<?xml-model href='http://www.tei-c.org/release/xml/tei/custom/schema/relaxng/tei_all.rng' schematypens='http://relaxng.org/ns/structure/1.0'?><TEI xmlns="http://www.tei-c.org/ns/1.0">
	<teiHeader>
		<fileDesc>
			<titleStmt><title level='a'>Nature-inspired orientation-dependent toughening mechanism for TPMS ceramic architectures</title></titleStmt>
			<publicationStmt>
				<publisher>Springer Nature</publisher>
				<date>04/01/2025</date>
			</publicationStmt>
			<sourceDesc>
				<bibl> 
					<idno type="par_id">10698395</idno>
					<idno type="doi">10.1557/s43577-024-00831-5</idno>
					<title level='j'>MRS Bulletin</title>
<idno>0883-7694</idno>
<biblScope unit="volume">50</biblScope>
<biblScope unit="issue">4</biblScope>					

					<author>Luca D’Andrea</author><author>Ting Yang</author><author>Ming Dao</author><author>Pasquale Vena</author>
				</bibl>
			</sourceDesc>
		</fileDesc>
		<profileDesc>
			<abstract><ab><![CDATA[Triply periodic minimal surfaces (TPMSs) have been extensively studied in many fields of engineering, including bone tissue scaffolds. Recent advancements in manufacturing have enabled the three-dimensional printing of ceramic porous architectures; however, their intrinsic brittleness limits its practical applications. It has been observed that the ossicles of the knobby starfish exhibit a mineralized TPMS structure with lattice distortions (i.e., dislocations), which effectively deviate the crack propagation and enhance the fracture energy. In this article, the aforementioned toughening mechanism has been introduced in a TPMS architecture. We employed finite element models to analyze the effective mechanical properties of the structures under compression, both in the elastic and post-elastic regimes. Our analysis reveals that the introduction of the dislocation induces variations in both elastic and fracture properties of the structures. With particular reference to the fracture behavior, a suitable oriented edge dislocation is able to alter the crack nucleation and propagation, resulting in a tougher structure. Both the elastic and fracture phenomena can be enhanced or reduced by changing the dislocation density.]]></ab></abstract>
		</profileDesc>
	</teiHeader>
	<text><body xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xlink="http://www.w3.org/1999/xlink">
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="1">Introduction</head><p>Triply-Periodic Minimal Surfaces (TPMS) has gained significant attention across various engineering disciplines, including heat exchangers, energy absorbers, microwave absorption, and bone grafting. From a structural point of view, the TPMS's analytical definition provides them with a morphology such that local stress concentrations are avoided <ref type="bibr">[1]</ref>.</p><p>Concerning Bone Tissue Engineering (BTE) applications, porous structures are required to guarantee the necessary fluid flow transport for cell proliferation <ref type="bibr">[2]</ref>.</p><p>Ceramic materials like hydroxyapatite (HAP) guarantee high biocompatibility and bioactivity <ref type="bibr">[3,</ref><ref type="bibr">4]</ref>. Nowadays, Additive Manufacturing (AM) technologies are able to produce three-dimensional scaffolds for BTE exhibiting micrometric features, ensuring high printing fidelity and repeatability <ref type="bibr">[5]</ref>. The main drawback of ceramic scaffolds is represented by their intrinsic brittleness, especially in load-bearing applications. Furthermore, the strength and the fracture energy are dramatically reduced because of the porosity required in BTE applications. Different methods have been already proposed to enhance the strength of porous ceramic scaffolds <ref type="bibr">[6]</ref>.</p><p>In nature TPMS are observed under various circumstances, such as cubic membranes <ref type="bibr">[7]</ref> and biological photonic crystals <ref type="bibr">[8]</ref><ref type="bibr">[9]</ref><ref type="bibr">[10]</ref>. Distortions from TPMS ideal geometries have been observed in gyroid mesophases of amphiphilic ternary fluids, that exhibit edge dislocations <ref type="bibr">[11]</ref>. Notwithstanding, the majority of investigations into TPMS structures have predominantly concentrated on non-mineralized materials at the nanometer scale, and seldom exploring the mechanical role of the structural defects. An exception to this trend is a study investigating starfish ossicles, which exhibit a TPMS structure at the microscale composed of mineralized materials <ref type="bibr">[12]</ref>. In particular, starfish ossicles exhibit a mineralized diamond TPMS