Design and manufacturing by fused filament technique of novel YSZ porous grafts infiltrated with PCL/PVA/AgNPs for large bone defects repairing

Journal of Materials Research and Technology 29 (2024) 3393–3408 Available online 12 February 2024 2238-7854/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Design and manufacturing by fused filament technique of novel YSZ porous grafts infiltrated with PCL/PVA/AgNPs for large bone defects repairing Gemma Herranz a , * , Javier Hidalgo a , Victoria Axelrad a , Ernesto J. Delgado-Pujol b , Cristina Berges a , Juan Alfonso Naranjo a , Juan Pinilla d , e , Bel´ en Begines c , Ana Alcudia c , Yadir Torres b a Universidad de Castilla–La Mancha, INEI-Escuela T´ ecnica Superior de Ingeniería Industrial, Av. Camilo Jos´ e Cela S/n, 13071, Ciudad Real, Spain b Departamento de Ingeniería y Ciencia de los Materiales y del Transporte, Escuela Polit´ ecnica Superior, Universidad de Sevilla, 41011, Seville, Spain c Departamento de Química Org´ anica y Farmac´ eutica, Facultad de Farmacia, Universidad de Sevilla, 41012, Seville, Spain d Departamento de Traumatología y Cirugía Ortop´ edica, Trauma and Sports Medicine, Hospital General Universitario de Ciudad Real, C. Obispo Rafael Torija, S/n, 13005, Ciudad Real, Spain e Universidad de Castilla–La Mancha, Facultad de Medicina, Edificio Polivalente, Cam. Moledores, S/n, 13071, Ciudad Real, Spain ARTICLE INFO Handling Editor: P⋅Y. Chen Keywords: Fused filament fabrication Additive manufacturing Porous grafts YSZ nanopowder Biopolymer PVA/PCL Antimicrobial activity ABSTRACT Addressing large bone segmental defects remains a profound challenge in orthopaedic surgery, demanding graft designs that mirror patient-specific defects while ensuring vascularization, osteoconduction, and structural integrity. Conventional approaches fall short in achieving this level of customization and functionality. This work leverages fuse filament fabrication (FFF) to fabricate innovative yttria stabilize zirconia graft designs that combine several strategies to tackle these multifaceted challenges. Graft designs ingeniously incorporate dense and porous structures mimicking the intricate architecture of natural bone for improved structural integrity and different ridges and protrusions to enhance the fixation. Additionally, these grafts undergo infiltration with PCL/ PVA/AgNPs composite polymers, enriching biocompatibility, structural integrity, and antibacterial properties, fostering conducive conditions for optimal bone growth. Attaining printable filaments with YSZ powders below 50 nm and loads reaching 50 vol% marks a pioneering advancement in bone graft fabrication via Fused Filament Fabrication (FFF). This enabled high densification and low grain size at reduced sintering temperatures, imperative to achieve high mechanical performance. The biodegradation capacity of biopolymers in PBS and their role in the mechanical behaviour of grafts have been assessed, demonstrating the beneficial potential of the multicomponent graft proposed reaching strength values of 90–105 MPa. 1. Introduction Bone defects in which a large part of the bone is damaged are among the biggest challenges in orthopaedics surgery and traumatology. These lesions are very common but have a complex solution since results after the application of the available treatments depend directly on the lesion and patient idiosyncrasy. The injured bone is commonly replaced by a graft with interconnected porosity that acts as a guide during the patient’s regeneration of bone. The gold standard is to use a graft made of human bone, in which autologous grafts are preferred due to the high compatibility. However, this solution is only adequate for small lesions, generates discomfort in the donor zone and commonly results in comorbidities [1,2]. For larger injuries, heterologous bone grafts from deceased donors are a popular option, in which the donor’s bone is carved to adapt it to the lesion geometry. The heterologous bone graft is not exempt from complications and in multiple cases, the graft does not integrate with the receptor bone and the risk of infections and transmission of diseases increases [3]. Besides, the patient’s immobilization and unloading times are generally high to favour graft integration and access to donor banks can be limited considering that more than 2 million patients worldwide need bone grafting every year [4]. This urgently calls for synthetic alternatives which should be capable of mimicking the structural properties of the bone concerning bioactivity, fraction, size and topological distribution of porosity, interconnectivity and mechanical strength [5]. The research efforts seeking synthetic alternatives have resulted in a * Corresponding author. E-mail address: [email protected] (G. Herranz). Contents lists available at ScienceDirect Journal of Materials Research and Technology journal homepage: www.elsevier.com/locate/jmrt https://doi.org/10.1016/j.jmrt.2024.02.057 Received 16 November 2023; Received in revised form 25 January 2024; Accepted 6 February 2024 Journal of Materials Research and Technology 29 (2024) 3393–3408 3394 bunch of materials. Among them, the most commonly used materials include third-generation bioactive and resorbable polymers and ceramics with excellent osseointegration and osteoconductive properties [6–12]. Unfortunately, these materials cannot withstand high loads, which limits their use for critical-size skeletal defects and commonly entails the use of auxiliary plates to keep the integrity and position of the graft. Removing auxiliary plates requires additional surgery when the lesion is healed. Hence, orthopedical surgeons await a solution that eliminates this need. High-strength ceramics such as high-purity alumina [13], Zirconia Toughened Alumina (ZTA) [14] and Yttria Stabilized Zirconia (YSZ) [15,16] can meet the graft’s mechanical strength requirements and are biocompatible as well as non-cytotoxic. These structural ceramics are bioinert [17] and typically require the application of bioactive/biodegradable coatings to promote the adhesion of new bone to the graft walls. The open porous architecture and surface features play a crucial role in the graft’s structural and functional performance and thus are crucial parameters to be considered during its design. The interconnectivity of pores with a suitable size, density and distribution is critical to enable bone tissue ingrowth and the distribution of nutrients and metabolites [18]. The micro and macrostructures will be equally determinant in the graft’s mechanical performance. The apparent elastic modulus of the graft can be tailored by modifying the density and pore topology [19], which can be used to match the bone elastic modulus. However, the increase in the relative volume covered by pores, which typically entails a decrease in the mechanical properties, might conflict with biological requirements. Thus, a compromise between the functional and biomechanical properties should be considered during the design. Various manufacturing processes, such as polymeric sponge replication and salt leaching, have been utilized to produce ceramic scaffold grafts that mimic the porous structure of trabecular bone [20–22]. However, controlling the replicability of these methods and achieving the desired mechanical properties of the obtained scaffold grafts remains challenging. Additive manufacturing techniques have opened up new possibilities for designing and fabricating bone scaffold grafts. Additive manufacturing enables engineering open porosity architectures to