2.1 Sample Preparation

High-purity ruthenium powder (99.995 wt%) with an average particle size of 5 μm provided by the Kunming Institute of Precious Metals, China, was used as the raw material. Ruthenium samples were compacted under a pressure of 40 MPa at 1250°C in a vacuum of 10^–3 Pa at a heating rate of 15°C min^–1 for 0.5 h, 1 h, 2 h, 3 h and 4 h. After sintering, the power was turned off, the tablets were cooled in the furnace to room temperature before being taken out. The prepared samples had the shape of a tablet with a diameter of 30 mm and a height of 4 mm.

2.2 Microstructure and Properties Characterisation

The density of the samples was measured by Archimedes’ method (17). The phase content of the samples was examined with the help of the X’Pert PRO X-ray diffractometer (PANalytical, The Netherlands) and the SmartLab 9 Kw (Rigaku Corporation, Japan) operated by copper K_α irradiation. The morphology of the ruthenium samples and their fracture surfaces were studied on the Sirion 200 field emission scanning electron microscope and the Versa 3D^TM DualBeam^TM (FEI Company, USA) including EBSD. The fracture surface was prepared as follows: firstly, a ruthenium tablet of 10 × 4 × 0.5 mm was cut by electric discharge machining (EDM); then the disc was polished by abrasive paper until the thickness was ~200 μm; finally, the thin slice was bent by hand. A Leica EM TIC 3X ion beam milling system (Leica Microsystems GmbH, Germany) was used to prepare the sample’s surface for EBSD analysis. The Vickers microhardness of the samples was measured with the help of the HXS‐1000A microhardness tester (Shanghai Highwell Optoelectronics Technology Co, Ltd, China) with a load of 100 g.

3.1 Microstructure Characterisation

The density results are shown in Figure 1. It can be seen that the sample density increases with sintering time between ~0.5–2 h, then decreases after 4 h. The maximum density of ruthenium was 12.2 g cm^–3 with a sintering time of 2 h. The XRD spectra of the ruthenium samples are given in Figure 2. Ruthenium powder exhibits random orientation of the particles corresponding to the standard PDF card of ruthenium (PDF #06-0663) (18). The intensities of and peaks decrease sharply, the intensity of (0002) peak increases, while the intensities of (0002) and peaks are the same in the samples sintered for 0.5 h. The samples sintered for 1 h show similar XRD patterns. This indicates that a stable grain structure forms in the ruthenium samples after 1 h. The intensity of peak began to increase after sintering for 2 h and 4 h. The intensity of peak increased again in the sample sintered for 4 h. A stable (0002) texture was formed in the ruthenium samples sintered for 0.5–2 h; however, it disappeared after 4 h. This finding has shown that the texture is controlled by varying the sintering time.

Ruthenium powder is shown in Figure 3. It consists of particles having irregular shapes with sizes from 1 μm to 15 μm (Figure 3(a)). The large particles are aggregates of the small particles. There are many pores or voids in the particles (Figure 3(b)), which were caused by gas released during the chemical reduction of ruthenium. The apparent density and the tap density of ruthenium powder were 1.9 g cm^–3 and 3.2 g cm^–3, respectively.

Morphology of the fracture surfaces of ruthenium samples are given in Figure 4. Their fracture mode is attested as brittle intergranular fracture (19). This finding agrees with the conclusion that ruthenium behaves like a brittle solid even at elevated temperatures (20). This is a puzzling behaviour because a hexagonal close-packed (HCP) metal is a ductile material (21). The slip of dislocations on both prismatic and basal planes could happen in ruthenium single crystals at room temperature (22). However, basal slip is the main deformation mode of ruthenium single crystals at room temperature under tension and their fracture mode is brittle transgranular fracture (23). The grains in the samples are homogeneous having a size of ~4–5 μm, which does not depend on the sintered time.