structure. According to the research by Yang et al. <ref type="bibr">[12]</ref>, the TPMS configuration in starfish ossicles displays some defects, like 60 &#8226; edge dislocations and screw dislocations. Consequently, deformation bands often deviate from their course and are restricted by regions with high dislocation density. This factor forestalls the catastrophic failure of the lattice and facilitates a more controlled or graceful failure behavior. A comprehensive computational analysis of the elastic symmetry of the starfish ossicles has been presented by Chen et al. <ref type="bibr">[13]</ref>. In an other paper by Yang et al. <ref type="bibr">[14]</ref> the post elastic behavior in see urchin microstructure is investigated. In particular, the improved energy absorption capability introduced by organic component of the constituent material and the formation of damage bands was investigated. Other examples in nature suggest that the deviations from ideality (i.e., crystal misorientation) increase the toughness of materials <ref type="bibr">[15]</ref>.</p><p>The present work takes inspiration from the toughening effect of dislocation in the starfish ossicles: the Schwarz-P function has been modified, introducing a 1-dimension defect, namely a single edge dislocation, with the purpose of showing how suitably located defect can enhance the mechanical properties of the scaffold at the macroscopic scale, with particular reference to the toughness and fracture energy. To this purpose, Finite Element Models (FEM) have been used to simulate the mechanical response upon a compression load. The fracture response of the microstructure has been simulated by making use of the phase-field method. The nucleation and the propagation of the cracks have been investigated as well with the purpose of optimizing the scaffolds' mechanical performance. Proof-of-concepts experiments on physical models have been performed to corroborate the results obtained through the finite element models.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2">Materials and Methods</head></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2.1">Design of TPMS</head><p>The Schwarz-P structure is designed using Matlab 2020b, through the generation of a three-dimensional black and white binary model based on the following implicit trigonometric function</p><p>where x, y and z denote the spatial cartesian coordinates, the variable b represents a constant parameter, while c x , c y and c z signify the number of repetitive unit cells along the x, y, and z axes, respectively. The term n defines the number of pixels used for discretizing the function within the spatial domain. For this specific analysis, the parameters are set as follows: b = 0.8, c x = N , c y = 2N , c z = 1 and n = 60N + 1, where N is an integer number that assumes values from 2 to 6.</p><p>Notably, the initial function f (x, y, z) = 0 delineates a surface (i.e., zero thickness) in the three-dimensional space. To transform this surface into a volumetric entity, a specific criterion is employed for volume conversion <ref type="bibr">[16]</ref> -&#948; &lt; f (x, y, z) &lt; &#948;</p><p>(2)</p><p>The voxels satisfying the previous inequalities represent the solid fraction of the structure and the &#948; parameter has been set to 0.4.</p><p>In this study, the structural configuration comprises a single layer of unit cells along the z-axis, while varying the value N (2N ) in the direction x (y), as previously defined. To introduce an edge dislocation into the model, the parameter c x parameter has been redefined to be a function of the y coordinate. The functional relationship is established as follows:</p><p>N + 2 11 cyy-2(n-1)</p><p>where the function 3(b) provides a smooth transition between the initial and final value of c x . The obtained geometry is subjected to a mirror process about the yz-plane to achieve a square domain in the xy-plane (figure <ref type="figure">1</ref> shows the case of N = 4). Subsequently, the voxelized volume, delineated by the binary configuration of the image stack, underwent a smoothing process. This was accomplished by employing the Taubin filter within Meshlab with default parameters <ref type="bibr">[17]</ref>. After the smoothing, the geometry was converted into an STL file. Furthermore, the volumetric fraction, denoted as &#961;V , of the structure has been calculated as follows &#961;V = V V T OT <ref type="bibr">(4)</ref> where V is the volume enclosed by the surface and V T OT is the volume of the bounding box containing the surface.