Measurements of physiological parameters such as pulse rate, voice, and motion for precise health care monitoring requires highly sensitive sensors. Flexible strain gauges are useful sensors that can be used in human health care devices. In this study, we propose a crack-based strain gauge fabricated by fused deposition modeling (FDM)-based three-dimensional (3D)-printing. The strain gauge combined a 3D-printed thermoplastic polyurethane layer and a platinum layer as the flexible substrate and conductive layer, respectively. Through a layer-by-layer deposition process, self-aligned crack arrays were easily formed along the groove patterns resulting from stress concentration during stretching motions. Strain gauges with a 200-µm printing thickness exhibited the most sensitive performance (~442% increase in gauge factor compared with that of a flat sensor) and the fastest recovery time (~99% decrease in recovery time compared with that of a flat sensor). In addition, 500 cycling tests were conducted to demonstrate the reliability of the sensor. Finally, various applications of the strain gauge as wearable devices used to monitor human health and motion were demonstrated. These results support the facile fabrication of sensitive strain gauges for the development of smart devices by additive manufacturing.
Because the GF and recovery time improved as the printing resolution (thickness) decreased, FSGs printed with a resolution below 200 μm showed enhanced GF and recovery time. To investigate the minimum printable resolution, a TPU substrate was printed with 100-μm resolution and compared with other substrates, as shown in Fig. 5b. Compared to the uniform and regular groove patterns seen on substrates printed with 200, 300, and 400-μm resolution, the TPU substrate with 100-μm resolution showed an irregular surface profile. In addition, a comparison of top-view images for the 100-, 200-, 300-, and 400-μm TPU substrates showed incomplete features on the 100-μm TPU substrate (see Figure S1(a) in Supplementary Information), which indicated unstable printing resolution in the FDM printing method used to generate crack-based FSGs. Although the digital light processing (DLP) printing method was used to manufacture a substrate with a smaller printing resolution (25 μm), the substrate was not flexible and showed a relatively smooth surface without groove patterns due to the characteristics of the DLP method, as shown in Figure S1(b); hence, DLP was not appropriate for creating FSGs. Therefore, the minimum printable resolution of 200 μm was selected as the optimal condition for sensitive strain sensors and further analysis, including application examples.
Table 1 shows a comparison of device characteristics, including materials, fabrication method, recovery time, GF, and corresponding strain range, applicable to the present work and previous studies. Compared to chemical synthesis3,4,13 and photolithography16,17, the formation of self-aligned cracks on a flexible substrate was more easily realized using a 3D printing method without complicated manufacturing processes. In addition, the use of 3D printing technology has the advantage of directly printing customized 3D-shaped substrates, which can be used for finger gloves, wrist bands, and masks where curved surfaces are required, rather than simple rectangular and planar substrates2,7,15,20. Although a high GF (~2000) in a small strain range was achieved through advanced prestretching3, it was noteworthy that the crack-based strain gauges developed in this study exhibited sufficiently high GFs with faster recovery times compared to other devices1,5,16, thus supporting real-time monitoring applications.
As such, computational models are desired for fatigue crack nucleation and propagation that alleviates the complexity of re-meshing and can track the crack tip in complex microstructures, while at the same time can be efficiently implemented in an efficient computational framework. The phase field model technique in conjunction with isogeometric analysis that utilizes the geometric model, can provide the solution which does not require any criteria for crack initiation and propagation under random spectrum loading including environmental effects.
Phase field model development will be required in order to link spatial and temporal evolution of complex crack patterns to the external applied load by utilizing finite element and iso-geometric analysis (IGA) discretization methods. Starting with an initial 2D analysis, the PFM model has to describe the complex phenomena of 3D crack evolution at a microscale as well as the final fracture at the macroscale. The proposed PFM model may include an appropriate plasticity model to study load interactions occurring in complicated loading situations such as variable amplitude loading. Finite element based numerical implementations of the PFM crack propagation under dynamic loading is desirable. Furthermore, the application of PFM to dynamic ductile fracture needs to be further explored, addressing the limitations and assumptions and enhancements as needed.
