Multi-material additive manufacturing technologies for Ti-, Mg-, and Fe-based biomaterials for bone substitution

The growing interest in multi-functional metallic biomaterials for bone substitutes challenges the cur- rent additive manufacturing (AM, = 3D printing) technologies. It is foreseeable that advances in multi-material AM for metallic biomaterials will not only allow for complex geometrical designs, but also im- prove their multi-functionalities by tuning the types or compositions of the underlying base materials, thereby presenting unprecedented opportunities for advanced orthopedic treatments. AM technologies are yet to be extensively explored for the fabrication of multi-functional metallic biomaterials, especially for bone substitutes. The aim of this review is to present the viable options of the state-of-the-art multi- material AM for Ti-, Mg-, and Fe-based biomaterials to be used as bone substitutes. The review starts with a brief review of bone tissue engineering, the design requirements, and fabrication technologies for metallic biomaterials to highlight the advantages of using AM over conventional fabrication methods. Five AM technologies suitable for metal 3D printing are compared against the requirements for multi- material AM. Of these AM technologies, extrusion-based multi-material AM is shown to have the greatest potential to meet the requirements for the fabrication of multi-functional metallic biomaterials. Finally, recent progress in the fabrication of Ti-, Mg-, and Fe-based biomaterials including the utilization of multi-material AM technologies is reviewed so as to identify the knowledge gaps and propose the directions of further research for the development of multi-material AM technologies that are applicable for the fabrication of multi-functional metallic biomaterials. Addressing a critical bone defect requires the assistance of multi-functional porous metallic bone sub- stitutes. As one of the most advanced fabrication technology in bone tissue engineering, additive manufacturing is challenged for its viability in multi-material fabrication of metallic biomaterials. This article reviews how the current metal additive manufacturing technologies have been and can be used for multi-material fabrication of Ti-, Mg-, and Fe-based bone substitutes. Progress on the Ti-, Mg-, and Fe-based biomaterials, including the utilization of multi-material additive manufacturing, are discussed to direct future research for advancing the multi-functional additively manufactured metallic bone biomaterials.


Bone tissue engineering
Bone takes part in the key functions of the human body for locomotion, protection of soft tissues and organs, and mineral home- through which old, microcracked bone is resorbed and replaced by new bone [4] . Despite being able to regenerate, bone loss from traumas or diseases often leads to non-unions and critical-size bone lesions ( i.e. , > 5 cm, e.g. , in the femur and tibia [5] ). A critical-size bony lesions is defined as a lesion that cannot be healed naturally during the patient's lifetime [6] . In such cases, bone replacements are necessary.
Until today, autografts, or bone replacements collected from and implanted in the same person, remain the gold standard of the clinical treatments of critical-size bony lesions [7] . However, the volume of autografts that can be harvested to replace critical-size bony lesions, for example, from the posterior iliac crest, is limited to 33.82 cm 3 [8] , while the volume of tibia defects may amount to 122 cm 3 [9] . Although allografts ( i.e. , bone replacements from donors) can compensate the shortcomings of autografts, the risks of disease transmission as well as immunogenic incompatibility pose other types of challenges. Driven by the need to provide an alternative bone replacement approach, bone tissue engineering, which focuses on regenerating bone tissue using synthetic biomaterials, has shown promising progress in the reconstruction of large-scale bony lesions [10] .

Metallic biomaterials for load-bearing bone substitutes
Ideal synthetic biomaterials for bone substitutes must be biocompatible and designed to mimic the extracellular matrix of the natural bone. They should provide bone regeneration environment, facilitate complete bone regeneration, withstand dynamic mechanical loading, and degrade along with the formation of new bone [11] . A successful bone substitute first allows mesenchymal stem cells (MSCs) to adhere onto its surface and differentiate into bone cells, followed by inward mineralization [12] . In the design of an ideal bone substitute, the mechanical properties, pore sizes, pore distribution, porosity, surface characteristics, and biodegradability of the biomaterial should all be taken into consideration.
Synthetic bone substitutes can be made using metallic biomaterials, bioceramics, biopolymers, or composite biomaterials. Metallic biomaterials have been widely used for functional and load-bearing bone substitutes due to their mechanical properties that are superior to those of bioceramics, biopolymers and composite biomaterials. Most of bulk metals, however, are known for having mechanical properties that are much higher than those of the native bone, which induces bone tissue resorption after implantation as a result of stress shielding [13] . But, the excessively high mechanical properties of metals can be tuned by increasing their porosity. Increasing the porosity of metallic biomaterials cannot only reduce their mechanical properties but also improves their permeability for cells and nutrients and facilitates angiogenesis and bone ingrowth [14] . The porosity of metallic biomaterials should be at least 50% while pore sizes > 300 μm are recommended to ensure better progress of the bone tissue regeneration process [15 , 16] . The porous design of metallic biomaterials should be tailored to ensure an optimum trade-off between the mechanical properties for load bearing and the porous structure needed for vascularization.
In addition, the surface of metallic biomaterials must be osteoconductive to promote adhesion, proliferation, and differentiation of the relevant cells such as MSCs. An osteoconductive surface can be obtained through generating certain surface characteristics, such as surface roughness and curvature. As bone tissue regeneration occurs more likely on concave surfaces as compared to convex or planar surfaces, large concave surface curvatures are preferred for an enhanced growth rate [17] . Moreover, nanofeatures and surface nanotopography can also induce the osteogenic differentiation of MSCs and encourage osteoblasts adhesion [18][19][20] .
At the early stage of bone defect regeneration, the entire mechanical support relies on the implanted biomaterial. The mechanical integrity of the biomaterials should be sustained for about 3-12 weeks to support the upper limb healing process, while the lower limb requires approximately 12-24 weeks [21] . As the regenerated bone slowly regains its strength, the biomaterial is allowed to degrade gradually. Therefore, the biodegradation rate of metallic bone substitutes should be adjusted such that the loss in the load-bearing capacity of the implanted biomaterial is balanced by the gain in the structural integrity of the bone. Furthermore, the biodegradation products must be delivered in acceptable quantities and cytotoxicity levels to ensure that they are tolerated by the human body.
With all the requirements for ideal bone substitutes fulfilled, the regeneration of bony lesions assisted by metallic biomaterials could result in healthy and functional bone. To obtain such ideal biomaterials, advanced fabrication technologies that enable the precise arrangement and control of complex geometries are required.

