Fatigue performance of hierarchically porous titanium scaffolds produced by additive manufacturing and its possible improvement by gas nitriding

This contribution focuses on the nitriding of hierarchically porous titanium scaffolds to enhance their fatigue behaviour. Firstly, recent experimental findings that demonstrate the benefits of intra-filament porosity in improving fatigue resistance are discussed, providing details on crack growth shielding micromechanisms. Subsequently, the study explores the application of titanium scaffolds nitriding as a promising technique to prolong fatigue crack initiation. The scaffolds, prepared using the direct ink writing method with intra-filament porosity of ~ 6% and inter-filament porosity of ~ 68%, underwent gas nitriding at 1100 °C for 2 h. This process resulted in the formation of a consistent 42 μm thick nitriding case across the entire structure. Preliminary experiments showed a minimal decrease in fatigue strength within the low cycle fatigue region, attributed to the fracturing of a thick brittle compound zone under high applied loading. These results suggest that nitriding has the potential to improve fatigue performance after process optimization.


Introduction
Porous metallic structures (scaffolds) prepared by additive manufacturing (AM) are currently receiving significant attention in advanced industrial sectors.The reasons are evident: the possibility to reproducibly fabricate metallic meta-materials (materials with architected structure), enabled by modern AM methods, provides specific properties to the final products.For instance, porous meta-biomaterials can be produced with a unique shape and complex porosity that fit patient's specific needs and support implant osseointegration by mimicking the properties of surrounding bone [1].
Direct ink writing (DIW) is a rapidly evolving AM method that enables the production of hierarchically porous metallic structures.DIW is based on the automated extrusion of a suspension containing metallic particles in a liquid binder (ink) to create scaffolds, which are then sintered for consolidation [2,3].This process allows for the independent control of macroporosity (inter-filament porosity) and microporosity (intra-filament porosity) through digital design and sintering conditions, respectively.DIW is well-suited for fabricating structures with varying degrees of microporosity, including open intra-filament porosity (see figure 1a), which is expected to enhance biomechanical compatibility by reducing stiffness and can also serve as a drug reservoir to aid in the healing process.On the other hand, micropores, along with surface roughness, can act as fatigue crack initiators and pose a significant challenge for AM products that function as critical load-bearing components [4,5].
Interestingly, recent fatigue experiments conducted on commercially pure titanium DIW scaffolds have yielded intriguing findings.Contrary to expectations, it was discovered that high volumetric microporosity does not necessarily deteriorate fatigue behavior; instead, it can actually enhance it [6].The study revealed that scaffolds with 14.3% microporosity (porous) endured nearly an order of magnitude more cycles compared to those with 5.9% microporosity (compact).Another notable distinction between the porous and compact scaffolds was the difference in grain size (53 vs. 92 m) and surface topography (developed interfacial area ratio [7] of 1.8 vs. 0.6).Detailed examination of fracture surfaces and analysis of the stress-strain response during cycling loading demonstrated that smaller grains and higher porosity more than compensate for somewhat earlier initiation of fatigue cracks in porous scaffolds by impeding their growth.
In pure titanium, smaller grains contribute to prolonged crystallographic growth of short cracks [8].As the crack crosses grain boundary and accommodates to remote mode I (opening) loading, its size becomes comparable to micropores, and it begins to interact with them.As a result, the crack is locally deflected, bifurcated, blunted, and widened as it advances through the pore space, which is prevalent in filaments of porous structures, figure 1b.Crack deflections and secondary cracks enlarge the crack path, increase energy absorption, and intensify beneficial crack growth shielding effects related to premature contact of rough fracture surfaces (roughness-induced crack closure) and the irregular crack front (geometrically induced crack front shielding) [9].This mechanism has proven to be highly effective since crack growth in AM materials significantly contributes to the total fatigue life to failure, Nf, where nanocracks are nucleated quickly, unlike in traditional dense samples with smooth surfaces.The number of cycles associated with nucleation, Nc, can be estimated using the revised Tanaka-Mura model [10]: where μ is the shear modulus, ws is the surface energy, R is the roughness factor, b is the Burgers Further enhancement of the fatigue resistance of these scaffolds can be achieved by carefully optimizing the pore space and grain size.Additionally, exploring novel types of powders for preparing DIW inks and employing different sintering processes and trajectories can help achieve the desired densification.For instance, the pressureless spark plasma sintering method can rapidly consolidate titanium scaffolds without significant grain coarsening [11].Another approach involves employing case hardening to delay the nucleation of fatigue cracks, similar to the strategy used for sintered steels in the automotive industry, which are typically case hardened [12].This study focuses on analyzing the latter approach and provides initial data on gas nitriding.
Gas nitriding of titanium is a preferred case hardening process due to its ease of application to both titanium and its alloys.This process has been well established in bulk dense samples and offers several advantages.Titanium nitrides formed during gas nitriding exhibit high biocompatibility [13] and are wear resistant [14], making them suitable for novel meta-biomaterials.The nitriding procedure involves the adsorbtion of nitrogen species and their subsequent diffusion into the material, resulting in the formation of a solid solution strengthened diffusion zone and a micrometric surface compound layer composed of titanium nitrides.The compound layer provides hardness, while the diffusion zone harbours compressive residual stresses due to a distorted crystal lattice, thereby retarding the initiation of fatigue cracks.In this study, we present our initial evaluation of the feasibility of gas nitriding for titanium scaffolds and outline its potential to enhance the fatigue resistance of titanium DIW scaffolds based on the preliminary fatigue experiments.