enhance biological integration and optimize mechanical performance. It also allows customized designs by capturing defect shapes through tomographic scanning and translating them into 3D printable files. Direct methods like laser-based sintering or melting of ceramic powders face difficulties due to low absorbance, high melting temperature, high melt viscosity, and poor thermal shock resistance [23–26]. Indirect methods involving binder usage are preferred for additive manufacturing ceramic components. Direct ink writing (DIW) [15,27, 28] and stereolithography (SLA) [29–31] are popular 3D printing techniques for ceramic graft fabrication, but they come with challenges like powder dispersion, ink formulation, shape retention, and limited variations of raw ceramic powders. Binder jetting (BJ) offers an alternative, requiring a low binder percentage but resulting in low integrity of as-printed green parts and undesired internal porosity [32–36]. Fused filament fabrication (FFF) emerged as a promising additive manufacturing alternative to produce ceramic components [37–42]. FFF stands out for its simplicity and the low equipment investment required compared to other technologies. Besides, FFF benefits from the ample existing knowledge of the well-established ceramic injection moulding (CIM) technology, which shares most of the process steps with FFF and the feedstocks can be implemented to both technologies [43–47]. The principles of FFF are similar to those of DIW, though instead of a viscoelastic paste, a thermoplastic filament is extruded to produce the desired geometry. This has the advantage of providing better shape retention and strength to the printed green specimen. The mixing of the binder with the powder before printing results in better powder particle packing and thus, sintered ceramic parts with high densification and mechanical properties can be achieved. Faulty printed parts can be straightforwardly recycled into new raw materials. This is also an interesting benefit of FFF compared to DIW, SLA, BJ and other direct printing technologies, exhibiting complications of reusing the printing waste or unsuccessful printings. The main difficulty of FFF technology is producing filaments with a high fraction of ceramic powder that has adequate properties for printing, i.e., sufficient flexibility to be spooled, adequate rigidity to avoid buckling during printing and good fluidity when melted. This is critical when particle sizes fall below the micrometre range since it significantly diminishes the fluidity of the feedstock and the permissible powder load [43]. The use of low ceramic powder loads generates more shrinkage, makes dimensional control difficult and can compromise the structural integrity of the bone implant (brown part). On the other hand, nanometric powders enhance sintering mechanisms, leading to the production of highly densified parts compared to that attained with larger powder sizes at equal loads [44,45]. This high densification can be attained maintaining a fine grain size, which is crucial to meet the standards of grain sizes below 400 nm [46]. Low-temperature-degradation (LTD) of YSZ and subsequent implant ageing has been proven to decrease closely with decreasing grain size [47] which also calls for minimizing the powder size to the limits allowed by the FFF technology. Recent works on FFF of YSZ used specially designed powders with an average particle size of 100–300 nm and binder systems enabling printable filaments with 45–50 vol% solid loading [40,41,48,49]. These particle sizes are within the limit to both, meet the standard after the grain growth associated with sintering and effectively reduce LTD, the latter being particularly challenging to address in 3% mol. YSZ. YSZ powder with particle sizes as low as 40 nm is commercially available, which to the best of the authors’ knowledge has not been used in this technology to date. Producing functional highly-loaded filaments with such low particle size powder will be a breakthrough in FFF allowing processing of ceramic grafts or other technical applications with unprecedented mechanical performances. Furthermore, this achievement would also mark a significant breakthrough in the processing of zirconia itself and its application in bone repair. The reason for this is that the immaturity of zirconia processing techniques has been a significant hindrance to the commercial viability of zirconia scaffolds for load-bearing applications [50]. Besides, addressing potential complications associated with implantation procedures, particularly the formation of pathogen biofilms, is also pivotal in reducing the probability of implant failure [51]. In this regard, incorporating antimicrobial properties into the implant surface becomes a cornerstone of implant development. In this sense, several strategies can be employed, including passive and active techniques, to achieve this goal [52,53]. Passive methods involve modifying the surface texture and properties of the implant to inhibit initial bacterial adhesion and subsequent proliferation [54]. However, these methods have limitations in completely preventing bacterial colonization, and this is where active techniques, such as antimicrobial coatings are more convenient [55]. Also, silver ions (Ag + ), widely recognized for their antimicrobial efficacy against bacteria, fungi, and viruses, are commonly integrated into these coatings [56,57]. The characteristic high surface-volume ratio of silver nanoparticles (AgNPs) enhances their antimicrobial properties, making them a promising option for implant applications [58–60]. Additionally, polymeric matrices have proven effective as smart drug delivery systems, with polyε -caprolactone (PCL) standing out for its slow biodegradation rate, biocompatibility, and favourable mechanical properties [61]. Nonetheless, PCL’s hydrophobic nature poses challenges to its attachment to natural bone surfaces. Blending PCL with hydrophilic and non-toxic polyvinyl alcohol (PVA) addresses this limitation, as PVA exhibits excellent mechanical behaviour and unique degradation rates compared to PCL [62]. The use of PCL/PVA-based biodegradable biocomposites filled with silver nanoparticles presents a compelling approach to address the limitations and harness the benefits of these materials for YSZ grafts, thereby offering potential advantages for implant applications by a multifunctional approach, integrating antimicrobial activity, tailored mechanical properties, apart from conveniently controlled drug release. In addition, the G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3395 degradation balance of the composite would allow the replacement of the degraded polymer by osteoblasts, leading to the bone tissue to colonize the inner part of the grafts. This work investigates and discusses the use of FFF technology for the manufacturing of customized medical-grade YSZ grafts. Two innovative designs are presented, with protrusions and cavities to facilitate scaffold fixation to the damaged bone, preventing dislocation during consolidation. Pore distribution patterns based on the bone cortical and trabecular structures are proposed, with a solid outer shell to provide structural support and a porous core allowing bone ingrowth and biofunctionalization via biopolymer infiltration for better patients’ recovery. For this purpose, the development of a suitable filament with a high load of ~40 nm YSZ powder is addressed and the effect of different processing conditions on the filament features is investigated. The graft’s dimensional control and microstructure depending on the printing and sintering conditions are assessed and adequate conditions are selected based on these studies. Some processed grafts are then infiltrated with a variable ratio (80:20, 50:50) of PCL and PVA, combined with silver nanoparticles. This allows for modulating biodegradation kinetics and thus the release of the bactericidal agent depending on the patient’s needs. Mechanical properties of the designed grafts with and without polymeric infiltration are tested, and polymer degradation studies in phosphate-buffered saline (PBS) solution are conducted to compile data to evaluate the performance for the best design of a large bone-repairing device fabricated with FFF techniques. 