Fabrication technologies for metallic biomaterials used as bone substitutes
Over the years, porous metallic biomaterials have been produced using conventional fabrication technologies, mostly based on powder metallurgy, such as metal injection molding [22] or the space-holder method [23 , 24] . Even with the remarkable progress made in these fabrication technologies, certain limitations, such as the impossibility to control the geometry and distribution of pores precisely, as well as dimensional inaccuracies, remain.
Additive manufacturing (AM) has recently emerged as a powerful method for the fabrication of biomaterials, including metallic biomaterials aimed for bone tissue regeneration [25] . AM technologies enable high-precision fabrication with high flexibility in the internal and external macro-and micro-architecture of orthopedic implants [26] . Through controlled AM fabrication processes, geometrical and topological porous characteristics of metallic biomaterials can be precisely tuned, leading to improved bone-mimicking mechanical properties [27 , 28] , altered biodegradation kinetics [29 , 30] , enhanced bone tissue regeneration rates [31][32][33] and the formation of an extensive, interconnected osteocyte lacunocanalicular network [34 , 35] . However, some other properties, including hardness, wear resistance, anti-ferromagnetic properties, or antibacterial properties, cannot be simply adjusted through geometrical design, as they require the adjustments of the properties of the underlying base material(s) prior to AM processing.
As for the biocompatibility, most metallic biomaterials have a relatively low intrinsic osteogenic and osteoimmunomodulation potential as compared with, for example, Ca-P-based bioceramics [36] . Their presence as foreign body objects is often a risk factor for prolonged chronic inflammation [37] . The high surface area to volume ratio of porous AM metallic biomaterials raises another challenge in preventing bacteria colonization. These issues are currently being addressed by taking an additional post-AM step of surface biofunctionalization to add bioactive agents to the porous surfaces in order to improve the performance of AM biomaterials including their osteogenic properties, and to prevent infection [38] .
The existing technologies for the fabrication of multi-functional porous AM metallic biomaterials strongly rely on the availability of pre-alloyed materials and the post-AM surface biofunctionalization, thus, being a two-step process. There is a strong need to upgrade the AM technologies so that they can realize the desired spatial distribution and bonding of multiple materials, thereby enabling the in situ synthesis of multiple materials in one single AM process.
Utilizing multi-material AM technologies will enable the fabrication of multi-functional porous AM metallic biomaterials with region-specific performance such that the material types and compositions can be specifically placed at different scales within the biomaterials design. The mechanical properties of porous AM metallic biomaterials, including hardness, can be realized, in accordance with those of cortical and trabecular bone by changing not only the porosity but also the material types or compositions. Enhanced osteoconductivity and antibacterial properties with appropriate inflammatory responses can be obtained through the right compositions and distributions of bioactive agents on the metallic surfaces. Varying the compositions and types of materials that have distinct biodegradation properties, for example, from bio-inert or slowly biodegrading biomaterials to fast biodegrading materials, will create complex profiles of biomaterials properties over time, as clinically required.
Despite the fact that numerous properties of porous AM metallic biomaterials can be improved by using multi-material metallic AM technologies, the currently available literature on these technologies is relatively scarce [39] . Recent literature on the AM technologies for bone tissue engineering is mainly focused on the commercially available AM processes and the choices of existing biomaterials, including biocompatible metals [40][41][42][43][44] , and geometrical and topological designs of porous AM metallic biomaterials [45][46][47][48][49] in relation to their mechanical and biological performance. Regarding the multi-material AM technologies, multi-material polymers are advancing the frontiers of multimaterial AM concepts due to the relative simplicity of the involved processes and the compatibility of those materials with multiple AM technologies [50] . In situ multi-material AM for metals and metal-ceramic composites have been mainly reviewed for industrial applications, such as in aerospace and automotive applications [51][52][53] , but not specifically for biomedical applications.
In this review, we present the prospects of using the currently available metallic AM technologies for in situ or ex situ multimaterial fabrication of multi-functional Ti-, Mg-, Fe-based bone substituting biomaterials. The key principle of each of the AM technologies as well as their its advantages and limitations for multi-material fabrication are described, analyzed, and compared. Recent research on Ti-, Mg-, and Fe-based biomaterials and the (multi-material) AM technologies for these biomaterials is also reviewed. Finally, the current challenges and future perspectives in metallic multi-material AM technologies are provided.

Multi-material metallic AM technologies
To build multi-functional porous AM metallic biomaterials using multi-material AM technologies, multiple materials must be delivered during the fabrication processes, and strong bonding between biomaterials of different types or compositions must be ensured. Multiple material delivery systems and their bonding processes vary, depending on the particular AM technology used. Meeting these two requirements is essential for successful multimaterial fabrication, as they strongly influence the performance of the resulting multi-materials, especially at the interfaces. Some AM technologies only allow the interchange (and bonding) of different materials between layers to lead to the generation of planar multi-material interfaces, while others can deliver and bond any materials throughout the build to achieve complex multi-material interfaces.
There are 5 AM technologies, according to the ASTM F2792-12a standard [54] , that are suitable for the fabrication of metallic biomaterials, including powder bed fusion, directed energy deposition, sheet lamination, binder jetting, and material extrusion. Each of these AM technologies has its own capabilities and limitations, and may or may not meet the requirements for multi-material fabrication.

Powder bed fusion
Powder bed fusion works by delivering a layer of metal powder and then selectively melting or sintering the desired area by using laser or electron beam. The process iteratively continues by lowering the build plate, spreading another metal powder layer over the previous layer, and selectively melting or sintering this layer and the layer(s) beneath. Selective laser melting (SLM) is the most common powder bed fusion technology for fabricating metallic biomaterials. This technology usually utilizes only one powder bed dispensing system for one single metal powder, which makes the fabrication process challenging to accommodate the in situ delivery of multiple materials, unless a multi-material powder mixture is prepared beforehand. By using blended metal or metal-ceramic powder mixtures in the powder bed ( Fig. 2 a), multi-material SLM has been applied for the multi-material fabrication of Ti-based [53][54][55][56][57][58][59][60][61][62][63][64] and Mg-based [67][68][69][70] biomaterials.
In situ multi-material SLM for Ti-based biomaterials enhanced the hardness and wear resistance through the in situ reinforcement of titanium with SiC [55] , Si 3 N 4 [56] , TiB 2 [57 , 58] , or hydroxyapatite (HA) [59 , 60] . In addition, the in situ SLM of titanium with Mo [61] , Nb [62 , 63] , and Ta [64] , improved biomechanical compatibility with bone in terms of elastic modulus, in comparison with the elastic moduli of pure Ti and Ti-6Al-4 V. Moreover, the in situ alloying of Ti-6Al-4 V with Cu [65 , 66] has been shown to upregulate the angiogenesis-related genes and demonstrate antibacterial properties.
For Mg-based biomaterials, the in situ alloying of Mg with Zn during SLM [67] has been explored and the relationships between Mg-Zn composition, defects, and mechanical characteristics have been studied. In addition, pre-alloyed Mg-based biomaterials ( e.g. , ZK60 and Mg-3 Zn) have been in situ alloyed with rare earth elements ( e.g. , Nd [68] and Dy [69] ) during SLM, for improved corrosion resistance and prolonged biomaterials integrity. Furthermore, the in situ alloying of ZK60 with Cu [70] has been found to provide the base alloy with antibacterial properties and improved the compressive strength. Furthermore, for Fe-based biomaterials, SLM of pre-milled Fe and Mn powders enhanced the biodegradation rate and mechanical properties of the iron [71 , 72] .
To obtain the intended multi-functionality for the biomaterials, optimum SLM processing parameters must be selected to allow for the sufficient diffusion of alloying elements into the base materials [55 , 56] . Tuning laser energy density, while avoiding the formation of undesirable internal pores and excessive melting, is challenging due to the distinctly different thermal properties of multiple materials. Some alloying elements have very high melting points and cannot be completely melted and diffused, and as a result partially melted material remains next to the matrix [62 , 64] . To bridge the gap in thermal properties, the particle size distributions of multi-material powders can be varied, with a higher melting point material having smaller particle sizes [63] . In addition, post-AM heat treatment may be necessary to improve the multi-material diffusion and homogeneity in chemical composition [62] . Furthermore, the delivery of multi-material powders to the powder bed is an equally important aspect to ensure a uniform multi-material distribution within the resultant biomaterial [64] . Since powder bed fusion operates with high thermal energies, this AM process is also prone to high temperature gradients, which can cause cracks and distortions in the multi-material structure.
Using powder bed fusion for building in situ multi-functional AM metals or metal-ceramic composites requires complex and comprehensive process optimization to ensure proper selection of the laser power and the other process parameters that have to be aligned with the physical properties and powder characteristics of each of the materials. In addition, the steps required for the recycling and reuse of the leftover materials in the powder bed need to be considered.

