**4. Discussion**

The proof-of-concept experiment reported herein enabled important understanding of the SR X-ray radiation-induced damage to the integrity of bone tissue. The residual strain accumulation caused by SR X-ray radiation was quantified for the first time using DVC applied to in situ SR-microCT images, and the effect of the environmental temperature on the SR irradiation-induced damage in bone tissue was addressed. It is known that irradiation has a deleterious effect on the structural and mechanical properties of bone as a result of collagen matrix degradation due to the formation of collagen cross-links and eventual rupture of the collagen fibres [20,27]. Several studies have addressed the effect of high-energy SR X-ray radiation on the mechanical properties of bone [20,21,26,48], and safe dose values (35 kGrays) were defined to preserve bone mechanics [20]. However, during in situ SR-microCT studies, a reduction of the dose is related to a reduction in the total exposure to SR radiation and, therefore, the signal-to-noise ratio of the acquired tomograms, with a consequent reduction in image quality and increased DVC errors [14]. Therefore, new protocols need to be defined in order to preserve bone tissue while maintaining good image quality. Furthermore, whether bone integrity can be preserved by controlling the temperature during in situ SR experiments still remains unexplored.

The overall change in temperature during image acquisition was minimal (ΔT = 0.4 ◦C) (Figure 2a) and in line with previous reports on SR beam heating [49,50]. Wallander and Wallentin [51] showed that X-ray-induced heating can lead to significant temperature increase (i.e., nanowire at 8 ◦C above room temperature) at typical synchrotron beamline fluxes. As a strategy for reducing the X-ray-induced heating, it was suggested to improve the heat transfer of the target material to the surroundings, for example, by immersing the samples in liquid [51]. However, it still remains unclear whether that thermal gradient in a very short period of time (opening/closing of the beam shutter) may induce collagen degradation. As specimens were held between the loading stage platens during the in situ test, the effect of the X-ray beam on the stress-relaxation behaviour of the specimens could be observed (Figure 2b), similar to the data reported in [52]. With only a fixed preload applied, an increase in the load was identified for the trabecular bone specimen at each cyclic period that corresponded with the opening of the X-ray shutter. Both trabecular and compact bone exhibit a highly viscoelastic behaviour; however, this is more evident for trabecular bone due to the large content of bone marrow in its cavities. Thus, the loadcell of the loading stage was not accurate enough to capture any changes in the stress-relaxation behaviour for the compact bone specimen. Heat causes a transformation of the collagen molecule, known as the collagen shrinkage phenomenon [53], whereby the collagen molecule develops a contractile force that is held constant [54,55] at a given temperature (shrinkage temperature). This shrinkage behaviour is related to the cross-links in the collagen and its stability [53]. Even though the specimens in the current study were kept at a constant temperature (~23 ◦C), the beam-induced temperature rise of 0.4 ◦C may contribute to the activation of a similar contractile force, which is a clear indicator of the harmful effects of the SR irradiation on bone tissue.

The results obtained from the current study have shown that reducing the temperature to 0 ◦C notably reduced the irradiation-induced microdamage and residual strain in trabecular bone specimens (Figure 6). However, minimal effect was observed for compact bone (Figure 3). Nguyen et al. [30] reported that the mechanical properties of compact bone were decreased by a lower dose than that affecting trabecular bone. However, it has been shown here (Figure 3) that the structural integrity of compact bone tissue was not compromised, as microcracks were not detected as in the trabecular bone tissue. In any case, specimens were not mechanically tested; thus, whether the regions of high strain concentration found in compact bone (Figure 3) influence the mechanical properties is still unknown. Furthermore, Peña Fernández et al. [14] showed that the presence of microcracks was not always related to changes in the apparent elastic properties of the irradiated bone.

Although the overall residual strain in compact bone imaged at room temperature was low, with peak strain values below 1000 με for <sup>ε</sup>p<sup>1</sup> (Figure 3), some highly strained regions were identified in close proximity to Haversian and Volkmann's canals (Figure 4). Canals and osteocyte lacunae are known to act as stress concentrating features in specimens subjected to mechanical load [6,12]; however, the effect of irradiation on these specific sites has never been considered. Haversian canals contain unbound water [56], and as ionising radiation produces the release of free radicals via radiolysis of water molecules [29], it is expected that a larger number of free radicals, which could interact with the collagen and induce cross-linking reactions, are found in proximity to the canals due to the higher water content.