</p><p>The dislocation density (&#948;) has been computed as the ratio between the number of cells in the length (N ) of the dislocation and the total number of cells in the structure (4N 2 ), resulting in</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="2.2">Finite Element Models</head><p>The geometrical models have been imported in ABAQUS/Standard, where the volume has been discretized using linear tetrahedral elements (C3D4 for elastic analyses and C3D4T for fracture analyses). In the simulation, the Young's modulus of HAP has been assigned to E 0 = 100 GP a <ref type="bibr">[18]</ref> and the Poisson coefficient has been set at 0.3.</p><p>The homogenization process has been conducted in order to obtain the effective elastic parameters <ref type="bibr">[19]</ref>. This involved consideration of a single cell layer along the z-axis (c z = 1), which inherently limit the analysis to the extraction of 2D macroscopic mechanical properties within the xy-plane. More specifically, three simulations have been performed: two involving direct strain and one involving shear strain, each applying a unit strain. A plane strain condition has been imposed by restricting displacement along the z-direction on the two surfaces perpendicular to the z-direction. Subsequent post-processing focused exclusively on the macroscopic stress components in the xy-plane.</p><p>The calculation of the effective stiffness, denoted as &#274;, is carried out in accordance with the equation 6, which incorporate considerations of plane-strain conditions at the macroscopic level.</p><p>in which D ijkl represents fourth order compliance tensor derived by inverting the elastic stiffness tensor. Meanwhile, n is the unit vector that specifies the directions within the xy-plane through the angle &#952;</p><p>The fracture propagation process has been simulated by using FEM combined to a phase-field approach <ref type="bibr">[20]</ref>. This methodology has been previously applied in TPMS single cell microstructures <ref type="bibr">[21]</ref> and in ceramic architected materials <ref type="bibr">[22]</ref>. In order to account for the tension/compression strength mismatch typical of ceramic materials <ref type="bibr">[23,</ref><ref type="bibr">24]</ref>, the phase-field formulation reported in <ref type="bibr">[25]</ref> which integrates the Drucker-Prager failure criterion has been used in the current study. According to its expression in terms of invariants of stress, it states</p><p>where A and B depend on the intrinsic tensile strength (&#963; T ) and intrinsic compressive strength (&#963; C ) of the material</p><p>In this work, &#963; c has been imposed 20 times higher than &#963; T . This makes the failure mode dominated by the tensile stress, a characteristic mostly observed in ceramic materials <ref type="bibr">[26]</ref>. This assumption corresponds to B = -0.53. Furthermore, since the phase-field approach is based on Griffith's energy balance, the quantity G c has been computed as</p><p>where <ref type="bibr">27]</ref> and E 0 = 100 GP a <ref type="bibr">[18]</ref>. Thereby G c = 0.01 N/mm. Here a simplifying assumption of crack propagation on the basis of the mode I parameter (K Ic ) has been adopted.</p><p>Compression tests incorporating fracture propagation are simulated until complete rupture. Two distinct load cases are considered, differentiated by the orientation of the compression stress relative to the dislocation: i) the first load case involve compression stress aligned at a 0 &#8226; angle, signifying a direct alignment between the macroscopic stress and the dislocation line; ii) the second load case has the compression stress at a 90 &#8226; angle, indicating a perpendicular alignment between the macroscopic stress and the dislocation. For these simulations, a displacement control boundary condition has been implemented, characterized by a constant macroscopic strain increment of 3.81 &#8226; 10 -7 . This specific strain increment is determined based on a convergence analysis targeting macroscopic strength of the material. The criterion for selecting this increment is its effect on the strength value; the increment is adjusted downwards until any change in the strength value is lower than 1%.