The Interface Zone of Explosively Welded Titanium/Steel after Short-Term Heat Treatment

As Ti is characterized by the high reactivity with oxygen and nitrogen at elevated temperatures, the explosive welding (jetting phenomenon) serves as the optimal solution for its welding with steel. However, the extreme conditions of explosive welding and the differences in the physical properties of Ti and steel result in the presence of large residual stresses within the interface region, thus influencing the durability and the mechanical properties of the weld.[9] Therefore, the importance of the microstructure observations should not be omitted. The microstructure of the weld with a characteristic wavy interface obtained in this study is shown in Figures 1(a) and (b). The melted regions (an example is shown by the arrow in Figure 2(a)), where strong intermixing of the bonded materials took place as revealed by elemental map distribution (Figures 2(b) through (d)), are located on both sides of the wave crest mostly in wave vortexes surrounded by steel. They are, however, not that common in the examined weld. Cracks were observed in the vortex region resulting from the shrinkage of melt during the solidification without further propagation to the neighboring areas (Figure 2(a)). Based on the chemical point analysis, the composition inside the melted regions (such as one showed with arrow in Figure 2(a)) corresponded to the stoichiometry of FeTi and Fe₂Ti intermetallic phases, as was previously shown in References 1, 2, and 4. As a result of explosive welding, the area of the material near the bonding interface has highly deformed grains and is strongly strengthened, which can be especially observed for steel grains in Figure 1(a). These grains are elongated along the explosive direction, which is in agreement also with the previous reports (see, for example, Reference 6). The heat treatment was applied to reduce the internal stresses by inducing the recovery and recrystallization processes. As mentioned previously, such annealing can lead to the growth of the intermetallic phases and Ti carbides at the interface of the joined materials and, therefore, results in deterioration of the mechanical properties. Nevertheless, SEM/EDS examination did not reveal such a phenomenon under the applied annealing conditions in the areas located between the waves, as shown in Figure 1(c). Thus, further inspection of these interface regions was performed using TEM.

Figures 3 and 4 show two sets of TEM data collected for the Ti/steel interface after EXW and heat treatment probed from two different places located between the waves. In the first region (Figure 3), surprisingly, the analyzed interface had a form of a remelted layer with thickness even up to 1 μm, which has broadened underneath the surface of the interface. It was composed of nanocrystalline TiC and α-Fe as evidenced by the electron diffraction shown in Figure 3 (zone marked “Z” in Figure 3). In contrast to this, in the second investigated area, the transition zone was only of 350 to 400 nm in thickness (Figure 4) and contained several sublayers. Their identification was done based on the electron diffractions and high-resolution images presented in Figure 4. The obtained patterns were indexed using crystallographic information found in the Inorganic Materials Database.[10] Next to the steel base material (SAED of α-Fe marked “1” in Figure 4), the layer of the FeTi intermetallic compound was identified with the cubic structure and lattice parameter of a = 0.2972 nm (“2” in Figure 4). A subsequent layer with amorphous-like contrast was found to be composed of nanocrystals of TiC cubic phase (a = 0.4327), as confirmed by electron diffraction (“3” in Figure 4).