Lowering the environmental temperature to 0 ◦C had a positive effect on the DVC-measured residual strain in trabecular bone, which showed a peak principal strain value below 1000 με (Figure 5); furthermore, no microdamage was visually detected on the reconstructed tomograms. These results are consistent with medical studies on the effect of gamma irradiation, where it was shown that irradiating bone specimens while frozen did not affect the mechanical properties of bone [31,34]. In fact, decreasing the temperature reduces the mobility of the water, and, therefore, decreases the mobility of highly reactive oxygen free radicals produced by high-energy X-ray radiation. Impairing that mobility protects the collagen by reducing cross-linking reactions within its molecules [57,58]. The effect of freezing on the mechanical properties of bone has been previously studied [59–62] and no statistical differences were found after freezing, nor after several freeze-thaw cycles [63,64]. It should be noted that, during the proposed experiment, specimens were immersed in saline solution at 0 ◦C, and ice crystals, which may cause structural damage to the tissue [63], were not observed at any stage of the experiment.

The irradiation-induced damage in the trabecular bone imaged at room temperature resulted in microcracks that were visible in the tissue even if the specimen was not subjected to any mechanical load. At the nanoscale, SR irradiation-induced free radical attack of the collagen network results in a cross-linking reaction that degrades the structural integrity of the collagen fibres [20,29,30]. Previous studies using atomic force microscopy have shown that crack formation and bone fracture occur between the mineralised collagen fibrils. Fantner et al. [65] proposed that the mineralised collagen fibres are held together by a nonfibrillar organic matrix that acts as a glue. The glue resists the separation of the mineralised collagen fibrils, avoiding the formation of cracks, when a load is applied to the bone. During the formation of microcracks, work that stretches the glue molecules

would be required to separate the mineralised collagen fibrils. Irradiation may affect that mechanism by damaging the sacrificial bonds, as a result of the observed shrinkage behaviour, which could lead to the rupture of those bonds after prolonged exposure to irradiation and consequent microcrack formation. At the macroscale, DVC-computed displacements (Figure 7) sugges<sup>t</sup> a vortex motion around the microcracked region, which results in a shrinkage process of the material and the formation of a microcrack that follows an unusual pattern in fracture mechanics. The denaturation of the collagen may not be homogeneous throughout the bone tissue; therefore, crack propagation would follow the degeneration process of the collagen.

DVC was successfully used to understand crack formation and propagation in bone. Christen et al. [12] investigated the initiation and propagation of microcracks in cortical bone using DVC; however, full-field displacements and strains were only evaluated in terms of magnitude, but the directions were not explored. Additionally, specimens were pre-cracked before mechanical testing; thus, crack initiation and propagation was expected around the notch region. In this study, microcracks were not induced by mechanical loading, but by SR irradiation instead. Discontinuities in the displacement field (Figure 8a) corresponded to high-orientation changes in the strain field (Figure 9a) that could indicate crack formation. Furthermore, perpendicularity of displacement (Figure 8b) and principal strains (Figure 9b) to the crack might be related to a crack propagation front. Similar crack formation mechanisms were observed in clay deformation using digital image correlation (DIC) following desiccation [66,67]. Like the results herein reported, in opening mode, the direction of the crack was perpendicular to that of <sup>ε</sup>p1, whereas, for cracks in mixed opening-sliding mode, <sup>ε</sup>p<sup>1</sup> was found parallel to the direction of the crack (Figure 9b). Those studies [66,67] concluded that cracks formed a network which is found after thermal shocks, and the authors emphasized the need to develop a multiscale approach to better understand crack formation and propagation. Similar to those findings, irradiation-induced microcracks need to be further investigated at different dimensional levels to properly understand the formation mechanisms.

This study has some limitations. First, only one specimen per bone type was tested at each temperature, and the mechanical properties of the bone specimens were not evaluated after irradiation. Residual strain maps suggested that a decrease in the temperature had a beneficial effect on preserving bone integrity and mechanics, but specimens were maintained far below physiological conditions (~37 ◦C); thus, it could be argued that the mechanical properties of bone tissue could have been altered. Further analyses are needed to properly assess the effect of the environmental temperature during in situ SR-microCT experiments, translating the findings of the proposed methodology to in situ SR-microCT bone mechanics. Moreover, a combination of techniques at different dimensional scales would enhance the knowledge of the irradiation-induced damage in bone tissue.