</p><p>The stress &#963; at each load increment has been calculated by dividing the total reaction force on the face where the displacement was applied by its apparent crosssectional area. Additionally, in order to characterize the stress drops observed during the fracture propagation process, the parameter &#8710;&#963; &#951; has been defined to quantitatively represent the percentage drop in stress relative to the maximum stress &#963; M AX (macroscopic strength) &#8710;&#963; &#951; = &#951;&#963; M AX <ref type="bibr">(11)</ref> &#8710;&#963; &#951; has been calculated as the reduction in stress expressed as a percentage &#951; of the peak stress &#963; M AX attained during the loading process.</p><p>The fracture energy per unit volume (W &#951; ), at given &#951;, has been calculated as</p><p>where the integral provides a comprehensive measure of the energy absorbed by the material up to the point of fracture and the strain &#949; &#951; has been specifically defined as the macroscopic strain that corresponds to the stress reduction denoted as &#8710;&#963; &#951; in the stress-strain diagram. The value of &#949; &#951; has been identified at the point on the curve where the stress experiences a decrement equal to &#8710;&#963; &#951; , indicating a significant alteration in the material's resistance to the applied stress.</p><p>Experimental tests have been performed on cubic-shaped samples (with 2N = 8 in each direction) 3D printed using a brittle polymeric resin resembling the brittle behavior of the HAP. Both the ideal and the geometries exhibiting the edge dislocations have been tested under monotonic uniaxial compression until failure. The fracture energy per unit volume at different &#951; has been evaluated as for the numerical models.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="3">Results and Discussion</head><p>In this section, except when explicitly reported, all considerations are referred to structures with 2N = 8. The normalized Young's modulus ( &#274;/E 0 ) is depicted in figure <ref type="figure">2</ref>, which elucidates the dependency of the macroscopic stiffness on the orientation of the structure. Specifically, the orientation at 45 &#8226; is identified as the one possessing the minimal macroscopic stiffness. This phenomenon can be attributed to the discontinuous distribution of the solid matrix in this orientation, which leads to a bending dominated deformation modalities. Indeed, when the loading axis is not aligned with the principal axes of the microstructure (i.e., the Cartesian directions), the bending dominated mechanisms become more pronounced. This phenomenon is universally observed in materials characterized by grid-like microstructures. According to the analytical definition of the Schwarz-P TPMS, the macroscopic stiffness along the two cartesian directions results in &#274;(0 &#8226; ) = &#274;(90 &#8226; ), whereas they are different in case of edge dislocation ( &#274;(0 &#8226; ) &gt; &#274;(90 &#8226; )). Since both microstructures have been designed to have the same overall volumetric fraction (25%), the observed variation in macroscopic stiffness is attributable to distinct deformation modes induced by the presence of dislocation. The elastic mismatch is also dependent on the number of cells constituting the geometry, and thereby they dependent on the dislocation density &#948;.</p><p>Figure <ref type="figure">3</ref> shows that increasing the number of cell in the structure, i.e. decreasing the dislocation density, the elastic anisotropy ratio (defined as &#274;(90)/ &#274;(0)) progressively approaches to 1. This indicates that the single dislocation becomes less relevant as &#948; decreases and the macroscopic behavior progressively approaches that of a structure without dislocation. In particular, when 2N &gt; 10 the elastic ratio is higher than 90%.</p><p>Comparative analysis reveals that the stiffness derived in our study for a microstructured material with volumetric fraction of 25% is consistent with the Fig. <ref type="figure">3</ref>: Ratio of the elastic mismatch of the two stiffness along the cartesian directions with respect to the number of repetitive units per edge. Note: each unit cell is represented by the same number of elements. findings of Jia et al. <ref type="bibr">[28]</ref> for Schwarz-P TPMS. Furthermore, our findings agree with Gibson and Ashby's theoretical predictions for bending-dominated microstructures. <ref type="bibr">[29]</ref>, especially for the loading directions which are not aligned with the main microstructure directions. Mirzaali et al. <ref type="bibr">[30]</ref> have also explored the influence of defect distribution within regular microstructures, demonstrating that the modulation of randomness can effectively adjust the Young's modulus in two-dimensional metamaterials.