Sample preparation
Cylindrical scaffolds were printed using commercially pure ASTM Grade 1 titanium powder (TLS-Technik, Germany) and a gelatin-based binder.The printing process was carried out using a commercial printer (Pastecaster, Fundació CIM, Spain) equipped with a conical dispensing nozzle with an aperture of 610 m.Subsequently, the scaffolds were sintered in an inert argon atmosphere at 1400 °C for 10 h to achieve a macroporosity of 68% and a microporosity of 6%.Gas nitriding was conducted in a tubular furnace (GSL-1700X-UL MTI, USA) at 1100 °C for 2 h.Pure nitrogen (100 ml/min) was used as the working gas, with argon purges.The objective was to achieve a nitriding case thickness of approximately 40 m based on previous research that reported gas nitriding of titanium in ammonia at the same temperature [15].

Microstructural characterization
Microstructural characterization was performed using a scanning electron microscope (SEM; Tescan Lyra3, Czech Republic) equipped with an energy-dispersive X-ray (EDX) spectrometer (Max N 50, Oxford Instruments, UK).Additionally, microhardness measurements were conducted using a hardness tester (QNESS 60A+ EVO, ATM Qness GmbH, Germany).The scaffolds were longitudinally cut and prepared following standard metallographic procedures.mid-profiles were obtained for one exterior filament and one centre filament, figure 2b.Hardness measurements were taken at the centre and both exteriors, analysing three filaments at each location.Indentations were made in the outer (nearsurface) and inner regions of the diffusion zone, as well as in the core material, for each filament.

Fatigue experiments
Fatigue experiments were carried out using a servo-hydraulic testing machine (Instron 8874) under ambient air conditions at room temperature.Prior to testing, the upper and lower surfaces of the scaffolds were polished to ensure flatness and parallelism.Cyclic compression loading, with a sinusoidal waveform and constant stress amplitude, was applied at a frequency of 10 Hz and a compressive stress ratio of 0.1.Following specimen failure, the scaffolds were examined in SEM to analyse micromechanisms of fracture.

Microstructural analysis
Figure 2a demonstrates that the nitriding process resulted in a case with an average thickness of 42 ± 7 m, exhibiting local variations attributed to surface irregularities and pores.In certain regions, complete encapsulation of near-surface pores was observed (as seen in the right part of the filament in figure 2a), indicating the potential for achieving volumetric nitridation in materials with open microporosity.The nitriding case comprised a compound layer, 4.2 ± 1.4 m thick, and a relatively deep diffusion zone.EDX analysis confirmed the presence of titanium nitrides and oxides in the compound layer (location 1 -70Ti 30N and location 2 -63.2Ti 27N 9.8O in figure 2b), whereas the core material remained pure titanium (location 3 -100Ti), with localized contamination of iron forming distinct bright phases (location 4 -95.4Ti 4.6 Fe).While the presence of oxides is typically anticipated unless a highvacuum furnace is utilized, the presence of iron resulted from powder contamination.Line mapping of nitrogen, titanium, oxygen, and carbon elements in the centre and exterior filaments (figure 3a) revealed that the nitriding of the scaffolds was homogeneous, with no significant differences observed between the centre and the exterior, except for a slightly higher scatter in the exterior filament towards one of its ends (D/D0 = 1), likely representing the outermost free surface of the scaffold.As expected, the nitrogen content increases rapidly while the titanium content decreased near the surface of the filament.The distinct concentration rises and drops correspond to the presence or close proximity of micropores, cf.figure 2b.The hardness data (figure 3b) also supported the conclusion of the homogeneity of the nitriding procedure.Near the surface, the hardness ranged between 780 and 930 HV0.1.Deeper in the diffusion zone, it decreased to 600 to 780 HV0.1, while the core material exhibited a hardness of 350 to 570 HV0.1, regardless of whether it was the centre or the exterior filament.
It is important to note that the scatter observed in these data is caused by the difficulties of placing the indent where the influence of surrounding features is minimal, which, however, cannot be avoided in the presence of unseen subsurface pores.