The next layer, when moving toward the Ti (α-Ti—SAED denoted “6” in Figure 4), was composed of the grains larger in size, with the diffraction contrast visible during bright-field observations. The electron diffraction for this region indicated that the layer also consisted of TiC particles (“4” in Figure 4). Finally, the last sublayer, located in the neighborhood of Ti, formed larger grains of TiC phase up to 100 nm in size. The example of a high-resolution image collected from one such grain is presented in Figure 4 (marked “5”). As is shown, the thickness of the remelted layer was not constant and varied along the joint depending from which place the lamella had been taken (Figure 3vs Figure 4). However, it should be emphasized that Figure 4 can be considered as the representation of the Ti/steel interface region, as such type of interface prevails. Therefore, the scheme of the microstructure observed in the thinnest parts of the bonding zone can be schematically drawn, as shown in Figure 5. The performed analysis revealed that two phases are formed between α-Fe and α-Ti plates in this case: nanocrystalline FeTi intermetallic phase next to the steel base material and predominant nanocrystalline TiC carbide as a subsequent layer, which coarsens toward the α-Ti plate. This coarsening is not gradual but, rather, sharp interfaces could be distinguished between the various TiC sublayers. It should also be noticed that both types of formed phases, i.e., FeTi intermetallic phase and carbide, have a cubic structure. The lattice mismatch between the structure of the base plate (α-Fe, a = 0.2866 nm) and the subsequently formed FeTi phase (a = 0.2978 nm) is ~ 4 pct, that between FeTi and cubic TiC (a = 0.43227 nm) is ~ 45 pct, and that between TiC and α-Ti (a = 0.29486 nm, c = 0.467 nm) is ~ 9 pct (if the c parameter of α-Ti is considered). Internal stresses induced by the lattice mismatch may also influence the properties of the bond formed between two such plates. Formation of carbides at the interface is generally related to the diffusion of carbon in the steel plate toward Ti being a result of the EXW process and subsequent heat treatment.[4,11] It was clearly evidenced by Haitao et al.[11] who examined explosively welded Ti/low-carbon steel (Q235B) being subjected to different rolling processes in the temperature range 700 to 850 °C. It is even more pronounced in the case of a sufficiently higher carbon content in the steel.[12] In the present study, the increased carbon concentration near the Fe/Ti interface was evidenced especially within the melted areas (see the map of element distribution presented in Figure 2(d)). Wachowski et al.,[4] who examined Ti Gr. 1/S355J2 + N steel after the same short heat treatment as the one used in this work (1.5 hours at 600 °C), also observed decarburization of the areas adjacent to the weld due to the diffusion of carbon toward Ti. In the analyzed area reported in Reference 4, fine α-Fe precipitates and TiC were found to grow toward the Ti direction with a thickness of approximately 3.5 µm. This is consistent with the observations made for the interface located between the waves with a higher width of remelted layer (up to 1 µm) found in the present study (Figure 3) but not with the thinner one (350 to 400 µm in width). The authors in Reference 4 indicated that two effects favor the carbide formation instead of intermetallic phases. First is the Ti/C ratio of raw materials. Nevertheless, the studies of the clad plates reported in Reference 4 concerned steel with a similar carbon content (0.19 wt pct in References 4 and 13) as that used in this work (0.16 wt pct), yet they gave different results despite the same heat treatment being applied. Thus, the differences in phase composition of the interlayer in the current work and that found in References 4 and 13 may be associated with the localization of the thin lamella investigated with TEM. In Reference 4, the lamella was cut from the crest of the wave, where a thin melted layer can be seen with SEM, whereas in Reference 13, there is no information on the localization of FIB lamella. It can be speculated whether the difference between the interface region shown in Figures 3 and 4 can be associated with the type of interface that has been examined—either Ti/ferrite or Ti/perlite. In the case of perlite (0.77 wt pct C), even short heat treatment can lead to diffusion of carbon, resulting in the formation of a microstructure such as the one presented in Figure 3. The second factor that influences the phase formation at the joint interface according to Reference 4 is the change in Gibbs free energy as a function of temperature, which is much lower for TiC compared to Fe₂Ti or FeTi phase in the temperature range related to the welding process and annealing. Studies of the interface formed directly after the EXW process did not reveal the formation of Ti carbides, while this phase is always observed to form after the additional heat treatment, which brings the system closer to equilibrium. The temperature increase during the welding process favors the remelting of the material and formation of the intermetallic phases especially at the crest of waves, where the highest local increase in temperature is expected. One can notice that the formation of intermetallic phases in this area is accompanied by the formation of carbides in the transition zone adjacent to it after annealing. In the area of the interface less affected by the increase in temperature, such as the area between the waves, formation of intermetallics and subsequent diffusion of carbon (due to the heat treatment) will occur at a much smaller scale than in the previous case. In this case, the thin zone of intermetallics and carbides (after heat treatment) is expected, which is in agreement with our observations of interface microstructure. In the work of Jiang et al.,[5] Ti/steel explosive-rolling cladding plate was analyzed in terms of the influence of different heat treatments (650 to 950 °C for 60 minutes) on its microstructure and mechanical properties. In this case, the interface of the formed joint had a rather flat character and the parts and the interface were studied by TEM. It was found that Ti/steel explosive-rolling clad plate heated at 850 °C or below this temperature entails the formation of a TiC thin layer (300 nm) composed of TiC crystallites at the interface due to diffusion of carbon toward the Ti side and aggregation at the weld interface. Jiang et al.[5] reported that further diffusion of carbon into Ti was difficult due to the lower solubility of C in Ti alloy, while the low binding energy of TiC caused them to form on the interface, which suppressed the mutual diffusion of Fe and Ti, and the TiC was located next to the Ti side of the interface. Raising the rolling temperature to 950 °C resulted in the diffusion of Ti and Fe, and therefore, a layer of 200 to 300 nm composed of a Fe₂Ti and FeTi mixture of grains was formed along the interface of the Ti/steel clad plate. It was concluded that the temperature of 883 °C corresponding to the temperature of α \( \Leftrightarrow \) β transformation plays an important role in the formation and growth of the intermetallic compounds at the welded interface. This, in turn, influenced the shear strength of the weld, which decreased with the presence of FeTi-type intermetallics. Our study demonstrates that the temperature boundary (883 °C) pointed out in Reference 5 is not that sharp, and the FeTi compounds can be formed even after short-term annealing at 600 °C, as evidenced in the high-resolution image from TEM presented in Figure 4(c).