Directed energy deposition
In directed energy deposition, a metallic material is delivered in the form of powder or wire through a nozzle, which is then melted by using thermal energy. Laser engineered net shaping (LENS) is the most common commercial process used for directed energy deposition. This technology offers a straightforward multi-material delivery system through multiple nozzles and the capability of in situ deposition and synthesis of different materials to obtain multi-material gradients in the structure ( Fig. 2 b). In addition, compositionally graded structures can be produced by depositing one material and gradually replacing that material with another. Using multi-material LENS, Ti-based biomaterials have been in situ alloyed with boron [73 , 74] or reinforced with HA (under a N 2 atmosphere) [75] , CaP [76] , Nb-Zr-Ta elemental powders [77] , and Zr-BN [78] to improve the hardness and wear resistance of the base biomaterial. Compositionally graded Co-Cr-Mo coating on Ti-6Al-4 V surface [79] and TiO 2 coating on Ti surface [80] have been fabricated to minimize the wear-induced loosening of metal-on-metal implants.
Similar to the laser-based multi-material AM processes based on powder bed fusion, the vast differences in the thermal properties of multiple materials have to be comprehensively investigated, in order to match these with the laser parameters and process parameters to create multi-functional structures with no structural or metallurgical defects. Despite the fact that directed energy deposition has the ability of in situ deposition of multiple materials, this AM process is less suitable for fabricating structures with fine geometries or hollow passages, which are often required for complex porous biomaterials designs [81] . Apart from that, this technology suffers from the common drawbacks of laser-based AM processes, i.e. , high thermal gradients, which can induce residual stresses and metallurgical defects.

Sheet lamination
Sheet lamination is performed by stacking, bonding, and cutting foil materials into a 3D structure, after which an additional machining or milling process is required to make a specific structure. Ultrasonic consolidation is the most commonly used technique for bonding metallic foils [82][83][84] . Although the main advantages of this technology lie in the low operating temperature and the capability of producing large-scale structures at low costs, this technology has not yet been reported for the fabrication of multi-materials for bone substitutes.
Although sheet lamination accommodates simultaneous multiple foil feeding, due to the stacking and bonding of foils ( Fig. 2 c), multi-material interfaces occur only in the planar direction. The major drawback of this technology concerns the voids occurring along the foil interfaces due to excessive and/or insufficient welding, surface roughness, or inaccuracy in the manual positioning of the foils. Such defects create a low bonding strength between the layers and increase the possibility of delamination, which leads to poor mechanical properties of the resulting structure under shear and tensile loading. Considering the capabilities of this technology, the sheet lamination technology is more suitable for an embedded application, rather than multi-material AM for porous bone-substituting biomaterials, where mechanical integrity is one of the most important criteria.

Binder jetting
Binder jetting operates in a similar manner to powder bed fusion, but instead of applying thermal energy to fuse metal powder feedstock, an adhesive liquid is dispensed on the surface of the powder bed, bonding powder particles to form a desired structure. Since an adhesive liquid is delivered, the compatibility of the binder with metal powders, as well as the bonding properties are of great importance. Several types of adhesives are available, such as polymer binders, particle suspension inks, metallic salt compounds, and organometallic inks [85] .
For multi-material binder jetting, a powder bed similar to the one required in powder bed fusion processes is utilized, meaning that this technology requires a carefully designed multi-material powder handling system for the preparation of blended multimaterial powder prior to AM, as well as the systems for recycling and reusing the leftover powders after AM. Multi-material binder jetting with blended multi-material metal powders ( Fig. 3 a) has been studied for porous Fe-Mn [86] and Fe-Mn-Ca [87] biomaterials to enhance the biodegradation rate of iron and to develop anti-ferromagnetic properties.
Porous binder-jetted biomaterials ( e.g. , Fe-Mn [86] and Fe-Mn-Ca [87] ) are composed of adhesive-bound powders. As such, post-AM debinding and sintering are needed to remove the binder and subsequently fuse multi-material powder particles together. Since the binder-jetted adhesive-bound porous structure is built inside a powder bed, it is required to remove loose powder particles from pores before post-AM heat-treatment, without damaging the structural integrity of the biomaterial. If loose powder particles inside pores are not completely removed, they will fuse into the structure during sintering, compromising the interconnectivity of the pores and reducing the fidelity of the morphological properties of the final biomaterial. The as-printed structure is usually cured to strengthen the adhesive bonding [86] , then high-pressure air is applied to remove loose powder particles before sintering [86] .
Although removing loose particles in a green structure is demanding, there are no residual stresses created during the binder jetting process, due to the absence of direct laser heating and rapid cooling during fabrication. However, post-AM sintering causes structural shrinkage, as a consequence of binder decomposition and powder particle rearrangement and integration. As an example, Fe-Mn-Ca biomaterials shrank about 11.7% in all directions [87] .
In summary, the application of binder jetting for the fabrication of multi-functional porous AM metals or metal-ceramic composite biomaterials not only requires intricate multi-material powder handling systems before and after AM, but also necessitates some steps to remove loose powder particles from the adhesive-bound porous structure and to sinter the particles through a subsequent post-AM heat treatment.