</p><p>Figure <ref type="figure">4</ref> delineates the stress-strain response obtained through the FEM analysis of the compression test. Here, the ideal structure displays a compressive strength equal to 2.93 M P a. For structures with edge dislocations, the compressive strength measures 2.36 M P a and 2.29 M P a when the load is aligned (&#952; = 0 &#8226; ) or orthogonal (&#952; = 90 &#8226; ) to the dislocation, respectively. Analysis of the entire stress-strain curve (&#951; = 100%) reveals that fracture energy for the ideal structure is 1.98 &#8226; 10 -3 J/mm 3 , while for structures with dislocation at 0 &#8226; and 90 &#8226; , the values are 1.26 &#8226; 10 -3 J/mm 3 and 2.44 &#8226; 10 -3 J/mm 3 , respectively. By comparing the response to load parallel to the dislocation (figure <ref type="figure">4 (a)</ref>) and that to load perpendicular to the dislocation (figure 4 (b)), a higher energy dissipation is found for perpendicular load (2.44 &#8226; 10 -3 J/mm 3 versus 1.26 &#8226; 10 -3 J/mm 3 ). Notably, a slightly reduced maximum strength (&#963; M AX ) is observed when the load is applied perpendicularly to the dislocation. This finding is consistent the commonly noted trade-off between strength and fracture energy <ref type="bibr">[31]</ref>.</p><p>FEM serves as a powerful tool not only for the prediction of the effective mechanical properties, but also for elucidating local phenomena that are challenging to detect experimentally, such as the stress distribution and the fracture pattern. As shown in figure <ref type="figure">5</ref>(a), the stress distribution in the regular lattice reflects the periodic cell arrangement. Consequently, the initiation of crack patterns, as observed in figure <ref type="figure">5</ref>(b), also follows a periodic nature. This distinctive feature leads to a simultaneous failure at multiple locations, which ultimately result in a catastrophic failure in the structure. A similar crack pattern (i.e., vertical cracks in the region connecting two repetitive units) has been found by Lu et al. <ref type="bibr">[32]</ref> in experimental compression tests on Schwarz-P TPMS with similar volumetric fraction of 21.4% (versus 25% of the present study) made by alumina.</p><p>It should be emphasized that our results come from numerical simulations which does not account for local defects that could be present in actual manufactured structures, potentially influencing the localization of crack patterns <ref type="bibr">[33]</ref>. Consistent with findings in studies of regular structures <ref type="bibr">[34]</ref>, the fracture occurring in the scaffold's microstructure is promoted by localized tensile stress, even under macroscopic compression tests.</p><p>As illustrated in figure <ref type="figure">5(c</ref>), the presence of the 0 &#8226; dislocation substantially alters the stress distribution within the structure. In scenarios where the applied load is aligned with this dislocation, a notable amplification of tensile stress is observed, particularly at critical points of potential fracture initiation. These points are predominantly located at the junctures between the tip of the dislocation and the adjacent structural units, as indicated by arrows labeled 'A'. This specific alignment proves to be more unfavorable than the ideal scenario, highlighting the structural vulnerability to fracture. The resultant crack pattern, as depicted in figure <ref type="figure">5(d)</ref>, deviates from a periodic arrangement. Instead, fractures are observed to disseminate throughout the entirety of the solid fraction of the material. This widespread distribution of fractures compromises the structural integrity across virtually the entire material, starkly contrasting with the behavior of more regularly arranged microstructures. Figure <ref type="figure">5</ref>(e) shows the stress distribution proximal to a 90 &#8226; dislocation. In this instance, the distortion introduced by the dislocation disrupts the periodicity of the stress distribution within the structure. Distinct from the 0 &#8226; dislocation scenario, the stress profile on the right side of the structure, particularly at the juncture between the dislocation tip and the structure, is predominantly characterized by local compressive stress, as indicated by arrows 'B'. This unique stress distribution fosters the initiation of fractures predominantly on the left side of the structure (figure <ref type="figure">5(f)</ref>). This asymmetric fracturing effectively shields the right side from early-stage fractures. The plateau observed in the macroscopic stress-strain response (figure <ref type="figure">4(b)</ref>) is indicative of