2. Materials and methods Fig. 1 shows the summary scheme of the experimental design proposed in the present investigation, while the following sections develop the most important aspects related to the manufacture and characterization of the proposed bone implants. 2.1. Starting ceramic powder The YSZ powder was supplied by Tosoh Corporation (Japan) under the commercial name of TZ-3Y-E. Other YSZ grades, such as the TZ-3YS- E powder have been specifically conditioned by Tosho to facilitate ceramic injection moulding and equivalent processes as FFF, but exhibiting a higher particle size. According to the supplier, TZ-3Y-E is mainly composed of ZrO 2 with a 5.2 wt% (3 mol.%) of Y 2 O 3 and ~0.4 wt % Al 2 O 3 and has a d 50 of 40 nm. In this work, TZ-3Y-E has been selected because it is expected to exhibit a different sintering response and to yield sintered parts with a fine crystal grain structure that could exhibit improved strength, fracture toughness, wear resistance and LTD. The raw powder density was measured by a Micromeritics Accupyc II 1340 helium pycnometer. The size, morphology and phases of the ceramic powder are obtained using a JEOL JSM-6610LV scanning electron microscope and Philips X’pert Pro diffractometer, respectively. In this context, the measures were done at a potential of 40 kV and an intensity of 40 mA, in a 2θ configuration, with a secondary monochromator, Cu K α radiation (λ =1.5418 Å) and PIXcel solid state detector with an active length in 2θ of 3.347◦. The 2θ range covered is 20◦–70◦with a step size of 0.026◦and a step time of 600 ms. The Full Profile Search Match software [63] based on the Crystallography Open Database (COD) and Rietveld method for fitting XRD diffractograms was used for XRD peak indexing, phase fraction determination and crystal domine size calculation from the resulting diffractograms. 2.2. Manufacture and characterization of filaments for 3D printing The binder system is based on previous authors’ formulations [64, 65] and the studies concerning YSZ previously published in the literature [40]. This novel binder formulation was specifically developed to enable the use of nanometric YSZ powders in filaments with a balance of properties including enough flexibility and toughness to be spooled without breaking, but with enough rigidity to avoid buckling during printing. The binder is mainly composed of 22 vol% of a polypropylene (mp 112–140 ◦C ASTM D 87), 63 vol% of a thermoplastic elastomer (27.34 g/10 min melt index, 2 kg, 190 ◦C), 10 vol% of wax (mp 53–58 ◦C ASTM D 87) and 5 vol% of stearic acid; are additionally added to improve the feedstock’s rheology and act as surfactants. Feedstocks were produced by a Haake Rheocord 252p rheometer Fig. 1. Flow diagram of the manufacturing and characterization protocols addressed in this work. G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3396 with roller blades, which enabled monitoring of the torque required for mixing to assess the process evolution and the feedstock fluidity. The mixing temperature and speed were set to 160 ◦C and 40 rpm, respectively. A Dynisco LCR 7000 was used to extrude a filament from the feedstocks produced in the former step previously ground to 2–5 mm size pellets. The extrusion temperature and speed, and the capillary diameter were adjusted to produce filaments with a diameter close to the 1.75 mm recommended for the printer. The YSZ solid loading was varied up to 50 vol% of powder, equivalent to 86.1 wt% according to powder density (5.788 g/cm 3 ) and feedstock density (3.361 g/cm 3 ). This solid loading resulted in a suitable loading to maintain adequate filament properties to be printable. During the extrusion of the filament, different temperatures (115–130 ◦C), extrusion speeds (10–100 mm/s) and capillary diameters (1.6–1.8 mm) were tested until a filament with adequate characteristics to be printable reached. 2.3. Fabrication using FFF and characterization of ceramic grafts with simple geometries Different specimen designs were printed by an FFF printer (3D Prusa Steel Black Edition Mark II) with a hardened steel nozzle of 0.6 mm diameter, renewed if deviations in diameter were detected post-printing, employing 0.6 mm drill bits as precision tool. The designs were first done in Solid Works CAD software and exported to IdeaMaker software for the definition of the printing protocol and conditions. For the validation of the filaments, hexagons circumscribed in a 12 mm circumference, with 4.5 mm thickness and with 100% and 50% infill were first printed (solid and porous hexagons, respectively). These hexagons were also used to study the effect of the printing conditions and the process parameters on the defects generation and dimensional control of the parts from printing to sintering. The hexagonal geometry enabled a straightforward method of detecting any shape distortion during the printing process. Besides, in the context of FFF (Fused Filament Fabrication), the maximum achievable thickness is limited by the necessity of a flawless debinding step. Given that implant designs involve solid parts, employing solid hexagonal structures allows us to effectively test these limitations within the printing process. Once the process parameters were adjusted, the innovative graft designs described later (see section 2.4) were manufactured. A grid-type infill pattern with alternating 45◦ and 135◦printing direction layers was selected in all cases. An iterative process was carried out to adjust the printing parameters. The printing bed temperature of 80 ◦C and a layer height of 0.2 mm were common in all printing experiments. A minimum nozzle diameter of 0.6 mm was attainable with the processed filament. A low nozzle diameter is typically desired for increasing the printing resolution. However, decreasing the nozzle diameter below 0.6 mm required high forces during extrusion due to the high feedstock viscosity, as anticipated in the previous section. This resulted in buckling and a discontinuous flow, which could not be corrected by increasing the printing temperature to provide the feedstock with more fluidity. The printing temperature was varied in the 200 ◦C–300 ◦C range. Temperatures below 240 ◦C do not enable enough feedstock fluidity for proper printing. Temperatures above 240 ◦C resulted in high dilation of the feedstock at the nozzle exit producing poor printing results and low shape retention of the hexagons. Hence, 240 ◦C printing was chosen. Finally, the flow rate and printing speed were systematically varied until flawless specimens, with good shape retention and layer adhesion, were achieved. The best results were obtained with a 105 % printer feeding rate and 10 mm/s printing speed. Thereafter, the printed specimens were subjected to a thermal cycle in an air atmosphere to eliminate the binder system. The thermal debinding cycles were set in a Hobersal