Material extrusion
Extrusion-based AM technology works by continuously pushing metal or ceramic powder-based feedstock through a nozzle or nozzles to build a 3D structure layer by layer. The feedstock materials should have viscoelastic properties that enable them to flow through the nozzle, solidify immediately upon extrusion, and maintain the shape even when they are stretched over the space of the underlying layers [88] . The feedstock materials can be designed to comprise powdered materials of various compositions mixed with a polymer binder. The compatibility of the feedstock material with the binder should avoid undesired interactions during mixing, extrusion, debinding, and even sintering. For multi-material AM purposes, more than one nozzle can be integrated to deliver different f eedstock materials and achieve complex multi-material interfaces within the structure. Alternatively, one feedstock of multi-material can be prepared and delivered through one single nozzle.
Material extrusion with blended multi-material powder-based feedstock ( Fig. 3 b) has been pursued [89][90][91] , since it has no drawbacks inherent to the laser-based AM technologies that were described earlier, such as residual stresses, cracks, distortions, and even metallurgical defects. Using this technique, Fe-based biomaterials ( e.g. , Fe-CaSiO 3 ) has been realized for bone cancer treatment as well as for regenerating cortical bone defects [89] . The extruded Fe-CaSiO 3 composite is composed of an adhesive-bound multi-material powder mixture, which requires post-AM debinding and sintering. Since the adhesivebound structure is not built in the powder bed, it saves the powder handling steps of recycling and reusing multi-material powder, as well as the post-AM step to remove loose powder particles.
The main limitation of this technology lies in the printing resolution, which is strongly dependent on the powder particle size and the nozzle diameter. In addition, building a part with a large aspect ratio and an overhanging structure can only be achieved by controlling the composition of the feedstock and its rheological properties in order to ensure consistent flow and fast solidification. Designing suitable multi-material feedstock is the most crucial step to achieve complex structures with multi-material interfaces.

Selection of multi-material AM technologies for the fabrication of multi-functional metallic bone substitutes
Choosing a suitable multi-material AM technology for fabricating multi-functional porous metallic biomaterials requires a comprehensive knowledge of the capabilities and limitations of each of the processes described earlier. The basic requirements for a multi-material AM technology concern the capabilities of delivering multiple materials, achieving complex multi-material interfaces, and ensuring strong bonding between the materials of different types and compositions.
With respect to the material delivery systems that influence the multi-material interaction, three systems have been implemented in AM technologies, namely powder bed dispensers, nozzles, and foil material delivery systems. The AM processes based on powder bed fusion, binder jetting, directed energy deposition, and material extrusion allow for the building of complex multi-material interfaces, while sheet lamination can only create planar multi-material interfaces.
Moreover, the cost effectiveness in relation to the material delivery system is another issue to address. Using a powder bed dispensing system to build complex multi-material interfaces requires an advanced powder handling system to allow for the recycling and reuse of multi-material powder mixtures. A post-AM step is needed to remove the unbound or unmolten powder particles entrapped in the pores of structures. This step is demanding, especially for binder jetting, as the as-built structure is bound only by an adhesive material and is, thus, vulnerable. Multi-material deposition using nozzles in material extrusion and directed energy deposition processes is capable of controlling the quantity of the materials needed, while recycling and reusing the powder or a post-AM step of removing loose powder particles are not required. Nozzles are, therefore, a more suitable material delivery means for multi-material AM.
In addition to the material delivery system, bonding between multiple materials is of critical importance for the performance of multi-functional porous AM metallic biomaterials. Two material bonding approaches are usually used in multi-material AM technologies including direct heating by laser or adhesive bonding by using a binder, followed by post-AM heat treatment. The chosen type of bonding also determines the final microstructure of the multi-material. Using direct heating in powder bed fusion and directed energy deposition processes enables the in situ AM of multi-materials. This approach, however, requires advanced process control and management, due to the integration of laser energy into the AM machine and the vast differences in thermal properties between multiple materials. Adhesive bonding in binder jetting and material extrusion, on the other hand, is only capable of the ex situ AM of multi-materials and requires post-AM debinding and sintering. The biocompatibility of the residual binder entrapped inside the multi-material structure may be an important issue to consider.
AM using direct heating usually produces mechanically stable structures. However, there is a risk of having high thermal gradients and metallurgical mismatch, which may induce residual stresses or structural defects. Furthermore, this is a challenging approach for fabricating metals with low boiling points and high vapor pressures. On the other hand, the post-processing of adhesive-bound AM structures is usually conducted in a homogenously heated furnace, which allows the materials to diffuse slowly and bond strongly through sintering. Although sintering leads to the shrinkage of the structure, this can be compensated for during the design phase and minimized during fabrication. In addition, the parameters of the sintering processing can be optimized to yield high-density structures, which presents an opportunity to fabricate not only porous but also fully dense biomaterials for bone substitutes. Based on the advantages and disadvantages of each of bonding approaches, adhesive bonding followed by sintering seems to be a better choice for multi-material bonding.
Comparing the five available multi-material AM technologies, material extrusion appears to be the most straightforward option for multi-material fabrication of multi-functional bone substituting biomaterials [89] . Material extrusion is capable of building multimaterial interfaces. Moreover, material extrusion does not directly heat the feedstock materials to fuse the multiple materials, making it less probable to create metallurgical defects in the structure. In addition to Fe-CaSiO 3 scaffolds that were aimed for bone substitutes [89] , multi-material extrusion AM has been applied to fabricate materials for other types of application including hydroxyapatite-based surgical biomaterials, graphene-based materials for biomedical electronic devices, and other multi-element materials systems such as Fe-Ni-Cr, Al-Mg-Cr-Fe and Ag-Cu for various applications [90][91][92][93][94] .

Biomaterials for multi-functional metallic bone substitutes
An ideal metallic biomaterial for bone substitutes should not only be biocompatible and fulfill the design requirements in geometry, but also have mechanical and biological functionalities that are needed for bone replacement and regeneration. In the case of the bone substitutes aimed for permanent load-bearing applications, long-term mechanical properties, corrosion resistance, and osseointegration are the most important factors to consider. However, for temporary bone repair applications, biodegradability, short-term mechanical integrity, and biological functionalities for osteosynthesis are the critical considerations.
For bone substitutes to achieve multiple functionalities, more than one material is often needed. The concept of multi-material synthesis and processing for improved physical characteristics, chemistry, and mechanical performance, which may be achieved through powder metallurgy, and for enhanced biological performance such as cell attachment through surface modification, has been established long before the development of multimaterial AM technologies. Therefore, developing multi-material compositions specifically suitable for AM technologies requires multi-disciplinary knowledge about the intrinsic properties of each of the materials and the understanding of the interactions between different materials in order to meet the functional requirements of the synthesized biomaterials aimed to be used as bone substitutes.
In this section, recent progress in developing Ti-, Mg-, and Fe-based biomaterials for bone substitutes, including advances in (multi-material) AM technologies applied for the development of permanent or temporary bone replacements, is reviewed.