the local shielding effect exerted by the dislocation. Similar phenomena have been observed in studies of materials with random architectures, such as those by Wang et al. <ref type="bibr">[35]</ref> and D'Andrea et al. <ref type="bibr">[36]</ref>. Their findings demonstrate heterogeneous stress distributions can prevent the formation of a continuous damage surface <ref type="bibr">[35]</ref> and mitigate detrimental fracture patterns <ref type="bibr">[36]</ref>. However, the inherent randomness of these structures presents significant challenges in precisely controlling stress distribution during the design phase. The management of fracture propagation in microarchitected materials is an open issue: for instance, Manno et al. <ref type="bibr">[37]</ref> tailored the crack propagation under uniaxial tensile load on notched samples by modifying the lattice structure in a two dimensional domain. Their approach primarily leveraged variations in the material's solid fraction to guide cracks through weaker regions. A different approach is presented by Indurkar et al. <ref type="bibr">[38]</ref> which conceived a network of crack arrestors in the multimaterial microstructure with the aim to improve toughness, and specifically to break the link between toughness and strength. Unlike the above-mentioned studies, the current research focuses on manipulating fracture responses through controlled modifications in micro-architecture while maintaining a constant overall volumetric fraction. This approach offers a novel method for influencing fracture behavior in microarchitected materials. A similar strategy has been used by Lu et al. <ref type="bibr">[32]</ref>, that introduced a smooth transition between two different TPMS (i.e., Schwarz-P and Gyroid) in order to induce a layer by layer failure, concluding that some hybridization improves the fracture energy per unit volume of 13%. Similarly to this study, Zhang et al. <ref type="bibr">[39]</ref> took inspiration to crystal twinning for modifying TPMS structure; although such modification allowed to alter the deformation mechanisms of the structures, they are evenly distributed along the structure, whereas in our approach the defects is only present in one region of the structure. An other key difference with respect to our study is identified by the material constitutive relations of the scaffold's solid fraction: in fact, Zhang et al. dealt with a ductile polymer characterized by a plastic behavior that eventually affects the deformation mechanism in a different way with respect to a brittle material.</p><p>Figure <ref type="figure">6</ref> elucidates the variation of the fracture energy parameter, denoted as W &#951; , across 6 distinct values of the percentage stress drops (&#951;). The data presented in the figure reveals that the structure with the 90 &#8226; dislocation exhibits higher fracture energy compared to the ideal structure for stress drops larger that 15%. A significant fracture energy enhancement is achieved in structure with the imperfections, but only for 90 &#8226; orientation. This observation is further substantiated by the plateau observed in the macroscopic stress-strain response in figure <ref type="figure">4(b</ref>). On the contrary, for the structure with a 0 &#8226; dislocation, the fracture energy exhibits a near-constant trend at different values of &#951;, meaning that the stress drop is instantaneous, corresponding to an immediate failure (see the stress-strain response in figure <ref type="figure">4(a)</ref>). The toughening effect of the 90 &#8226; dislocation is particularly evident at &#951; = 15%, where the value of fracture energy is 1.85 times higher than the ideal case.</p><p>The dependency of the fracture properties on the number of cells per edge (or the dislocation density) has been reported in figure <ref type="figure">7</ref>. As already observed in figure 5(f), the dislocation shields the right side of the structure inducing the crack nucleation on the left side of the dislocation (starting from the dislocation tip). It can be observed that the toughening effect is strictly dependent on the dislocation density; in particular, the optimal value is represented by 2N = 8, where the value of fracture energy is significantly higher than the ideal structure. In case of 2N = 6 the length of the preferential path of the crack is too short (identified by the red arrows in figure <ref type="figure">7</ref>); in this case the cracks soon reach the boundaries of the sample, causing a premature failure. A different phenomenon is observed