In order to further reveal the sintering mechanism, EBSD was used to analyse the microstructure. Figure 5 is the grain boundary map of ruthenium. Firstly, it can be seen that the grain size of the ruthenium samples with different sintering time is ~4–5 μm, corresponding well with the fracture surface (Figure 4). The grain interior shows different morphologies with sintering time. For ruthenium sintered for 0.5 h, there are few twins and low-angle grain boundaries (~5–15°) in the grain interior (Figure 5(a)). For ruthenium sintered for 1 h, Figure 5(b) shows clean grain structure with fewer low-angle grain boundaries as compared with Figure 5(a). Twins and low-angle grain boundaries with high density appear inside grains when sintering time reaches 2 h (Figure 5(c)), and then the defect density decreases again in the sample sintered for 4 h. The statistical results of grain boundaries are summarised in Table I. From Table I, it can be seen that the total length of low-angle grain boundaries is only 1.8 μm for ruthenium sintered for 1 h which is the shortest of all the samples. The total length of low-angle grain boundaries is 42 μm for ruthenium sintered for 2 h, which is the longest of all the samples. The total length of boundaries increases from ~0.5–2 h, achieving a maximum value with sintering time of 2 h (263 μm), and then decreases to 141 μm with sintering time of 4 h. For some HCP metals, the activation of twins shows a strong dependence on the deformation temperature and strain rate. During the hot compression of titanium at temperatures from 673 K to 973 K, Zeng et al. found many twins at 723 K and 0.1 s^–1, few twins after deformation at 723 K and 0.01 s^–1 and no twins after deformation at 973 K and 0.01 s^–1 (24).

Ruthenium possesses an HCP lattice, where twinning is the important stress accommodation channel under mechanical loading (21). Experiment has shown that there are four active twin systems: , , and in the ruthenium samples. Figure 6 is the twin boundary map of ruthenium sintered for different times at 1250°C. For ruthenium sintered for 0.5 h, it can be seen there are a few twin systems in the grain interior (Figure 6(a)). Figure 6(b) shows that for ruthenium sintered for 1 h, there are fewer twins in the grain interior, and the percentage of twin systems decreases while the other twin systems of , and increase as compared with Figure 6(a). The total twin density increases again, and the percentage of twin systems increases when sintering time reaches 2 h (Figure 6(c)). The total twin density decreases again with a sintering time of 4 h, although the percentage of twin systems increases to a maximum value of 72.1%. The statistics of twin boundaries in ruthenium samples are summarised in Table II. It can be seen that the longest twin was 42.6 μm for ruthenium sintered for 2 h. The total length of twin boundaries increases from ~0.5–2 h while the total length of twin boundaries achieves a maximum value of 64.3 μm with a sintering time of 2 h, then the value decreases to 44.2 μm with a sintering time of 4 h. In HCP titanium, the ratio of the lattice constants (c:a) is 1.587, which is similar to that of ruthenium (c:a = 1.582) (23). Previous research has shown that deformation twins occur during hot compression of titanium at 723 K and 0.1 s^–1 (24).

Table II

Statistics of Twin Boundary Types in Ruthenium Samples

Time, h
	Length, μm	Percentage, %	Length, μm	Percentage, %	Length, μm	Percentage, %	Length, μm	Percentage, %
0.5	14.4	51.0	6.0	21.3	3.6	12.6	4.3	15.1
1	10.3	30.7	9.3	27.6	6.6	19.7	7.4	22
2	42.6	66.3	7.2	11.2	4.5	7.0	9.9	15.5
4	31.8	72.1	4.9	11.2	3.7	8.3	3.7	8.4

3.2 Hardness Characterisation

To examine the effect of microstructure on the mechanical properties of ruthenium, the Vickers microhardness of the samples was measured. The dependence of the hardness of the ruthenium samples on the sintering time is shown in Figure 7. The hardness increases at first and then it decreases with sintering time. The hardness of ruthenium sintered for 0.5 h was 447.2 HV, and it increased to a maximum hardness of 540.1 HV for ruthenium sintered for 1 h. The hardness decreased to 531.6 HV for ruthenium sintered for 2 h, then decreased to the minimum value of 407.8 HV for ruthenium sintered for 4 h. In a previous study of ruthenium hardness (3) the sintered tablets exhibited a hardness between 91 HV to 377 HV, while after hot working their hardness became 307 HV to 455 HV. The difference in the measurements may be explained by the fact that the samples in (3) were sintered without pressure and, as a result, their density was lower (9.69–11.88 g cm^–3).

3.3 Discussion

It was shown that the average grain size in the ruthenium samples is stable (~4–5 μm) and does not depend on the time of sintering at the process temperature of 1250°C (Figure 4). It seems that sponge particles could recrystallise under annealing at 1250°C for ~0.5–4 h, while this temperature is sufficiently low that the grains in the samples could begin growing during this short time. The initial powder size of materials may impact the grain size of the tablets during high pressure-high temperature sintering. For example, Shin et al. found that during the sintering of diamond there was no abnormal grain growth (AGG) for the initial powder size of 4 μm, while AGG happened for the initial powder size of 2 μm (25). Thus, the particle size of the present initial ruthenium powder may be suitable for the present sintering conditions.