In order to confirm the plastic properties of the weld after the applied heat treatment, the side bending tests comprising the elongation of 24 pct were performed followed by the observations of the tensile reference surface. The view of the sample after side bending is shown in Figure 6(a), where no cracks were observed in the macroscale (Figure 6(b)). Microstructural changes of samples subjected to plastic deformation were also examined with a scanning electron microscope. These observations (an example of such microstructure is presented in Figure 6(c)) allow exclusion of any material discontinuities, which confirms the high quality and plasticity of the weld after applied heat treatment. The microstructural observation results were similar for all tested samples, and the only crack-type discontinuities were observed within the volume of the melted regions (arrows in Figure 6(c)). This indicates the brittleness of the remelting zones due to the presence of intermetallic phases in a relatively large volume. Similar results, but with a different stress state, were obtained in References 8, 14, and 15, where also no crack propagation beyond the areas of remelting was observed.

The microhardness measurements were performed across the interface and included the tests of both base and clad plates at a given distance (up to 1.5 mm) from their interface. These measurements were performed across the interface avoiding a broad zone of the solidified melts, as this study is dedicated to the apparently sharp Ti/steel interface. The microhardness tests are the basic indicator of the effect of stress relief annealing influence on the reduction of the strengthening after the explosive welding process. In general, after the explosive welding process and near the weld, the materials are characterized by a highly deformed microstructure, which is associated with the phenomenon of strengthening and a decrease in plastic properties. Postcladding heat treatment allows leveling of the effects of strengthening and causes an increase of joint plasticity. Thus, it allows further plastic deformation of explosively welded materials, e.g., bending, stamping, rolling, and welding, and therefore, it is a basic stage of the technological process of manufacturing equipment made of explosively welded materials. As shown in this work with the aim of TEM as a result of diffusion processes during the short heat treatment of Ti/clad steel, creation of undesirable brittle Fe_xTi_y and TiC phases takes place; however, this region of interest is localized at the interface with the thickness below 1 μm. It was expected that, as a result of recrystallization and recovery processes, the defected microstructure is rebuilt and the grains grow, which will affect the obtained microhardness values. Figure 7 shows the summary of the microhardness tests of the samples after applied heat treatment. Microhardness values are significantly lower for the clad material in comparison to the base material. These values vary in the range of 130 to 170 HV0.1 for Ti and 200 to 240 HV0.1 for steel. In the area close to the interface of the steel side, there is an increase in microhardness, pointing out that the strengthening after this heat treatment has not been completely removed. These values can be compared to those in Reference 4, where thicker Ti flyer plate has been applied (6 mm) in comparison to that used in this study (2 mm). However, the microhardness values for Ti Gr. 1 presented in the graph in Figure 7 are similar to these obtained in Reference 4. Additionally, at the Ti side in the vicinity of the interface, reduced microhardness values were observed, similar to those in Reference 4, where they were explained by the recovery, recrystallization, and loss of adiabatic shear bands in the microstructure.