Fatigue experiments
The preliminary fatigue experiments revealed that the as-sintered scaffolds demonstrated slightly better fatigue performance compared to the nitrided scaffolds, figure 4a.At a fatigue life of approximately 2 × 10 3 cycles (Nf,1 = 1.8 × 10 3 , Nf,2 = 2.4 × 10 3 cycles), the fatigue strength of the nitrided samples was less than 5% lower than that of the as-sintered samples.It is worth noting that the high applied cyclic loading used in these experiments was chosen in the absence of prior experimental data and resulted in failure in the low cycle fatigue (LCF) domain [16].In this region, the scatter in fatigue life is generally low.These findings indicate that the nitriding of titanium scaffolds has the potential to enhance fatigue behaviour, particularly in the high cycle fatigue region (under low stresses), after process optimization.It is well recognized that the thickness of the compound layer plays a crucial role in determining the fatigue behaviour in nitrided materials, particularly in the LCF region.A thick compound layer, despite its hardness, can lead to the formation of deep and sharp initial brittle cracks, which have a detrimental effect on fatigue properties.On the other hand, a thin compound layer generally has a minimal impact [17,18].In the present study, the observed average thickness of the compound layer is 4.2 m, which is relatively large, especially when considering the presence of surface relief with peaks and valleys that contribute to stress concentration and premature fracture.This observation is supported by the obtained fractographic evidence, figure 4b.The fatigue cracks originated near the filament interconnections in regions experiencing the highest tensile stresses during the compression test [19].Notably, a variable surface rim was observed along the circumference of filaments, appearing microscopically very flat up to a depth of approximately 10 m, providing evidence of sudden fracture of the compound layer.As the initial crack propagated deeper into the diffusion zone under cyclic loading, it gradually became more complex, and in the core material, it exhibited characteristic features of pure titanium, including fatigue striations [16,20].The experimental evidence confirms that the reduction in LCF strength is related to the high thickness of the compound layer, which is prone to premature fracture under high applied stresses.Therefore, to improve the fatigue behaviour using nitriding, additional experiments are required to obtain an optimal nitriding case.

Conclusion
The hierarchically porous titanium scaffolds produced by direct ink writing exhibit promising fatigue properties, primarily attributable to the fatigue crack growth shielding effects induced by crack-pore interactions.In order to further enhance their fatigue resistance, gas nitriding has been proposed as a means to prolong the fatigue crack initiation period.The experimental results obtained from nitriding the scaffolds at 1100 °C for 2 h have demonstrated the formation of a homogeneous nitriding case, approximately 42 m thick across the entire structure.The nitriding case was locally influenced by surface relief and the presence of micropores.Preliminary fatigue testing of nitrided scaffolds has revealed that the brittle surface compound layer, with an average thickness of 4.2 m, leads to a minimal decrease of approximately 5% in fatigue strength within the LCF domain.This finding highlights the importance of optimizing the nitriding process.Consequently, future efforts will be focused on conducting experiments using a large sample set to explore diverse nitriding conditions and further optimize the process.

Figure 1 .
Figure 1.(a) CT rendering of intra-filament pores in pure titanium DIW scaffolds.Detected pores are shown in different colours.The open pore network (green) accounts for 12.6% microporosity of the porous structures.(b) Experimental evidence of extensive crack-pore interactions in scaffolds with porous design.White arrows point to secondary cracks in the interior (Detail A) and near-surface (Detail B) regions of a fractured filament.

Figure 2 .
Figure 2. (a) General overview of a centre filament (located in the inset) with marked placements and sizes of indents in the inner and outer diffusion zone and the core material.(b) Centre and exterior filaments with indicated locations and profiles analysed by EDX.

Figure 3 .
Figure 3. (a) EDX line mapping of Ti, N, O, and C in the centre and exterior filaments (figure 2a).The distance D along the profile is normalized by the diameter of the filament D0.(b) Vickers hardness of interior and exterior filaments in the outer and inner diffusion zone (DZ) and the core material.

Figure 4 .
Figure 4. (a) S-N data with conventional division between the very low cycle (VLCF), low cycle (LCF), and high cycle fatigue (HCF) regions.(b) Representative images of fatigue failure of filaments in nitrided scaffolds.