Ti-based biomaterials
Titanium and its alloys are promising biomaterials for permanent bone replacement due to their biocompatibility, lightweight, high strength, and superior corrosion resistance. To be able to use them as long-term bone replacements, titanium and its alloys should possess bone-mimicking mechanical properties, which can be obtained through the optimization of the geometry of the porous structure and realized using AM technologies ( Fig. 4 a) [ [95][96][97][98][99][100][101] . Porous Ti-based bone substitutes also improve osseointegration and can be made patient-specific ( Fig. 4 b) [102] , which is essential for preventing implant failure due to inadequate integration of the implant into the surrounding bony environment.
In craniofacial and orthopedic applications, patient-specific AM Ti-based implants have shown remarkable improvements over other alternatives, since they are fabricated to fit precisely into bony defects. This is beneficial, since it reduces the probability of implant failure due to loosening. In a six-month follow-up of a custom-designed AM titanium implant for a large skull defect, no complications were observed [103] . In addition, a clinical study on 21 patients with custom-made AM Ti-6Al-4 V implants during a follow-up of 6-24 months demonstrated osseointegration and skull symmetry [104] . For orthopedic applications, patient-specific AM titanium implants were reported to be effective with no complications, when replacing distal tibia bone defects with multiple fractures in the foot and talus [105] . Furthermore, a success rate of 86.7% has been reported in a 22 month follow-up study on 15 patients treated with patient-specific AM titanium implants for foot ankle deformity corrections and arthrodesis procedures [106] . Overall, AM patient-specific designs improved the mechanical stability of Ti-based implants.
Surface biofunctionalization of AM porous Ti-based biomaterials have been performed to improve their bioactivity and induce antibacterial properties to prevent implant-associated infections. The integration of Ca and P elements and immobilized Ag nanoparticles into the surface of porous AM Ti-6Al-4V implants ( Fig. 4 c) has been demonstrated to afford AM biomaterials with antibacterial properties against methicillin-resistant Staphylococcus aureus (MRSA), while stimulating the growth of human MSCs [38] . Moreover, the surface bioactivity of AM porous Ti-6Al-4V biomaterials has been shown to be enhanced with the incorporation of Sr, which induces more bone formation and stimulates bony ingrowth [107] .
The in situ SLM of Ti powders with SiC (in a molar ratio of 8:3) and Si 3 N 4 (in a molar ratio of 9:1) has been reported to result in viable TiC-Ti 5 Si 3 and TiN-Ti 5 Si 3 composites [55 , 56] . During the process, both TiC and TiN phases were formed through a dissolution/precipitation mechanism and subsequent grain growth [55 , 56] . The TiC phase reached the most refined dendritic morphology and a dendritic trunk length of 4.5 μm, at a laser power of 80 W and a scanning speed of 0.2 m/s ( Fig. 5 a) [55] . The TiN phase exhibited an optimum refined near-round morphology at a laser energy density of 5 kJ/m and a scanning speed of 0.2 m/s [56] . As a result of the grain refinement strengthening, the TiC-Ti 5 S 3 composites achieved a lower friction coefficient of 0.2 as compared to pure Ti (~1.3) and a microhardness value of 980.3 HV, which is three times higher than that of pure Ti [55] . Similarly, the TiN-Ti 5 S 3 composites had an improved, uniform wear rate of 6.84 × 10 -5 mm 3 /Nm with a low friction coefficient of 0.19 [56] .
Moreover, the fine eutectic TiB grains formed through the in situ dissolution of Ti and B and subsequent precipitation were distributed along the Ti grain boundaries ( Fig. 5 b). The in situ LENS Ti-TiB composites made of pure Ti and 1.6 wt% B powders enabled the design of a quasi-continuous network microstructure when fabricated with 200 W laser power [73] . The homogenous microstructure of the Ti-TiB composites led to a greater hardness value of 392.6 HV and better wear resistance with a wear rate of 2.4 × 10 -3 mm 3 /Nm as compared to those of pure Ti (345.5 HV and 6.1 × 10 -3 mm 3 /Nm, respectively) [74] . In addition, the in situ SLM of Ti-6Al-4V with 3 wt% TiB 2 powders fabricated using a laser power of 240 W created needle-like TiB grains of 0.5-1 μm in size. The Ti-6Al-4V-TiB composites possessed two times higher wear resistance compared with the Ti-6Al-4V alloy [57] .
A series of in situ SLM trials to reinforce pure titanium with 2 − 5 wt% HA (nano-sized) under different process conditions resulted in varied microstructures ( Fig. 5 c) and improved microhardness and nanohardness values [59] . However, the fracture toughness of the composites decreased from 3.41 MPa m 1/2 to 0.88 MPa  [64] with permission from Elsevier). m 1/2 as the HA content increased [60] . The reduced toughness was due to the formation of the fragile CaTiO 3 and Ti 5 P 3 phases that are susceptible to crack initiation and propagation. Nevertheless, the toughness of the Ti-HA composites was still in the range of those reported for the native trabecular bone. Additionally, the in situ LENS of Ti-6Al-4V with 5 wt% HA [75] and pure Ti with 10 wt% CaP [76] resulted in significant ( i.e. , up to 95%) improvements in wear resistance. In in situ Ti-6Al-4V-HA and Ti-CaP composites, the CaTiO 3 and Ca 3 (PO 4 ) 3 phases were developed, which formed a tribological layer protecting the biomaterial surface [75 , 76] .
Apart from hardness and wear resistance, the in situ formation of β-phase Ti alloys ( Fig. 5 d) is also desired to improve the match between the elastic moduli of Ti-based biomaterials and the bone.
In situ SLM using blended elemental powders of Ti-6Al-4V-10Mo [61] , Ti-40.5Nb [63] , Ti-35Nb [62] and Ti-50Ta [64] (in wt%) have been shown to result in significantly decreased elastic moduli ranging between 73 and 84.7 GPa as compared with α-phase pure Ti or dual-phase Ti-6Al-4V (105 − 120 GPa). In addition, the β-phase Ti alloy (Ti-35Nb-7Zr-5Ta) fabricated through in situ LENS of elemental powders has been found to demonstrate a higher trans-passive potential of > 5 V, confirming its superior corrosion resistance when compared to pure titanium ( i.e. , ~1.5 V) [77] . A higher trans-passive potential can be attributed to higher adherent oxides including not only TiO 2 but also Nb 2 O 5 , Ta 2 O 5 , and ZrO 2 that are developed on the alloy surface as the alloy is corroded [77] .
Regarding the biological performance, in situ SLM Ti-6Al-4V-Cu made from Ti-6Al-4V and 6 wt% Cu powders not only maintained a high in vitro proliferation rate and alkaline phosphatase (ALP) activity ( Fig. 6 a-b), but also controlled the inflammatory responses [66] . Macrophages were found to be less active on the surfaces containing Cu and the pro-inflammatory cytokine of IL-6 was significantly suppressed ( Fig. 6 c-d) [66] . Furthermore, in situ Ti-6Al-4V-6Cu allowed the proliferation of human umbilical vein cells (HUVECs) and remarkably upmodulated the angiogenesisrelated gene expressions of VEGF-A ( Fig. 6 e-f) [66] . The inclusion of 6 wt% Cu also demonstrated antibacterial properties against E. coli and S. aureus ( Fig. 6 g-h) [65] .
AM Ti-based alloys have displayed a promise for the treatment of critical bone defects. The state-of-the-art in situ multi-material AM using powder bed fusion and directed energy deposition has demonstrated the ability to develop multi-functional Ti-based biomaterials. Most of the abovementioned studies on multi-material AM Ti-based materials were focused on the improvements in microstructure and mechanical characteristics, leaving a large scope for validation in terms of biocompatibility. Further research should be conducted to clarify the achievable multi-functionalities that also include the biocompatibility of additional alloying elements and the constituents that are formed in the multi-material Ti environment.