furnace (12PR450/SCH PAD P). This cycle was designed and adjusted based on the thermogravimetry analysis (TGA) performed in LABSYS evo TGA equipment (see Fig. 2). Brown parts were sintered in a tubular furnace Hobersal furnace (ST 186030) under an air atmosphere. The sintering cycles were carried out at a heating rate of 2 ◦C/min and sintering temperatures ranging from 1100 ◦C to 1500 ◦C were evaluated to have a record of the densification and dimensional shrinkage of the specimens for process design. Sintering time was set at 1 h for all the sintering temperatures and cooling was done freely at the furnace-cooling rate. 2.4. Designs and fabrication of ceramic implants (“tailor-made”) The geometrical features of the two innovative grafts proposed in this study are represented in a summary scheme (see Fig. 1). The first conceptual incorporation to the existing graft designs consists of adding a solid outer shell to provide better structural stiffness to the current fully porous designs, which only focused on providing a graft for the growth of healthy bone and lack of fixation zones. These grafts are intended to connect two sections of a long bone such as the femur, and thus as a first approximation, a cylinder with an ellipsoidal base shape is considered. Two types of anchoring zones are proposed and thus added to the base cylindrical model. These designs, in combination with the use of YSZ, are planned to improve the structural integrity and get rid of the use of a temporary plate fixation that needs additional surgery to be removed when the lesion is healed [66]. Design 1 considers a peripherical joint with the outer area of the cortical bone, and thus stretched protrusions are radially added to the outer shell of the graft at both bases, substituting the grid pattern of the middle part. The cortical part will be embedded by the graft shell and protrusions will attach to the bone providing rotational stiffness to the system. This design can be adequate in bone fractures exhibiting high fragmentation. Fig. 2. (a) TGA curves of the YSZ filament, and (b) The thermal debinding cycle. Air atmosphere was used in both processes. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3397 In Design 2, the porous zone of the graft is extended from both bases and devoid of the outer shell. These two branches will be introduced into the bone marrow, and thus the cortical part will directly stand over the outer shell of the graft. In this case, the cortical bone provides rotational stiffness. This design might be particularly effective in tumorous bone extractions resulting in clean and flat surface bone sectioning. This anchoring zone is designed to promote osseointegration and thus enhance biofunctionality. However, the mechanical properties of this porous anchoring zone may be inferior to the anchoring zone of Design 1. The mechanical behaviour of the two “tailor-made” designs strongly depends on their porous zone (Design 1) as well as on the role of the outer shell (Design 2). In this context, compression tests using simple geometries with these porous and dense outer shell architectures allow us to evaluate the influence of the particularities of both designs on mechanical strength. Cylindrical specimens of equal dimensions were fabricated with the same processing conditions described before, the one associated with Design 1 consisting of a dense outer shell and the other (associated with Design 2) consisting only of the 50% rectilinear printing pattern. Both cylinders were mechanically evaluated with a uniaxial compression test. The dimensions of these two preliminary prototypes (Fig. 3) do not correspond with a human femur. In the experimental approach of this work to assess biocompatibility, it is foreseen to implant the grafts in small rodents to subsequently transfer the designs to the needs of larger mammals until reaching the measurements of a human. For the porous zone, the rectilinear grid pattern used in porous hexagons is maintained. The size and fraction of the porous grid are conceptualized to enable good vascularization and osteoconductivity and are based on established practices in graft design [67–69]. These are preliminary designs to show a probe of the concept. In future designs, the complex and exact shape of the bone can be straightforwardly reproduced by 3D printing and extrapolated from these preliminary models when a tomography scan is available. The combination of solid and porous fractions can be tailored in the printed graft to better reproduce that of the two connecting parts of the damaged bone. Finally, before printing, the dimensions of the CAD design were modified, considering the preliminary shrinkage results in hexagons, to match the dimensions which represent the expected dimensions after the sintering stage. 2.5. Synthesis and infiltration of the composite biopolymers (PVA/PCL- AgNP) in porous implants All chemical products were used directly from suppliers without further purification. Polyethylene glycol 3000 (PEG 3000), polyethylene glycol 200 (PEG 200), polyvinyl alcohol (PVA, Mn =30.000–70.000 g/ mol, 87–90% hydrolysed), polyε -caprolactone (PCL, average Mn 80,000 g/mol), silver nitrate (AgNO 3 ) and phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich (Madrid, Spain). Dichloromethane was acquired from Honeywell in Badalona, Spain. On the one hand, Madhavan et al. [70] methodology for AgNPs synthesis was employed. As a result, nanoparticles with high circularity and average size of 20 nm were obtained [60]. During the synthesis process, a magnetic stirrer (MR Hei-Standard magnetic stirrer from Heidolph Instruments, Germany) was used to mix 10 ml of PEG-200 and 3.42 g of PEG-3000 at 353 K. Once dissolved, the temperature was lowered to 303 K and 100 mg of AgNO 3 were added. Magnetic stirring was maintained throughout the process. On the other hand, the PCL/PVA composite was prepared in different composition relations, specifically 80/20 and 50/50. First, the PVA was dissolved in 120 ml of dichloromethane at room temperature. Subsequently, the PCL was added to the mixture; and finally, the AgNPs were poured into the blend in a relation of 0.4% w/w. Thereafter, the solution was heated to 303 K to remove dichloromethane until reaching a weight reduction of 70% relative to the original weight. Finally, for sample infiltration, a 5 mL syringe was used to pour the composite polymer through the 3D graft. Subsequently, the sample was left to dry in an oven at 300 K; then the infiltrated sample was frozen and lyophilised. 