Early in the sintering process, after a sintering time of 0.5 h (Figure 5(a)), the pressure and high temperature caused particle rearrangement, sliding and metallurgical bonding. There were a few twins in some grain interiors suggesting the initial inhomogeneous deformation of ruthenium. The inhomogeneous state may be caused by different grain orientations in which some orientations deform easily or by areas with closely spaced grains and metallurgical bonding which deform first. As the sintering time reached 1 h (Figure 5(b)), the grains showed clean and uniform grain structure. In VHP samples, the particles were pressed together and kept in contact (26). Thus, there were more diffusion paths to promote atomic migration and induce sintering in multiple directions (27). With the appropriate pressure, temperature and holding time, voids were further crushed and ruthenium particles contacted and bonded with each other to form fully metallurgical bonding across grain boundaries. The density of ruthenium also increased slightly from 0.5 h to 1 h.

With the increase of sintering time to 2 h, ruthenium formed metallurgical bonding in almost the whole bulk material (Figure 5(c)). Defects such as twins and low-angle grain boundaries appearing inside grains point to some plasticity in the ruthenium samples. The total length of grain boundaries and twin boundaries reached a maximum value for a sintering time of 2 h, implying the maximum plastic deformation for the present ruthenium sample. Thus, the density also reached a maximum value after sintering for 2 h. As sintering time further increased to 4 h (Figure 5(d)), there were fewer defects and boundaries inside the grains as compared with ruthenium sintered for 2 h, and the density also reached a minimum value in all samples sintered between ~0.5–4 h.

It is well known that annealing can induce crystal formation from such defects as dislocations and twins. Hence, the decreasing hardness of the samples sintered for 4 h could be caused by the annealing of twins which appeared in the material at earlier stages of sintering. During VHP, the ruthenium powders sustained the crushing of voids, formation of grain boundaries, grain growth, plastic deformation (formation of defects) and recovery of defects. As for the hardness of ruthenium with sintering time (Figure 6), the hardness increased first from 0.5 h to 1 h due to the crushing of voids and formation of tight boundaries. It was found that the maximum hardness was achieved for ruthenium sintered for 1 h, while the density reached its maximum value for ruthenium sintered for 2 h. According to Figure 1 and Figure 7, the samples’ density begins decreasing when sintered for 4 h. The hardness is prone to similar behaviour after processing for more than 0.5 h. This is normal behaviour for hardness because the twins’ density decreases after 4 h. It can be seen from the XRD pattern of ruthenium that the strongest peak is (0002) plane for ruthenium sintered for 1 h, while the strongest peak is plane for ruthenium sintered for 2 h (Figure 1). Since there was little difference in density and average grain size for the ruthenium samples sintered for 1 h and 2 h, the crystal orientation and defect density of ruthenium may play a role in the determination of hardness.

Ruthenium is a brittle metal due to its anisotropic HCP crystal structure that provides a limited number of independent slip systems, and twinning is an important deformation mode. Twinning has been observed in some metals and ceramics such as tungsten carbide, cubic boron nitrides and aluminium oxynitride ceramic during high-pressure and high-temperature sintering (28–30). Previous research has shown that deformation of ruthenium occurs by slipping on and twinning on , , and occasionally (3). In the present work, slipping and twinning occurred during VHP of ruthenium even at 1250°C, and the main twin system was . The transition of the main twin system to may be attributed to the high temperature and pressure. Any HCP structure is anisotropic in comparison with a cubic structure. As a result, the total plasticity of an HCP-metallic single-crystal depends on its crystallographic orientation. It is significant for zinc and cadmium at room temperature because only the basal slip is active in these metals under these conditions. The orientation anisotropy of plasticity in titanium and zirconium is not so visible, insomuch as the prismatic slip’s contribution is added to the basal slip. The contribution of twinning to the total plasticity of an HCP-metal is minor, while twinning could influence its hardness and, perhaps, work-hardening. Also, it should be noted that a lot of twin systems exist in HCP-metals, which are known to be ductile and malleable materials. The low malleability of ruthenium is connected with the low cohesive strength of grain boundaries. In the case of PM ruthenium, it is due to brittle intergranular fracture, whose source is likely to be non-metallic impurities rather than its intrinsic properties. Further work still needs to be done to understand the relationship between impurity elements and the plasticity of ruthenium.