Mg-based biomaterials
Magnesium and its alloys have been extensively studied for temporary bone fixation devices and have been reported to have a great potential for applications in low load-bearing bone substitution [108] . The elastic modulus of magnesium is relatively low, which aids in preventing mechanical failure due to stress shielding. Moreover, the biodegradability of Mg provides it with a unique potential for temporary bone replacement.
Mg-based biomaterials are expected to maintain mechanical integrity at the early stage of bone healing and to be eventually replaced by de novo bone tissue [109] . Magnesium degrades rapidly and releases hydrogen, which can be problematic. For example, excessive hydrogen gas creates mechanical interferences during bone healing. Additionally, the diffusion of hydrogen into subcutaneous tissue can lead to imbalance in blood parameters [110] , which is why the biodegradation rate of Mg should be controlled.
Surface biofunctionalization of bulk Mg alloys has been performed to improve their mechanical integrity during biodegradation and to enhance their biocompatibility [109] . Coating Mg alloys with Ca-P [111 , 112] , and Si [113] provides these alloys with protective layers that reduce their corrosion rates and prevent significant mechanical losses at an early stage. These surface modifications also contribute to enhancing bone growth during biodegradation. Another approach to controlling the biodegradation rate of Mg-based biomaterials is to alloy Mg with rare earth (RE) elements. The addition of RE elements, such as Dy [114] , Gd [115] , Nd [116] , and Y [117 , 118] to magnesium through conventional powder metallurgy methods has been commonly used to improve its corrosion resistance.  [118] with permission from Elsevier). WZ21 presents a combination of an appropriate degradation rate, mechanical integrity, and enhanced bone tissue regeneration performance. (b) Complete biodegradation of Mg-Ca-Zn screws (with a diameter of 2.3 mm and a length of 14 mm) and the regeneration of distal radius fracture after 12 months of implantation (Reproduced from [120] ).
An in vivo study, comparing fast and slowly degrading Mgbased biomaterials of ZX50 and WZ21 ( Fig. 7 a), was performed on rat femurs for up to 36 weeks to evaluate their biodegradation profiles and biocompatibility [118] . ZX50 that contained no RE elements had a rapid biodegradation profile. The deterioration of the mechanical integrity occurred in the first week and complete biodegradation occurred in 16 weeks. As a consequence, bone healing was disturbed due to massive hydrogen release ( i.e. , ~270 mm 3 daily) [118] . In contrast, WZ21 that contained the rare earth element Y could maintain its mechanical integrity for up to 4 weeks. The daily release of approximately 130 mm 3 hydrogen gas did not interrupt the bone healing process. Enhanced bone formation was observed on the surface of WZ21 specimens, implying the osteogenic properties of the biomaterial. When compared to ZX50, WZ21 was able to combine both good mechanical integrity and biocompatibility with an appropriate corrosion rate suitable for bone regeneration [118] .
The developments of Mg-based biomaterials with a controlled biodegradation profile and bone-mimicking mechanical properties have encouraged surgeons to consider the prospects of biodegradable implants for clinical applications. Currently, there are three Mg-based bone screws available in the market made from the alloys Mg-Y-RE-Zr [119] and Mg-Ca-Zn [120] as well as pure Mg [121] .
Clinical trials on bone screws made from the Mg-Y-RE-Zr alloy have shown therapeutic results that are comparable to titanium screws in terms of the American Orthopedic Foot and Ankle Score (AOFAS) for hallux, the range of motion for the metatarsophalangeal joint, and pain assessment for hallux valgus osteotomies [119] . In addition, the screws made from Mg-Y-RE-Zr showed high degrees of osseointegration after 6 months, with no revision surgeries required. In a different study, Mg-Ca-Zn screws successfully fixed the distal radius fractures of 53 patients [120] . Acceleration of bone healing was reported due to the accumulation of Ca on the screws during biodegradation. Complete bone regeneration was achieved within one year along with the biodegradation of entire screws ( Fig. 7 b). Furthermore, pure Mg screws were utilized for the treatment of 23 patients with osteonecrosis in the femoral head [121] . The Mg screws demonstrated good fixation during biodegradation and promoted osteogenesis.
Despite the positive clinical progress of bulk Mg screws, the utilization of the AM technology for magnesium alloys is challenging due to the intrinsic properties of magnesium that has a high oxygen affinity, a low boiling point, and a high vapor pressure. Indeed, there is only one comprehensive study on the in vitro biodegradation behavior and the evolution of the mechanical properties of the laser-based AM porous WE43 magnesium alloy during in vitro biodegradation [29] . The porous AM WE43 Mg alloy, fabricated using SLM, was designed to have a topologically ordered open-cell structure for enhanced interactions with the surrounding cells and tissue. The AM WE43 Mg alloy was able to preserve ~80% of its volume ( Fig. 8 a) and maintain its mechanical properties at a level of the mechanical properties of trabecular bone for up to four weeks of in vitro biodegradation in a revised simulated body fluid ( Fig. 8 b-c) [29] .
Utilizing adhesive-based AM processes for Mg-based biomaterials is challenging as well, as the polymer binder and solvent constituents have to be compatible with magnesium. Recent developments of solvent capillary-driven techniques using binder jetting have helped in minimizing the metallurgical complexities involved in the AM of Mg alloys [122] . The framework involves only a solvent to interact with the outermost layer of magnesium powder particles ( i.e. , MgO), forming a strong interparticle capillary-bond that assembles powder particles into a 3D structure ( Fig. 8 d-e). In the absence of polymer, prior to post-AM sintering, a debinding step is not required and the solvent will decompose during the sintering process. The AM process utilizing the binderless capillarybond approach has been shown to result in a Mg-Zn-Zr alloy with 1.3 g/cm 3 density, 27% apparent porosity, and 18.4 GPa of elastic modulus mimicking the human cortical bone ( Fig. 8 f) [123] .
In addition, several studies on in situ multi-material AM of Mgbased biomaterials have been conducted using the SLM technology with blended powders to achieve high relative density Mg alloys, controlled biodegradation rates, improved strengths, and antibacterial properties. In situ AM of binary Mg-Zn alloys with variations in the composition of the Zn powder was performed to understand the mechanical characteristics of the resulting materials [67] . Interestingly, only the Mg-1Zn alloy could reach a high relative density of 99.35% with almost no defects. As the Zn content increased from 2 to 12 wt%, the Mg-Zn alloys suffered from solidification cracking due to the presence of solidification shrinkage stresses that tore the liquid films composed of eutectic liquid phases. Among the in situ SLM binary Mg-Zn alloys ( Fig. 9 a), the Mg-1Zn alloy exhibited the best mechanical properties with a hardness value of 50 HV, an ultimate tensile strength of 148 MPa, and an elongation value of 11%, which were similar to those of cast Mg-1Zn counterparts [67] .
Furthermore, a series of in situ SLM experiments with prealloyed ZK60 added with 1.8 − 5.4 wt% rare earth Nd powder demonstrated an optimum biodegradation rate of 1.56 mm/year, with the in situ formed alloy containing 3.6 wt% Nd [68] . The enhanced corrosion resistance of the ZK60-3.6Nd alloy was attributed to the formation of fine Mg-Zn-Nd intermetallic phases along the α-Mg grain boundaries ( Fig. 9 b), which created tight junctions that prevented the propagation of corrosion. With a higher Nd concentration ( i.e. , 5.4 wt%), however, more Mg-Zn-Nd intermetallic phases were formed and more sites of galvanic corrosion occurred, which counteracted the beneficial effects from grain refinement and increased the corrosion rate [68] . In another research, the in situ addition of 0.4 wt% Cu to ZK60 during SLM resulted in a ZK60-0.4Cu alloy whose degradation rate was similar to ZK60, ( i.e. , 1.01 mm/year) but exhibited antibacterial properties as well ( Fig. 9 c-d) [70] . The addition of Cu also resulted in an enhanced compressive strength up to 158.3 ± 5.1 MPa, due to uniformly distributed Mg-Zn-Cu phases and grain refinement strengthening.
In conclusion, several studies focusing on the biodegradable behavior and mechanical integrity of Mg-based biomaterials have been performed. The clinical trials of Mg-based bone screws have shifted the paradigm of corrosion resistance towards a new perspective of temporary bone fixation implants. Following the positive outcomes, the research on (multi-material) AM technologies for Mg-based biomaterials is now advancing. Even though a few research groups have successfully produced (multi-material) AM Mg alloys [67][68][69][70] , more studies are required to better understand and control the AM processes and the biodegradation profiles, mechanical integrity, and biocompatibility of these materials.