2.6. Biofunctional and biomechanical behaviour of ceramic grafts The polymer water uptake study was conducted in a PBS solution for 8 weeks at room temperature. Samples of the different composition relations of the polymer with a volume of 0.628 cm 3 were prepared and weighted using a balance Mettler AE 100 (Mettler Toledo, Columbus, United States) with an error of ±0.0001 g. Subsequently, test specimens were immersed individually in a flask with 20 mL of PBS dissolution. Swelling measures were made at determined intervals, every 30 min during the first hour. Eventually, the measurement frequency was reduced to one every hour until it reached 6 h, where the mass seemed to oscillate around a certain weight for each sample. At this point, the measurements began to be carried out daily for a week. Afterwards, the frequency was decreased to one per week until the trial ended in the eighth week. To carry out this process, each sample was retired from the PBS solution and the liquid was blotted on the surface with a filter paper. The water absorption index was calculated from the difference of the initial weight and the post-swelling weight, concerning the wet weight. The degradation study was conducted over the ceramic 3D printed graft infiltrated with the polymeric composite PCL/PVA variant (50/50 & 80/20). The samples were individually immersed in a flask with 40 ml of PBS dissolution for 8 weeks. An Heidolph Unimax 1010 with an Heidolph Incubator 1000 (Heidolph Instruments GmbH & Co. KG, Schwabach, Germany) attached was used to maintain the samples under constant agitation of 120 rpm and 37 ◦C. Measurements of the weight of the samples were made weekly employing a Mettler AE 100 (Mettler Toledo, Columbus, United States). To do this, the samples were retrieved from PBS and dried for 12 h at 37 ◦C. In addition, surface morphology changes were evaluated by SEM images, which allowed an analysis of the evolution of the pore size using the ImageJ software. A chart of the mass loss curve was obtained by processing the recorded mass data. The polymer mass value was obtained from the difference between the infiltrated substrate and the initial weight of the sample. Additionally, this algorithm was repeated for each weight recorded weekly during the degradation period, thus obtaining the polymer mass after degradation. In the meantime, the relative weight loss was obtained using the initial mass of the infiltrated polymer and compared to the masses of the polymer after degradation. In this work, the mechanical tests are carried out by adapting the ASTM C1424-15 and ISO 13314: 2011 standards. The uniaxial compression tests of the ceramic implants with and without sample infiltration were performed using a Shimadzu AG-IS Autograph Series. Fig. 3. Technical drawing of the design 1 and design 2 implants with dimensions in mm. G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3398 Compression was carried out along the longitudinal axis of the sample with a load control of 100 N/s (to avoid static fatigue phenomena). The force was applied until the total failure of the sample to evaluate the mechanical properties that the grafts could have due to the complexity of the three-dimensional shape. The plotted data was standardized to stress-strain curves according to the area of the test specimen. Finally, a detailed study of the relationship between the applied mechanical stress and the structural integrity of the porous graft is carried out. 3. Results and discussion 3.1. Development and validation of YSZ filaments Fig. 4a shows the scanning electron microscopy image of the raw powder. The micrographs clearly show that the raw powder is agglomerated in spheres, whose characteristic size agrees with the parameters of D 10 =1.6 μ m, D 50 =20.5 μ m and D 90 =50.0 μ m. Further magnification of these agglomerates by SEM exposes the nanometric nature of the powder particles, which exhibit a particle size of approximately 40 nm as stated by the supplier. However, it is worth noting that particle sizes below the micrometre range can significantly decrease the fluidity of the feedstock, leading to a reduction in the admissible powder load. Additionally, low solid loadings may result in inadequate densification and high shrinkage during the sintering process, thereby hindering dimensional control. In this context, the presence of powder agglomerates could be beneficial for developing feedstocks used in filament production for FFF. These agglomerates, with their specific size and shape, mimic the ideal characteristics found in powder injection moulding (PIM) powders [43], which contribute to achieving favourable rheological properties in feedstocks. These ideal characteristics can be extended to FFF, using analogous feedstocks and process steps [38]. In Fig. 4b, the resulting diffractogram of the YSZ powder is presented, revealing the presence of cubic (~71%) and monoclinic (~29%) ZrO 2 as the main phases in the base powder. These findings are in line with expectations, as the tetragonal phase is anticipated to be stabilized with the inclusion of 3 mol% Y 2 O 3 during the sintering process. Indeed, certain minor peaks observed in the diffractogram may correspond to Y 2 O 3 present in the powder. The density of the ceramic powder measured with the Helium pycnometer is 5.788 ±0.00 g/cm 3 . Fig. 5 shows the evolution of the torque values during the mixing for a 50 vol% solid loading. The stepwise addition of components produces an increase in torque until a maximum is reached when all the components are inside the mixing chamber. From this point, the torque decreases as the powder is incorporated into the binder system, reducing the friction between particles. The powder incorporation and segregation from the binder system reaches an equilibrium state in which the feedstock can be considered macroscopically homogeneous. The 13 ±6 N m stabilization torque is high but in accordance with other equivalent loading YSZ (100–300 nm particle size) feedstock formulations [41]. This suggests that the agglomerates might be broken during the mixing process, as confirmed by SEM analysis of the feedstock by backscattered electron contrast, showing the homogeneous distribution of the mixture and the absence of micrometric agglomerates (Fig. 5b) and the presence of the ceramic in the form of nanometric particles at higher magnifications (Fig. 5c). Fig. 5a also shows the feedstock after opening the mixing chamber after the mixing process. The appearance is homogeneous, there are no signs of an excess of powder and the feedstock exhibits certain fluidity. In essence, 50 vol% is the selected solid loading even though the torque is relatively high because the stabilization and homogeneity of the feedstock were evidenced. The next step is the evaluation and adjustment of the filament extrusion conditions, in which different combinations of parameters were tested. Combinations of high speed (>150 mm/s) and high temperatures (>140 ◦C) resulted in thin filaments (<1.65 mm) far from the 1.75 mm thickness objective. This could not be straightforwardly corrected by increasing the capillary diameter to 1.8 mm. Besides, these conditions tend to produce wavy filaments with high variability in thickness as shown in Fig. 6a. However, they produce smooth surface finishes, which is explained by low friction with the capillary walls and the capillary end-edges as the filament is prone to shrink. Low speeds (<50 mm/s) and temperatures (<115 ◦C) resulted in an expansion of the extruded filament at the nozzle end, resulting in the filament’s diameter being higher than that of the capillary and with high rough surfaces and cracks (see Fig. 6b). These cracks make the filament susceptible to breaking when spooled. The set-up parameters of 120 ◦C, 50 mm/s and 1.7 mm of extrusion temperature, speed and capillary diameter, respectively, resulted in satisfactory filaments (Fig. 6c) with soft surface finishes, a constant diameter of 1.74 ±0.01 mm, and properties enabling them to be spooled without breaking. A processing parameter window around these values, in which filaments with adequate properties are attainable, should be further determined. Fig. 6d shows a coil with the spooled filament resulting from the adjusted combination of extrusion parameters, demonstrating the flexibility of the filament. A closer look at the filament’s cross section by SEM in Fig. 6e reveals a homogeneous distribution of all the feedstock constituents. The spherical powder agglomerates have been crushed and disaggregated by the shear forces during the mixing as expected, and the nanometric particles have been homogeneously dispersed into the binder. This again explains the high torque values recorded during the mixing experiment in which the advantageous agglomerates size and morphology are lost and the high specific area of the low particle size powder generates high internal friction forces in the system. It is worth noting that the extrusion of a practical filament at good spooling rates was possible at temperatures far below the mixing temperature. Hence, the low fluency of the feedstock in this particular case was not a determinant for achieving Fig. 4. Raw ceramic powder characterization: a) SEM images, and b) DRX pattern. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3399 good extrusion results. Indeed, the resulting filament is valid for printing using the printing parameters described in Section 2.3. Fig. 7 shows stereoscope microscope images of the lateral and top sides of a solid and porous hexagon in the green state. Layers of around 0.2 mm thickness piles along the hexagon thickness exhibit good adhesion and no apparent piling faults. The deposited filament width is close to the nozzle diameter, though it tends to slightly expand due to the pressure exerted by the nozzle. The dimensions of the printed hexagons deviate 0.4 ±0.1% in thickness and 4 ±0.3% in the parallel sides from the CAD design dimensions. This corresponds to ~0.02 mm and ~0.5 mm deviation in thickness and width, respectively. Taking into account the 0.6 mm nozzle, this can be considered good shape retention and process variability. Ongoing research is being carried out to adjust the composition of the filaments without compromising other requirements, to enable a better resolution by using 0.4 mm or lower nozzle diameters. Regarding the debinding step, the open porosity of the grafts facilitates a single-step thermal debinding minimizing the formation of defects, hence the use of an intermediate solvent debinding was in principle avoided. The thermal decomposition of the organic components of the feedstocks determined by TGA occurs in four well-defined stages, corresponding to the derivative curve valleys numbered in Fig. 2a. This result agrees with the multi-component nature of the binder system. Given the TGA results, the thermal cycle of Fig. 2b was used, in which a heating rate as low as 0.2 ◦C/min was necessary to avoid the formation of blisters and cracks in the brown solid hexagons. to produce flawless brown specimens. This thickness threshold should be considered Eventually, the thickness of such solid hexagons (100% infill) had to be reduced to 2.5 mm in the design of thicker wall grafts, which might require a solvent debinding step. Nevertheless, the porous hexagons did not exhibit visible problems during the thermal debinding with heating rates as high as 1◦C/min. The effectiveness of the debinding, in terms of Fig. 5. a) Evolution of torque with time during the mixing process of 50 vol% YSZ powder feedstock at 160 ◦C and 40 rpm. Note: Images of the feedstock as extracted from the mixing chamber are included. b) Comparison between commercial ceramic particles (left; SEM image by secondary electron detector; agglomerates according to D 10 =1.6 μ m, D 50 =20.5 μ m and D 90 =50.0 μ m) and ceramic particles once incorporated into the mixture (right; SEM image by backscattered electron detector; note that lighter shadow corresponds to ceramic particles), c) exposure of the nanometric nature of particles incorporated in the mixture (SEM image by secondary electron detector; note that lighter shadow corresponds to ceramic particles). Fig. 6. Influence of temperature and extrusion speed on the diameter, structural integrity, flexibility and surface roughness of filaments suitable for 3D printing: (a) high speed and temperature; (b) low speed and temperature, (c) optimal extrusion parameters; (d) coil with the spooled filament and (e) SEM of the filament cross-section. G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3400 generated defects, is evaluated in detail in the sintered specimens, as the brown specimens are prone to crumble due to the powder particles not jet physically jointed. The same thermal cycle as for the solid hexagons was conducted in porous hexagons though, for direct comparison and to ensure good results. Difficulties encountered in the debinding step can be explained by the nanometric size of the powders. The vaporous species generated during the thermal decomposition have to be evacuated from the system throughout the channels of interconnecting pores left by the contacting particles. These channels are reduced in size as the particle size decreases, hindering binder removal. Nevertheless, practically 100% of the binder system is eliminated when the 500 ◦C temperature is reached, with the remaining weight percentage (~84 wt%) falling within the error window of the thermogravimetry equipment for the YSZ powder percentage in the mixture (Fig. 2a). Any defect generated during printing or debinding could not be corrected during the sintering stage. Fig. 8a shows the effect of the sintering temperature on the volumetric and linear (x-y and z-axis) shrinkage relative to the dimensions of green porous and solid hexagons. The shrinkage during sintering is pronounced due to the consolidation of the powder particles by diffusive mechanisms. Indeed, the sintering process of YSZ ceramics consists of two concurrent fundamental phenomena: 1) densification and 2) grain growth. Both phenomena rely on diffusion mechanisms occurring during sintering and thus are activated by increasing temperature. During densification, voids resulting from powder particles packing are closed producing a shrinkage of the parts. Porous hexagons experience a more prominent shrinkage than solid hexagons, though the sintering behaviour is similar in both cases. In the temperature range below 1300 ◦C, the sintered specimen dimensions significantly decrease with the increment of the sintering temperature. Above 1300 ◦C, there is not a significant change in the shrinkage values compared to the studied sintering temperatures. The densification with temperature is comparable to other sintering studies in 3Y-TZP parts processed by other technologies and base powders [71–73]. In all the consulted references, Fig. 7. Stereoscope microscope images of the green (as-printed) hexagon-shaped implants (solid and porous). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Fig. 8. Effect of sintering temperature on specimen shrinkage (a) and densification (b) parameters.. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3401 the maximum densification rate occurs in 1200–1250 ◦C and densification is completed in 1300–1350 ◦C. Oh et al. [74] reported linear shrinkage values in the range of 20–22% in uniaxial press-formed YSZ compacts sintered at 1450 ◦C. Ji et al. [75] measured an anisotropic linear shrinkage in dense YSZ parts processed by stereolithography, with a diameter shrinkage of ~15% and a thickness shrinkage between 25 and 30% in the 1150–1550 ◦C temperature range. Hence, it can be highlighted that the current FFF process produced an isotropic shrinkage, with slight differences between thickness and width, and shrinkage values equivalent to other processing methods counterparts. Fig. 9 shows the SEM micrographs of solid hexagons sintered at different temperatures. In each case, the average grain size determined from SEM micrographs applying the linear intercept method is included. In the 1100 ◦C specimen, nanometric-size grains are observed along with a large population of homogeneously dispersed pores with a size ranging from that of the grains. The high porosity is in line with the low shrinkage and relative density (86.5 ±2.1% of that of the tetragonal zirconia) and suggests that early stages of the sintering were achieved at the selected condition. As the sintering temperature is increased, the size of the YSZ grains increases and the observed porosity is reduced. At 1300 ◦C, just a few pores could be distinguished and the densification at 1400 ◦C seems almost complete. The relative density for both conditions is respectively, 88.2 ±4.13% and 91.5 ±1.27%. These results are in agreement with the shrinkage trend in Fig. 8 and the porosity observed at SEM, but future work enhancing densification will be conducted in terms of reducing printing defects. The grain growth is considerably activated at 1500 ◦C, and the grain size slightly exceeds the standard for biomedical applications. Additionally, XRD experiments were conducted to determine the phases present in the microstructure. Fig. 10 corresponds to the diffractogram of a 100% infill specimen sintered at 1400◦C, which is representative of the specimens sintered at other temperatures. The tetragonal ZrO 2 is the predominant phase in all the sintered specimens, which is in good agreement with the 3 mol% of Y 2 O 3 present in the base powder. In some cases, the specimens exhibit traces of XRD signals that could not be identified due to their low intensity and that can be attributed to a negligible presence of secondary phases. Table 1 shows the calculated crystal domain size from the XRD results of the different sintered specimens. The crystal domain increases with the sintered temperature in agreement with the SEM observations. Given these results, 1400◦C is selected as the sintering temperature as it resulted in high densification, a near full tetragonal ZrO 2 phase and maintains a small grain size which is considered positive for the mechanical performance [71–74] and in agreement with the standards of YSZ parts for biomedical applications (grain sizes below the 400 nm) [46]. Nanoparticles tend to inhibit grain growth due to their high surface energy, which acts as a barrier to the movement of grain boundaries. This suppression of grain growth helps to maintain a fine microstructure and enhances the mechanical properties of the sintered part, this great final grain size was mainly possible due to the nanometric powder sizes used in the filaments. Once the YSZ filaments are validated, the full potential of FFF to create such complex geometries is discussed in this next section, with the proposition and fabrication of innovative graft designs. Fig. 11 shows the dimensional change of the grafts throughout the FFF process and Table 2 the variation of the base cylinder external diameter of the designed grafts at different stages. The values are an average of the dimensions obtained from 3 different specimens. The deviation from the CAD dimensions of the green specimen in Design 1 can be associated with the shrinkage occurring due to the solidification of the extruded melt and thermal contraction. In the case of Design 2, the diameter of the green specimen is higher than that for the CAD design and thus an expansion is measured. Assuming that the thermal shrinkage also occurs in Design 2, the printing inaccuracies are large for this design. The printing parameters were not varied from that adjusted in Fig. 9. SEM micrographs of the surface of specimens sintered at different temperatures (15k magnification in all micrographs). GS: grain size determined by the linear intercept method. Fig. 10. XRD diffractogram of a hexagon sintered at 1400 ◦C. The Rietveld fitting profile and the theoretical profile of different phases present in the microstructure are outlined in different colours. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) G. Herranz et al. Journal of Materials Research and Technology 29 (2024) 3393–3408 3408 [56] Knetsch MLW, Koole LH. New strategies in the development of antimicrobial coatings: the example of increasing usage of silver and silver nanoparticles. Polymers 2011;3:340–66. https://doi.org/10.3390/polym3010340. [57] Alcudia A, Begines B, Rodriguez-Lejarraga P, Greyer V, Godinho VCF, Pajuelo E, Torres Y. Development of porous silver nanoparticle/polycaprolactone/polyvinyl alcohol coatings for prophylaxis in titanium interconnected samples for dental implants. Colloid and Interface Science Communications 2022;48:100621. https://doi.org/10.1016/j.colcom.2022.100621. [58] Burdușel A-C, Gherasim O, Grumezescu AM, Mogoant˘ a L, Ficai A, Andronescu E. Biomedical applications of silver nanoparticles: an up-to-date overview. Nanomaterials 2018;8. https://doi.org/10.3390/nano8090681. [59] Gaviria J, Alcudia A, Begines B, Beltr´ an AM, Rodríguez-Ortiz JA, Trueba P, Villarraga J, Torres Y. Biofunctionalization of porous Ti substrates coated with Ag nanoparticles for potential antibacterial behavior. Metals 2021;11:692. https:// doi.org/10.3390/met11050692. [60] Begines B, Ortiz T, P´ erez-Aranda M, Martínez G, Merinero M, Argüelles-Arias F, Alcudia A. Polymeric nanoparticles for drug delivery: recent developments and future prospects. Nanomaterials 2020;10. https://doi.org/10.3390/ nano10071403. [61] Raina N, Pahwa R, Khosla JK, Gupta PN, Gupta M. Polycaprolactone-based materials in wound healing applications. Polym Bull 2022;79:7041–63. https:// doi.org/10.1007/s00289-021-03865-w. [62] Gao C, Gao Q, Li Y, Rahaman MN, Teramoto A, Abe K. Preparation and in vitro characterization of electrospun PVA scaffolds coated with bioactive glass for bone regeneration. J Biomed Mater Res 2012;100:1324–34. https://doi.org/10.1002/ jbm.a.34072. [63] Lutterotti L, Pilli` ere H, Fontugne C, Boullay P, Chateigner D. Full-profile searchmatch by the Rietveld method. J Appl Crystallogr 2019;52:587–98. https://doi. org/10.1107/S160057671900342X. [64] Naranjo JA, Berges C, Campana R, Herranz G. Rheological and mechanical assessment for formulating hybrid feedstock to be used in MIM & FFF. Result Eng 2023;19:101258. https://doi.org/10.1016/j.rineng.2023.101258. [65] Naranjo JA, Berges C, Gallego A, Herranz G. A novel printable high-speed steel filament: towards the solution for wear-resistant customized tools by AM alternative. J Mater Res Technol 2021;11:1534–47. https://doi.org/10.1016/j. jmrt.2021.02.001. [66] Vidal L, Kampleitner C, Krissian S, Brennan M´ A, Hoffmann O, Raymond Y, Maazouz Y, Ginebra M-P, Rosset P, Layrolle P. Regeneration of segmental defects in metatarsus of sheep with vascularized and customized 3D-printed calcium phosphate scaffolds. Sci Rep 2020;10:7068. https://doi.org/10.1038/s41598- 020-63742-w. [67] Chen Y, Li W, Zhang C, Wu Z, Liu J. Recent developments of biomaterials for additive manufacturing of bone scaffolds. Adv Healthcare Mater 2020;9: e2000724. https://doi.org/10.1002/adhm.202000724. [68] Chen Z, Yan X, Yin S, Liu L, Liu X, Zhao G, Ma W, Qi W, Ren Z, Liao H, Liu M, Cai D, Fang H. Influence of the pore size and porosity of selective laser melted Ti6Al4V ELI porous scaffold on cell proliferation, osteogenesis and bone ingrowth, Materials science & engineering. C, Materials for biological applications 2020;106:110289. https://doi.org/10.1016/j.msec.2019.110289. [69] Cheng A, Humayun A, Cohen DJ, Boyan BD, Schwartz Z. Additively manufactured 3D porous Ti-6Al-4V constructs mimic trabecular bone structure and regulate osteoblast proliferation, differentiation and local factor production in a porosity and surface roughness dependent manner. Biofabrication 2014;6:45007. https:// doi.org/10.1088/1758-5082/6/4/045007. [70] Madhavan RV, Rosemary MJ, Nandkumar MA, Krishnan KV, Krishnan LK. Silver nanoparticle impregnated poly ( ε -caprolactone) scaffolds: optimization of antimicrobial and noncytotoxic concentrations. Tissue Eng 2011;17:439–49. https://doi.org/10.1089/ten.TEA.2009.0791. [71] Ji M, Xu J, Li L, Yu D, Chen M, Geier N, El Mansori M. Investigation of material removal mechanisms and ductile-brittle transition zone of zirconia ceramics sintered at various temperatures. J Mech Behav Biomed Mater 2022;125:104944. https://doi.org/10.1016/j.jmbbm.2021.104944. [72] Ji M, Xu J, Yu D, Chen M, El Mansori M. Influence of sintering temperatures on material properties and corresponding milling machinability of zirconia ceramics. J Manuf Process 2021;68:646–56. https://doi.org/10.1016/j. jmapro.2021.05.012. [73] Othman S, Ramesh S, Teng W. Sintering of commercial yttria-stabilized zirconia. 