The SEM images of interface characteristics at different bonding temperatures are shown in Fig. 2, and the higher magnification images reveal the interfacial void morphology. As seen from Fig. 2(a), the long interfacial voids with scraggly edges separated by metallurgically joined areas are clearly visible at a temperature of 1000 ° C. As the temperature increases to 1020 ° C, it is obvious by comparing Fig. 2(a) and (b) that the void size decreases, and the edges of interfacial voids tend to be smooth. To further increase temperature to 1050 ° C, the long voids disappear completely and the joined areas increase significantly, as shown in Fig. 2(c). When joined at a temperature of 1080 ° C, there are occasional isolated ellipse voids in the bonding interface, as shown in Fig. 2(d). As the temperature increases up to 1100 ° C, the interfacial voids, which are round in shape and small in volume, can be seldom found from Fig. 2(e), and the EDS is used to study any compositional variations in the bonding interface compared with those of base material. The EDS profiles across the bonding interface show that the composition distributions of Fe, Cr, W, Mo, V, C and Ni are virtually constant, and no oxides and carbides exist in the bonding interface. Not only qualitative but also quantitative assessment of interfacial voids can be obtained in this study. The effect of bonding temperature on the interface bonding ratio is shown in Fig. 3. It is apparent that the interface bonding ratio increases rapidly from ~48% to 99% as the temperature increases from 1000 ° C to 1100 ° C. It is traditionally believed that it is difficult to produce a joint with 100% interface bonding ratio. Thus, the highest interface bonding ratio approaching ~100% confirms that a sound joint with few defects have been obtained.

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. SEM images of the interface characteristics at different bonding temperatures: (a) 1000 ° C, (b) 1020 ° C, (c) 1050 ° C, (d) 1080 ° C, (e) 1100 ° C; (f) EDS profiles across the bonding interface.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. Interface bonding ratio at different bonding temperatures.

During diffusion bonding, the void shrinkage as an important process for achieving high quality joints is generally ascribed to several physical mechanisms^{[17, 24]}: (1) plastic flow of materials around voids, including plastic deformation and creep deformation; (2) atomic diffusion, including surface diffusion, interface diffusion and volume diffusion. An increase in temperature will reduce the material yield strength, enhancing the plastic flow of materials, thereby forcing more mass from adjacent regions into voids. Simultaneously, void shrinkage is further accelerated by increased atomic diffusion. The diffusion coefficient D (m²/s) is a material property and is the most useful parameter for characterizing the capability of atomic diffusion, which is, generally, described as follows^[25]:

D=D₀exp(-Q/RT)(1)

where D₀ (m²/s) is the diffusion constant, Q (J/mol) is the activation energy for the atomic diffusion, R (J/(mol K)) is the gas constant, and T (K) is the temperature. It is apparent from Eq. (1) that the diffusion coefficient is a strong function of temperature, and an increase in temperature enhances the capability of atomic diffusion and will, in turn, accelerate void shrinkage.

Based on the well-defined interfacial voids presented above, the evolution mechanism of interfacial voids can be analyzed in detail. The actual specimen surface is rather irregular as a result of the machining process. It can be seen from Fig. 4 that the oscillogram of the surface roughness shows a bimodal distribution of asperities, which means that short wavelength asperities are superimposed on long wavelength asperities. After assembling two surfaces abutting each other, only a few long wavelength asperities contact to form large scraggly voids, and many convex and concave surfaces are superimposed on the void surface, as schematically shown in Fig. 5(a). As the bonding pressure is applied to the joined specimen, the deformation of contacted asperities will decrease the void height, resulting in the contact of some close convex surfaces on the opposite sides within one void, and this could subsequently divide the large void into two or more voids, as schematically shown in Fig. 5(b). However, as the joined areas increase to a certain stage, the plastic flow of materials cannot make the residual distant convex surfaces on the opposite sides within one void contact with each other. In this case, the surface diffusion would operate to smooth out the irregularly shaped voids. The details are that the atoms beneath a curved surface have their chemical potentials altered by the curvature of the surface, and this difference in chemical potential will drive the diffusional flux of atoms to reduce the free energy of the system. The convex surface has a higher atomic concentration than the concave surface, resulting in a difference in the chemical potential between convex surface and concave surface. Thus, the difference in the chemical potential will drive the diffusional flux of atoms from convex to concave region. This atomic diffusion process will flatten convex and concave surfaces, and then transform the original scraggly voids to smooth voids. Mullins et al.^[26] stated that the surface diffusion has a distance limit in a definite time. Therefore, the short scraggly void would be easily transformed into an elliptic void under the action of surface diffusion, while the long scraggly void would be transformed into a penny-shaped void, as schematically shown in Fig. 5(c). Surface diffusion only alters the void shape, but not the void volume^[27]. The reduction of void volume is chiefly attributed to the combined effects of plastic flow of materials around voids, interface diffusion and volume diffusion. As the bonding process proceeds, the volume of penny-shaped void decreases rapidly. Then, the penny-shaped void will also be transformed into an elliptic void, as schematically shown in Fig. 5(d). When the void becomes an ellipse void, the curvature of void top (point B) is greater than that of void neck (point A). This difference in curvature will drive the diffusional flux of atoms from point B to point A, resulting in the formation of rounded void with uniform curvature, as schematically shown in Fig. 5(e). The formation of rounded voids is generally considered to be a complete finish of the evolution of interfacial void morphology, because it has the most stable morphology and the lowest system free energy relative to any other morphology.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. Oscillogram of the surface roughness for 1Cr11Ni2W2MoV stainless steel specimens.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. Schematic evolution process of interfacial voids.