Fe-based biomaterials
Iron and its alloys have been studied for temporary loadbearing bone replacements. They combine biodegradable behavior with high mechanical strength and ductility [124] . As compared to Mg-based materials, Fe-based biomaterials have the advantage of not releasing hydrogen as they degrade. Their biodegradation products, being not completely dissolvable in physiological solutions, have been found to hinder the release of iron ions and slow down the biodegradation process [125] . In addition, the ferromagnetic nature of iron may need to be altered prior to using it for the fabrication of imaging-friendly implantable devices.
Various methods including alloying with Mn or noble metals and reinforcing with bioceramics through powder metallurgy techniques have been investigated to accelerate the biodegradation rate and reduce the magnetic properties of iron [125] . Among the alloying elements, manganese is considered promising, since it cannot only improve the biodegradation rate of iron [126][127][128] , but can also lead to anti-ferromagnetic Fe-Mn alloys [129] . A trace of manganese has been found to play an important role in osteogenesis and bone resorption [130] . Besides manganese, noble metals, such as Ag, Au, Pd, and Pt [131][132][133] have been used to create second phases in Fe-based alloys ( e.g. , Fe-Ag, Fe-Au, Fe-Pd and Fe-Pt), which can induce micro-galvanic coupling to stimulate Fe degradation.
Even though alloying pure iron can speed up its biodegradation in vitro , such improvements have not been observed in vivo . The non-invasive monitoring of pure iron and iron reinforced with 5 wt% bioceramics ( i.e. , HA), tricalcium phosphate (TCP), or biphasic calcium phosphate (BCP) implanted in sheep forelegs for 60 days has shown insignificant reductions in the sizes of these biomaterials [134 , 135] . Similarly, two Fe-based alloys, namely Fe-10Mn-1Pd and Fe-21Mn-0.7C-1Pd (in wt%), have displayed almost no structural changes after implantation in rat femurs for 52 weeks ( Fig. 10 a) [136] . Although signs of degradation were present, no remarkable weight reductions were observed ( Fig. 10 b). The inadequate biodegradation rate was explained by insufficient oxygen transport to the biomaterial surface, due to a dense layer of biodegradation products that wrapped around the surface and hindered further degradation of iron [136] . These findings suggest that Fe-based biomaterials should be designed to be highly porous so as to allow high permeability for accelerated biodegradation.
As for biocompatibility, in vivo studies on bulk Fe-based biomaterials have reported no systemic toxicity [124] . The Fe-10Mn-1Pd and Fe-21Mn-0.7C-1Pd alloys were found to show no signs of local toxicity or clinical abnormalities, when implanted transcortically in 38 rat femurs for 52 weeks [136] . Although Fe ions were present in the biomaterial's vicinity ( Fig. 10 c), no harm to the neighboring bone tissue was observed. The Fe-based biomaterials were well integrated and enveloped by a narrow flap of connective tissue. In addition, an in vivo study on pure iron and Fe-bioceramic composites including Fe-5HA, Fe-5TCP, and Fe-5BCP, have exhibited no systemic toxicity after 60 days of implantation in sheep forelegs [134] . Normal dynamic blood responses and no cellular stresses were observed throughout the in vivo study. Histological analysis determined the presence of inflammatory cells ( i.e. , macrophages, granular tissue, and fibrous tissue) surrounding the Fe-bioceramic composites, as the sign of active biodegradation and wound healing process ( Fig. 10 d) [134] .
On the other hand, an in vivo study on Fe-30 wt% Mn implanted in rat femurs reported that the biomaterial might act as a local irritant, although the finding was not statistically significant [137] . A small-scale necrotic bone was found to be engulfed in the corrosion products along with the presence of macrophages. bone, R -void after implant removal, -macrophages, -granular tissue and • -fibrous tissue (Reprinted from [134] © 2014 Wiley Periodicals, Inc.). (e) Bone remodeling occurs during the in vivo tests of Fe-Mn-Si alloys (periodic acid-Schiff stain). After two weeks, (i) a small bone necrosis region (black arrow, × 100) and (ii) empty lacunae (black arrow) with some osteocytes (white arrow) were found in bigger trabeculae ( × 900). After 4 weeks, (iii) new osteoid tissue and osteoblasts (black arrow) were present in the trabecular bone ( × 900) (Reprinted from [138] with permission from Elsevier). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Nonetheless, there were no adverse systemic effects determined [137] . Furthermore, the addition of silicon to Fe-Mn biomaterials has been found to improve their biocompatibility. In an in vivo study of a bulk Fe-Mn-Si alloy (with a ratio of 3:2:2) implanted in rat tibia for 28 days, the bone regeneration process was found to continue [138] . An increase in the ALP activity after 14 days of implantation suggested active bone formation. Moreover, histological analysis confirmed the start of bone remodeling after two weeks, when some apoptotic osteocytes were observed, followed by the appearance of osteoblasts at week 4 ( Fig. 10 e) [138] . Overall, a normal blood homeostasis was observed during the slow biodegradation of the Fe-Mn-Si alloy.
Recent developments of AM technologies have opened up new possibilities to improve the biodegradation profiles of Febased biomaterials through porous topological and multi-material designs. Ex situ binder-jetting of Fe-30 wt% Mn with an open porosity of 36.3% has been shown to result in a material with an electrochemical corrosion rate of 0.73 mm/year, which is ~11 times higher than that of pure iron with the same exposed surface area [86] . Moreover, the inclusion of 1 wt% calcium in the pre-alloyed binder-jetting of Fe-35 wt% Mn increased the biodegradation rate from 0.04 mm/year to 0.07 mm/year ( Fig. 11 a-b) [87] . Furthermore, AM topologically ordered porous pure iron with 80% porosity has demonstrated an electrochemical corrosion rate of 1.18 mm/year, which is ~12 times higher than that of cold-rolled iron, and has led to a lower impedance value as compared to its cold-rolled counterparts ( Fig. 11 c-e) [30] .
Although the electrochemical corrosion values of AM porous Fe-based biomaterials have been generally found to increase, validation through in vitro immersion tests is still required. After 4 weeks of static immersion in a revised simulated body fluid, porous AM pure iron achieved only 3.1% weight loss [30] . The small mass reduction was attributed to the dense corrosion products that covered the porous structure ( Fig. 11 f), preventing further biodegradation of porous iron from occurring. Similarly, in vitro immersion of binder-jetted Fe-30Mn resulted in a negligible weight loss 71 , 86] . These outcomes may also be due to an inadequate in vitro fluid flow, different atmosphere conditions from those occur in vivo , as well as the absence of blood cells and macrophages, which could have contributed to the faster biodegradation of biomaterials and the removal of the corrosion products. Obviously, performing a long-term in vivo biodegradation study on porous AM Fe-based biomaterials is of critical importance. However, no such studies have been performed to date. Recently, a 4-week in vivo study of SLM Fe-35 wt% Mn, having a 43% porosity, has shown an on-going new bone forming process with good bony-implant integration [71] .
In addition to an improved rate of electrochemical biodegradation, the progress in AM using multi-material has introduced other functionalities to Fe-based biomaterials including antiferromagnetic properties, improved osteosynthesis properties, and assistance in cortical bone cancer treatment. Ex situ binder-jetted Fe-30Mn [86] and Fe-35Mn-1Ca [87] have been shown to generate ε-martensite and γ -austenite Fe-Mn phases ( Fig. 12 a-b) during the post-AM sintering stage. The ε and γ -Fe-Mn phases possessed the intrinsic anti-ferromagnetic properties [139] . Furthermore, ex situ Fe-30CaSiO 3 (in wt%) composites fabricated using material extrusion not only improved the in vivo osteosynthesis significantly as compared to pure iron but also demonstrated the potential for bone cancer therapy ( Fig. 12 c). In the case of a Fe-30CaSiO 3 biomaterial synergized with laser and reactive oxygen species (ROS), the therapeutic effects of tumor treatment ( Fig. 12 d) were enhanced, as demonstrated in an in vivo study [89] .
Overall, Fe-based biomaterials have shown their potential as temporary load-bearing bone implants. Regardless of the systemic biocompatibility reported in all in vivo studies, Fe-based biomaterials may act as a local irritant, due to the toxicity of the corrosion products. Details on the interactions of Fe-based degradation products with bone tissue and inflammatory responses are still unclear. More research employing (multi-material) AM technologies for porous Fe-based biomaterials is required not only to improve the biodegradation profile, but also to understand the degradation behavior that strongly affects the biocompatibility of Fe-based biomaterials.