2006. undefined. [74] Oh G-J, Yun K-D, Lee K-M, Lim H-P, Park S-W. Sintering behavior and mechanical properties of zirconia compacts fabricated by uniaxial press forming. The journal of advanced prosthodontics 2010;2:81–7. https://doi.org/10.4047/ jap.2010.2.3.81. [75] Ji SH, Da Kim S, Park MS, Yun JS. Sintering process optimization for 3YSZ ceramic 3D-printed objects manufactured by stereolithography. Nanomaterials 2021;11. https://doi.org/10.3390/nano11010192. [76] Majhy B, Priyadarshini P, Sen AK. Effect of surface energy and roughness on cell adhesion and growth - facile surface modification for enhanced cell culture. RSC Adv 2021;11:15467–76. https://doi.org/10.1039/d1ra02402g. [77] Marin E, Zanocco M, Boschetto F, Santini M, Zhu W, Adachi T, Ohgitani E, McEntire BJ, Bal BS, Pezzotti G. Silicon nitride laser cladding: a feasible technique to improve the biological response of zirconia. Mater Des 2020;191:108649. https://doi.org/10.1016/j.matdes.2020.108649. [78] Kuboki Y, Jin Q, Takita H. Geometry of carriers controlling phenotypic expression in BMP-induced osteogenesis and chondrogenesis. J Bone Joint Surg 2001;83(A Suppl 1):S105–15. American volume. [79] LeGeros RZ. Properties of osteoconductive biomaterials: calcium phosphates. Clinical orthopaedics and related research; 2002. p. 81–98. https://doi.org/ 10.1097/00003086-200202000-00009. [80] Tsuruga E, Takita H, Itoh H, Wakisaka Y, Kuboki Y. Pore size of porous hydroxyapatite as the cell-substratum controls BMP-induced osteogenesis. J Biochem 1997;121:317–24. https://doi.org/10.1093/oxfordjournals.jbchem. a021589. [81] Perez RA, Mestres G. Role of pore size and morphology in musculo-skeletal tissue regeneration, Materials science & engineering. C, Materials for biological applications 2016;61:922–39. https://doi.org/10.1016/j.msec.2015.12.087. [82] Bernstein A, Niemeyer P, Salzmann G, Südkamp NP, Hube R, Klehm J, Menzel M, von Eisenhart-Rothe R, Bohner M, G¨ orz L, Mayr HO. Microporous calcium phosphate ceramics as tissue engineering scaffolds for the repair of osteochondral defects: histological results. Acta Biomater 2013;9:7490–505. https://doi.org/ 10.1016/j.actbio.2013.03.021. [83] Ghayor C, Weber FE. Osteoconductive microarchitecture of bone substitutes for bone regeneration revisited. Front Physiol 2018;9:960. https://doi.org/10.3389/ fphys.2018.00960. [84] Polak SJ, Rustom LE, Genin GM, Talcott M, Wagoner Johnson AJ. A mechanism for effective cell-seeding in rigid, microporous substrates. Acta Biomater 2013;9: 7977–86. https://doi.org/10.1016/j.actbio.2013.04.040. [85] Draenert K, Draenert M, Erler M, Draenert A, Draenert Y. How bone forms in large cancellous defects: critical analysis based on experimental work and literature. Injury 2011;42(Suppl 2):S47–55. https://doi.org/10.1016/j.injury.2011.06.007. [86] Delgado-Pujol EJ, Alcudia A, Elhadad AA, Rodríguez-Albelo LM, Navarro P, Begines B, Torres Y. Porous beta titanium alloy coated with a therapeutic biopolymeric composite to improve tribomechanical and biofunctional balance. Mater Chem Phys 2023;300:127559. https://doi.org/10.1016/j. matchemphys.2023.127559. [87] Gaviria J, Alcudia A, Begines B, Beltr´ an AM, Villarraga J, Moriche R, Rodríguez- Ortiz JA, Torres Y. Synthesis and deposition of silver nanoparticles on porous titanium substrates for biomedical applications. Surf Coating Technol 2021;406: 126667. https://doi.org/10.1016/j.surfcoat.2020.126667. [88] Hart NH, Nimphius S, Rantalainen T, Ireland A, Siafarikas A, Newton RU. Mechanical basis of bone strength: influence of bone material, bone structure and muscle action. J Musculoskelet Neuronal Interact 2017;17:114–39. [89] Currey J. 1 - the structure and mechanical properties of bone. In: Kokubo T, editor. Bioceramics and their clinical applications. Boca Raton: CRC Press; 2008. p. 3–27. [90] Morgan EF, Unnikrisnan GU, Hussein AI. Bone mechanical properties in healthy and diseased states. Annu Rev Biomed Eng 2018;20:119–43. https://doi.org/ 10.1146/annurev-bioeng-062117-121139. [91] Gdoutos EE, Raftopoulos DD, Baril JD. A critical review of the biomechanical stress analysis of the human femur. Biomaterials 1982;3:2–8. https://doi.org/ 10.1016/0142-9612(82)90053-9. [92] Singh Gangwar AK, Sudhakar Rao P, Kumar A. Bio-mechanical design and analysis of femur bone. Mater Today Proc 2021;44:2179–87. https://doi.org/ 10.1016/j.matpr.2020.12.282. [93] Touaiher I, Saadaoui M, Reynaud P, Reveron H, Chevalier J. Mechanical properties of additive-manufactured hydroxyapatite porous scaffolds and followup of damage process under compression loading. Open Ceramics 2023;16: 100498. https://doi.org/10.1016/j.oceram.2023.100498. [94] Ryan E, Yin S. Compressive strength of β-TCP scaffolds fabricated via lithographybased manufacturing for bone tissue engineering. Ceram Int 2022;48:15516–24. https://doi.org/10.1016/j.ceramint.2022.02.085. [95] Fu J, Goldsmith M, Crooks SD, Condon SF, Morris M, Komarova SV. Bone health in spacefaring rodents and primates: systematic review and meta-analysis. NPJ microgravity 2021;7:19. https://doi.org/10.1038/s41526-021-00147-7. [96] Miranda P, Pajares A, Saiz E, Tomsia AP, Guiberteau F. Fracture modes under uniaxial compression in hydroxyapatite scaffolds fabricated by robocasting. J Biomed Mater Res A 2007;83:646–55. https://doi.org/10.1002/jbm.a.31272. [97] Mohammadi M, Pascaud-Mathieu P, Allizond V, Tulliani J-M, Coppola B, Banche G, Chaput C, Cuffini AM, Rossignol F, Palmero P. Robocasting of single and multi-functional calcium phosphate scaffolds and its hybridization with conventional techniques: design, fabrication and characterization. Appl Sci 2020; 10:8677. https://doi.org/10.3390/app10238677. [98] Buj-Corral I, Vidal D, Tejo-Otero A, Padilla JA, Xuriguera E, Fenollosa-Art´ es F. Characterization of 3D printed yttria-stabilized zirconia parts for use in prostheses. Nanomaterials 2021;11:2942. https://doi.org/10.3390/ nano11112942. [99] Papini M, Zdero R, Schemitsch EH, Zalzal P. The biomechanics of human femurs in axial and torsional loading: comparison of finite element analysis, human cadaveric femurs, and synthetic femurs. J Biomech Eng 2007;129:12–9. https:// doi.org/10.1115/1.2401178. [100] Raffa ML, Nguyen V-H, Hernigou P, Flouzat-Lachaniette C-H, Haiat G. Stress shielding at the bone-implant interface: influence of surface roughness and of the bone-implant contact ratio. Journal of orthopaedic research official publication of the Orthopaedic Research Society 2021;39:1174–83. https://doi.org/10.1002/ jor.24840. [101] Kumar KN, Tandon T, Silori P, Shaikh A. Biomechanical stress analysis of a human femur bone using ANSYS. Mater Today Proc 2015;2:2115–20. https://doi.org/ 10.1016/j.matpr.2015.07.211. G. Herranz et al.