The effect of bonding temperature on the microhardness of joint and base material is shown in Fig. 6(a). It is apparent that the average microhardness of joint increases from ~400 HV_0.3 to 466 HV_0.3 and the scatter of the measured microhardness of joint decreases with increasing temperature from 1000 ° C to 1100 ° C. The obvious scatter of the measured microhardness of joint at low temperature should chiefly be attributed to the arbitrary variations in void distribution and void size resulted from the original irregular surfaces. When joined at 1100 ° C, a sound joint with average microhardness close to that of base material can be obtained and the scatter of the microhardness of joint cannot be found.

Furthermore, the lap shear test as the most convenient and frequently mechanical testing method was adopted to evaluate the strength of joint. In general, the absolute strength of joint could not be used to evaluate the extent of bonding because as-joined specimens having different thermal histories would have different shear strengths. The specific shear strength (μ μ ) is, thus, used to assess the shear strength of joint, and is defined as follows:

μ =τ _db/τ _b× 100%(2)

where τ _db (MPa) is the shear strength of joint, and τ _b (MPa) is the shear strength of base material at the same bonding condition.

The specific shear strength obtained at various bonding temperatures is given in Fig. 6(b), together with the shear strength of joint and base material. It can be found that the joint exhibited a general increase in the specific shear strength with increasing temperature up to the point at which the shear strength of joint can match that of base material. Islam and Ridley^[28] studied the diffusion bonding of microduplex stainless steel, and pointed out that a joint with specific shear strength of µ > 95% can be regarded as a sound joint. This indicates that a high strength joint has been obtained in the present work. In addition, it can also be seen from Fig. 6 that the mechanical properties of base material is slightly decreased as the bonding temperature increases from 1000 ° C to 1100 ° C. This decrease of mechanical properties of base material is attributed to the grain growth, but is so slight that is involved in an acceptable region.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. Effect of bonding temperature on microhardness (a) and shear strength (b) of both joint and base material.

A further assessment of joint is obtained by SEM examination of the fracture surfaces of shear test specimens, as shown in Fig. 7. It is apparent that the microscopic fracture surfaces of poor joints produced at low bonding temperature is featured by a large number of dark regions with flat plane as seen at a higher magnification in Fig. 8(a), identified as un-joined regions. As the bonding temperature increases, the bright regions with ductile dimples on the microscopic fracture surfaces as seen at a higher magnification in Fig. 8(b), identified as joined regions, significantly increase. When joined at 1100 ° C, the microscopic fracture surface exhibits a fully ductile failure, as shown in Fig. 7(e). Cá ceres and Selling^[29] proposed that the decrease in strength correlated with the area fraction of defects (the defects are mainly voids in this case) on the fracture surface. Thus, the dimples characteristic of ductile fracture over the entire surface implies few defects existing in the fracture surface, indicating that a high quality joint has been obtained.

Fig. 7.

	Figure Option ViewDownloadNew Window
	Fig. 7. Shear fracture surfaces of joints at different bonding temperatures: (a) 1000 ° C, (b) 1020 ° C, (c) 1050 ° C, (d) 1080 ° C, (e) 1100 ° C.

Fig. 8.

	Figure Option ViewDownloadNew Window
	Fig. 8. Dark regions (a) and bright regions (b) of the shear fracture surfaces of joint at a higher magnification.

The variation of the mechanical properties of joint is certainly attributed to the evolution of interfacial voids. The remaining voids distributed in the bonding interface have a detrimental effect on the mechanical properties of joint. Based on the detailed interfacial void morphology and mechanical properties of joints presented above, it is apparent that an increase in temperature accelerates the void shrinkage, resulting in an increase in interface bonding ratio, thereby obviously enhancing the mechanical properties of joint.