additional tool for understanding all phenomena. Additionally, it would be possible to reduce the experimental content and save time and money. Lastly, it was seen that the use of vacuum seems to directly influence the surface tension of the magnesium alloy for different casting conditions, which seems to be related to the reaction rate that can occur during the melting and pouring phases. Characterization of the surface tension of the liquid magnesium under different casting conditions is therefore necessary in order to better explain the obtained results. 132 REFERENCES [1] M. Hedin, ‘The origin of the word stent ’, Acta Radiol. , vol. 38, no. 6, pp. 937–939, Nov. 1997, doi: 10.1080/02841859709172106. [2] World Health Organization, ‘The top 10 causes of death’, 2020. https://www.who.int/newsroom/fact-sheets/detail/the-top-10-causes-of-death (accessed Jul. 15, 2021). [3] D. N. Ghista and F. Kabinejadian, ‘Coronary artery bypass grafting hemodynamics and anastomosis design: a biomedical engineering review’, Biomed. Eng. Online , vol. 12, no. 1, p. 129, 2013. [4] T. Simard, B. Hibbert, F. D. Ramirez, M. Froeschl, Y.-X. Chen, and E. R. O’Brien, ‘The Evolution of Coronary Stents: A Brief Review’, Can. J. Cardiol. , vol. 30, no. 1, pp. 35–45, Jan. 2014, doi: 10.1016/j.cjca.2013.09.012. [5] S. Amani, G. Faraji, H. Kazemi Mehrabadi, K. Abrinia, and H. Ghanbari, ‘A combined method for producing high strength and ductility magnesium microtubes for biodegradable vascular stents application’, J. Alloys Compd. , vol. 723, pp. 467–476, Nov. 2017, doi: 10.1016/j.jallcom.2017.06.201. [6] A. Schuesseler, ‘Manufacturing of stents: optimize the stent with new manufacturing technologies’, New Technol. Vasc. Biomater. Fundam. Stent II , pp. 93–106, 2007. [7] D. Mazzaccaro, M. T. Occhiuto, S. Stegher, P. Righini, G. Malacrida, and G. Nano, ‘New technologies in vascular surgery: San Donato’s experience in the last decades’, Eur. Heart J. Suppl. , vol. 18, no. suppl E, pp. E37–E41, Apr. 2016, doi: 10.1093/eurheartj/suw021. [8] Lim Ing Haan Cardiology Clinic, ‘Coronary Angioplasty and Stenting’. http://drliminghaan.com/treatments-and-services/coronary-angioplasty-and-stenting/ [9] S. Garg and P. W. Serruys, ‘Coronary Stents’, J. Am. Coll. Cardiol. , vol. 56, no. 10, pp. S43–S78, Aug. 2010, doi: 10.1016/j.jacc.2010.06.008. [10] A. G. Demir and B. Previtali, ‘Additive manufacturing of cardiovascular CoCr stents by selective laser melting’, Mater. Des. , vol. 119, pp. 338–350, Apr. 2017, doi: 10.1016/j.matdes.2017.01.091. REFERENCES 133 [11] L. Wang, G. Fang, L. Qian, S. Leeflang, J. Duszczyk, and J. Zhou, ‘Forming of magnesium alloy microtubes in the fabrication of biodegradable stents’, Prog. Nat. Sci. Mater. Int. , vol. 24, no. 5, pp. 500–506, Oct. 2014, doi: 10.1016/j.pnsc.2014.08.006. [12] L. Mao et al. , ‘A promising biodegradable magnesium alloy suitable for clinical vascular stent application’, Sci. Rep. , vol. 7, p. 46343, Apr. 2017, doi: 10.1038/srep46343. [13] H. Matsuoka et al. , ‘Bare metal stent implantation for in-stent restenosis with a drug-eluting stent’, J. Cardiol. , vol. 55, no. 1, pp. 135–138, Jan. 2010, doi: 10.1016/j.jjcc.2009.04.006. [14] H. Hermawan, D. Dubé, and D. Mantovani, ‘Developments in metallic biodegradable stents☆’, Acta Biomater. , vol. 6, no. 5, pp. 1693–1697, May 2010, doi: 10.1016/j.actbio.2009.10.006. [15] M. Moravej and D. Mantovani, ‘Biodegradable Metals for Cardiovascular Stent Application: Interests and New Opportunities’, Int. J. Mol. Sci. , vol. 12, no. 12, pp. 4250–4270, Jun. 2011, doi: 10.3390/ijms12074250. [16] H. Meng, J. Liao, Y. Zhou, and Q. Zhang, ‘Laser micro-processing of cardiovascular stent with fiber laser cutting system’, Opt. Laser Technol. , vol. 41, no. 3, pp. 300–302, Apr. 2009, doi: 10.1016/j.optlastec.2008.06.001. [17] R. Erbel et al. , ‘Temporary scaffolding of coronary arteries with bioabsorbable magnesium stents: a prospective, non-randomised multicentre trial’, The Lancet , vol. 369, no. 9576, pp. 1869–1875, 2007. [18] J. Wiebe, H. M. Nef, and C. W. Hamm, ‘Current Status of Bioresorbable Scaffolds in the Treatment of Coronary Artery Disease’, J. Am. Coll. Cardiol. , vol. 64, no. 23, pp. 2541– 2551, Dec. 2014, doi: 10.1016/j.jacc.2014.09.041. [19] D. Lim et al. , ‘Suggestion of Potential Stent Design Parameters to Reduce Restenosis Risk driven by Foreshortening or Dogboning due to Non-uniform Balloon-Stent Expansion’, Ann. Biomed. Eng. , vol. 36, no. 7, pp. 1118–1129, Jul. 2008, doi: 10.1007/s10439-008-95041. [20] R. A. Byrne, M. Joner, and A. Kastrati, ‘Stent thrombosis and restenosis: what have we learned and where are we going? The Andreas Grüntzig Lecture ESC 2014’, Eur. Heart J. , vol. 36, no. 47, pp. 3320–3331, Dec. 2015, doi: 10.1093/eurheartj/ehv511.