Bone quality is an important concept to explain bone fragility in addition to bone mass. Among bone quality factors, microdamage which appears in daily life is thought to have a marked impact on bone strength and plays a major role in the repair process. The starting point for all studies designed to further our understanding of how bone microdamage initiate or dissipate energy, or to investigate the impact of age, gender or disease, remains reliable observation and measurement of microdamage. In this study, 3D Synchrotron Radiation (SR) micro-CT at the micrometric scale was coupled to image analysis for the three-dimensional characterization of bone microdamage in human trabecular bone specimens taken from femoral heads. Specimens were imaged by 3D SR micro-CT with a voxel size of 1.4 µm. A new tailored 3D image analysis technique was developed to segment and quantify microcracks. Microcracks from human trabecular bone were observed in different tomographic sections as well as from 3D renderings. New 3D quantitative measurements on the microcrack density and morphology are reported on five specimens. The 3D microcrack density was found between 3.1 and 9.4/mm3 corresponding to a 2D density between 0.55 and 0.76 /mm2. The microcrack length and width measured in 3D on five selected microcrack ranged respectively from 164 µm to 209 µm and 100 µm to 120 µm. This is the first time that various microcracks in unloaded human trabecular bone - from the simplest linear crack to more complex cross-hatch cracks - have been examined and quantified by 3D imaging at this scale. The suspected complex morphology of microcracks is here considerably more evident than in the 2D observations. In conclusion, this technique opens new perspective for the 3D investigation of microcracks and the impact of age, disease or treatment.
Microdamage presents in the form of microcracks, whose size, morphology and localization are strongly related to the mechanical loading applied to bone , , . Microcracks are usually classified into four types : linear, parallel, cross-hatch and diffuse , . Initiation and propagation of microcracks have mostly been studied in cortical bone after applying ex-vivo mechanical loading. Four toughening mechanisms of bone tissue have been described: plastic deformation, bridging , creation of non-connected small linear microcracks and deflecting cracks with microstructure interfaces (cement lines) , . In addition to dissipating energy, microdamage is also hypothesized to drive bone remodeling by sending stimuli to osteocytes and plays a major role in the repair process. In particular, a decline in osteocyte lacunar density has been shown to be associated with an accumulation of microcracks .
Although many studies have been conducted in animal models, recent data have been reported on human bone specimens , , , . Less attention has been devoted to microcracks in trabecular bone ,  in the past, but this topic is now becoming the subject of growing interest . While the porosity and the anisotropy of trabecular bone tissue make it difficult to work with, studies on trabecular bone allow the analysis of microarchitecture simultaneously with microdamaging , , , .
The most widespread technique for in vitro investigation of microdamage consists in observing thin slices of bone by microscopy after staining, which requires histological sectioning of bone specimens , . Scanning Electron Microscopy has a much smaller field of view and is often used as a complement to observe a small number of microcracks  at very high resolution.
Microcracks are generally described as thin planar ellipsoids whose thickness is of the order of the micrometer . They are usually counted and measured manually, which is likely to be associated with some variability. In addition, two-dimensional (2D) sections provide incomplete information about the complex three-dimensional morphology and size of microcracks. Also, because microcracks can be relatively scarce, the measure of their density in cross-sections can be very sensitive to sampling effects . Although 2D measurements can be extrapolated to three-dimensional (3D) measurements by using statistical models , , such models require prior knowledge and may not necessarily be valid for an individual microcrack. Ideally, bone microcracks should therefore be observed, analyzed and measured in three dimensions with isotropic and sufficiently high spatial resolution.
3D observations of microcracks in cortical bone were obtained by reconstruction of microscopic images after serial sectioning , . Confocal microscopy can also produce micrometric 3D images of microdamage , ,  but with an anisotropic resolution and a small depth (typically 200 µm). Finally, contrast agents are being developed for 3D observations with standard micro-CT devices , , . Even if this technique is able to detect the presence, spatial location, and accumulation of microdamage, the spatial resolution of these images and the capacity of contrast agents to specifically bind to microcracks are still not sufficient to provide relevant 3D data on microcrack morphology. 2b1af7f3a8