Concluding remarks and future perspectives
The initial effort s of the bone tissue engineering community were mainly focused on fabricating biomaterials that mimic the macrostructure of the natural bone. Recent effort s have been mobilized to produce biomaterials using the AM technologies that allow for the realization of micro-architectured porous geometries and the placement of the right biomaterial at the right place for bone regeneration. Given the possibility of high-precision manufacturing of complex macro-and micro-architectured porous biomaterials, the AM technologies are undoubtedly linked to the future of metallic biomaterials for bone implants.
Clearly, AM technologies intrinsically provide the potential for multi-material fabrication that have not yet been extensively explored for bone implant applications. Using multi-material AM technologies, the functional requirements of biomaterials for bone implants, such as long-term mechanical properties for permanent use, short-term mechanical integrity, and biodegradation for temporary replacements, as well as biocompatibility, can be tailored not only through structural design but also via adjustment of material types or compositions.
Among the available multi-material AM technologies, extrusionbased AM, using multi-material powder-based feedstock, appears to be the most straightforward option due to its capability of fabricating complex multi-material interfaces with a simple manufacturing process. For the multi-material extrusion-based AM to be successful in fabricating metallic bone substitutes, the fabrication processes, implant design, and material choice need to be properly selected. Understanding the chemistry of multiple materials in the feedstock, which may or may not affect the AM process, structure, and material properties, is essential. In addition, ensuring a homogenous distribution of multiple materials in the feedstock prior to AM and in the fabricated biomaterials is equally important in order to achieve near-isotropic material properties. Together with appropriate AM processing parameters, the multi-material feedstock with shear-thinning behavior and free-standing characteristics should allow continuous, stable deposition of structures even when the aspect ratios are high. Post-AM debinding and sintering need to be adjusted with respect to temperature, time, and atmosphere in order to achieve specific microstructures and biomaterial functionalities. Even though the multi-material extrusion-based technology has been so far primarily applied to Fe-based bone-substituting [89] , it can, in principle, be applied to a wide range of powdered feedstock materials for various multi-functionalities, including metallic biomaterials, such as tantalum-based alloys and shape memory-based alloys The future research on the multi-material extrusion-based AM for metallic biomaterials should be directed towards enhanced abilities to introduce materials of different types or compositions locally at a few micrometer scale within the structure. This will advance the technology towards improved control over the resulting implant properties. Additionally, fabricating geometrically complex multi-functional biomaterials should involve easily dissolvable or decomposable sacrificial support materials that will pose no adverse effects on the performance and biocompatibility of the resultant biomaterials. Finally, combining multi-material extrusionbased AM with two-dimensional nanopatterning on each layer [140][141][142] during the fabrication process could further improve the functionalities of the final biomaterials through the optimization of surface nanotopography. Overall, multi-material extrusion-based AM technologies hold a great promise for advancing the state of the art in the fabrication of metallic multi-functional bone substitutes.
In short, substantial progress has been made in developing multi-functional Ti-, Mg-, and Fe-based biomaterials aimed for applications as orthopedic implants. AM Ti-based implants have been applied clinically, while the research is continuing on Mgand Fe-based biomaterials. Multi-material AM of Ti-based biomaterials including alloys other than Ti-6Al-4 V promises significant improvements in the microstructure, mechanical characteristics, and biological performance of the resulting materials including antibacterial properties, anti-inflammatory responses, and the upregulation of angiogenesis-related genes. For Mg-based biomaterials, multi-material AM has addressed the issue of decreasing the biodegradation rate of these alloys, while improving their mechanical strength and inducing antibacterial properties. Multimaterial AM for Fe-based biomaterials aims at achieving faster biodegradation rates, altering the ferromagnetic properties of such materials and enhancing their biocompatibility. In conclusion, multi-material AM technologies are expected to enable the proper selection of materials and compositions to improve the functionalities of Ti-, Mg-, and Fe-based bone substitutes.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.