Fig. 1. Stress-strain curve recorded during continuous straining plotted in terms of the diametral strain recorded at the centre of the sample. Necking initiated at 0.5 diametral strain. The inset shows the geometry and dimensions of the test specimen with the field of view (FoV) of the CT scan indicated by the red box.

Fig. 2. 3D visualisation of the synchrotron CT scan undertaken at a diametral strain of 2.83 showing the void population coloured according to their Euclidean distance to the centre of the sample (indicated by a + marker). The two dashed black boxes show two regions of interest containing a large void (inclusion 1) and the central region of the sample respectively which are tracked in Figs. 3 and 5.

Fig. 3. X-ray CT volume rendering showing growth of a typical large void (blue) around a 10 µm inclusion (black), the location of which is indicated in Fig. 2 as a function of diametral strain (see also Supplementary ‘Video 1′ for an animation showing many more frames in the sequence). The loading direction is vertical.

Fig. 4. (a, b) Evolution of size and shape of a representative set of the void population nucleating from large inclusions (inclusion 1, as shown in Fig. 3, corresponds to the black triangle). At zero strain the sizes of the voids are essentially given by the size of the inclusions. The equivalent diameter is calculated from the total volume of the void including the inclusion.

Fig. 5. X-ray CT volume rendering of the voids occurring as a function of strain in the high triaxiality region in the centre of the sample (location of this region is indicated in Fig. 2). At large diametral strains (εd>1.8), there is a sudden and significant appearance of many small voids in this region. (See also Supplementary ‘Video 2′ for an animation showing many more frames in the sequence).

Fig. 6. (a) Size and (b) shape of 6 individual voids representative of the region tracked in Fig. 5. The strain where coalescence took place was indicated by a dashed circle in (a) and the inset shows a histogram of void size in the volume at the onset of coalescence.

Fig. 7. (a-c) Longitudinal virtual cross sections of the same slice from the X-ray CT data with increasing strain demonstrating the process of void coalescence as a function of strain. (d) Void volume fraction and equivalent strain vs. distance from the central slice (the smallest cross-sectional area) corresponding to when the sample was strained to εd = 2.83 diametral strain.

Fig. 8. (a) SEM overview of the fracture surface. (b) Small dimples surrounding two deep large voids, visible in the top-middle of part of (a). (c) A high magnification view showing small voids with carbides 100 nm in size. (d) Region of coalescence of small voids related to internal crack formation.

Fig. 9. Volume rendering of X-ray CT data demonstrating different void growth behaviours from two types of tension specimen showing internal voids in red: (a) relatively smooth sided (10 mm radius notch) and (b) more severely notched (2 mm radius). The volume rendered CT images represent cores extracted from the centre of the fractured sample where the top of both samples is the fracture surface. Distinctly different void growth behaviours can be seen. Both samples are 500 µm in diameter.

Fig. 10. SEM images showing the fracture surface of a compact tension specimen [23]. (a) Evidence of dilatational growth around ～10 µm particles. (b) Evidence of dilatational growth around ～4 µm particles.