when 2N = 10, where the stress-strain response of the structure with the edge dislocation becomes similar to the ideal structure. Indeed, even if the crack nucleation is still on the dislocation tip, the bottom-right part of the structure (highlighted by the blue box in figure <ref type="figure">7</ref>(f)) exhibits the regular crack pattern of the ideal structure. This indicates that, as observed in the elastic behavior, as the dislocation density decreases, the fracture response approaches that of a structure without the defect. The number of cells per edge does not affect the fracture energy of the ideal structure, that is almost constant across the different number of cells per edge.</p><p>The results of the experimental compression tests, in terms of W &#951; are reported in figure <ref type="figure">8</ref>.</p><p>Although a punctual quantitative comparison between the experimental measures and the numerical results cannot be done, an overall consistency with the major findings highlighted by the models is found. As for the numerical models, the experiments have confirmed that compression aligned with the dislocation shows an approximately constant fracture energy for all values of &#951; and results in a sudden loss of bearing capacity. The ideal structure exhibits an intermediate behavior, coherently to the numerical results, increasing the W &#951; as &#951; increases. Finally, the structure with the dislocation perpendicular to the load shows the highest values of W &#951; as a result of the plateau shown by the force displacement curve. Although a good consistency between the experiments and the model has been shown here, a more insightful quantitative comparison is recommended. To this purpose, ceramic scaffolds should be manufactured in a consistent number in order to take into account for both the distortions present in the manufacturing process and the scattering of fracture properties of architected ceramics (that is beyond the aim of this study). In the picture representing the 3D-printed scaffold, the dislocation plane is colored in red.</p></div>
<div xmlns="http://www.tei-c.org/ns/1.0"><head n="4">Conclusions</head><p>In this work, the design of modified Schwarz-P scaffolds with improved fracture properties has been investigated through finite element models and preliminary laboratory experiments. Our primary modification involved introducing a distortion, akin to an edge dislocation, into the ideal periodic lattice structure. This alteration aimed to enhance the fracture energy of the scaffolds under uniaxial compressive loading conditions, achieved by means of a peculiar spatial distribution of the fractures guided by the presence of the defect. The influence of the dislocation was found to be contingent upon its orientation relative to the applied load. Specifically, when the dislocation and the load form a 0 &#8226; angle, both the strength and the fracture energy are reduced; in contrast, when the angle is 90 &#8226; , the reduction in strength of 21.8% is compensated by fracture energy gain of 23.2%. In scaffolds with low dislocation density, the enhancement in fracture toughness reported for loading perpendicular to the dislocation vanishes, and the dominant behavior exhibited by structures without dislocation is once more noted. These insights are particularly pertinent to the design of porous structures in brittle materials, such as bone tissue engineering (BTE) scaffolds, where avoiding catastrophic failure is paramount. Additionally, this research highlights the necessity to consider multiaxial loading conditions that are present in real-world applications of such scaffolds. Even if the load is expected to be highly directional, deviation to such orientation may occur, thereby a worst case loading condition must be investigated <ref type="bibr">[40]</ref>. Moreover, since properly arranged heterogeneity and/or inelastic anisotropy at different microstructural length scales can significantly toughen the material <ref type="bibr">[41]</ref><ref type="bibr">[42]</ref><ref type="bibr">[43]</ref><ref type="bibr">[44]</ref>, future investigations could explore the effects of multiple dislocations and their spatial arrangements, further enriching our understanding of the mechanical behavior of these complex microarchitected materials. Inspired by nature, the orientation-dependent stiffness, strength and energy dissipation behavior of the dislocation structure provide significant flexibility for designing bone scaffolds for desired and optimized properties.</p></div></body>
		</text>
</TEI>
