CVN impact energy and fracture characteristics correlations with different oxide nanoparticles improving submerged arc welds

Table 4 shows the WM chemical composition of WB-0, W-T, WB-S, WB-A and WB-M submerged arc welds.

It is expected that the greatest contribution in the chemical composition of the WM comes from the flux and the electrode (see table 1). If we analyze the content of C it is very close to or slightly higher than that of the electrode; This may be due to the fact that C reacts easily with the O from the oxides in the flux, which are dissolved in the molten metal, which in turn form CO and CO₂ and are subsequently eliminated into the atmosphere.

Mn and Si contents are also very similar to that of the electrode with a slight increase, especially in Mn, noting the contribution of the flux. According to Bennet et al [19], the content of MnO and SiO₂ in the flux increases the levels of Mn and Si in the MS, which contribute to the formation of complex oxides, which in turn form non-metallic inclusions and facilitate growth. of the AF phase in the submerged arc welds. Mn content of MS with Nps, WB-T, WB-S and WB-M is always greater than that of WB-0 by 17, 1 and 5%, respectively. Only for WB-A, the Mn content is less by 6%. In addition, the difference is made more noticeable by the addition of TiO₂-n, compared to other submerged arc welds. These results agree that the addition of Ti to the weld pool favors the transfer and retention of Mn [20, 21]. In contrast, submerged arc welds with added Nps have lower Si content, 65% on average compared to WB-0.

Meanwhile, the Cu content comes exclusively from the electrode, being retained in greater quantity by the submerged arc welds with Nps of TiO₂-n and Mn₂O₃-n. On the other hand, the Al that comes exclusively from the flux is retained at 125 ppm by all the submerged arc welds with added Nps.

On the other hand, the weld metal microstructure, as shown in figures 1(a)–(e), presents the allotriomorphic ferrite (AL-Fe) and acicular ferrite (AF) phases, the latter being distinguished by its interlocking needle-shaped grains.

Zoom In Zoom Out Reset image size

Figure 1 shows that the AF phase mainly grows in the dendritic solidification regions of the weld metal region. Given the importance that AF phase has in the mechanical properties of the different submerged arc welds, its morphological characteristics (length, width and surface density percent), as are shown in table 5, were directly determined on the OM metallographic images. The length and width were measured by ImageJ software v.1.51j8, reporting the averages obtained by processing more than 20 images, in which there were more than 40 AF needles. Meanwhile, the surface density percent was determined by placing a millimeter grid metallographic images, obtaining the ratio of the number of sheets occupied by the AF phase over the total number of sheets encompassing the millimeter grid.

In table 5, it is observed that both length, width and surface density percent of the AF are greater in the weld metal with added oxide-NPs compared to that of the WB-0. These increases of superficial density percents are framed in table 6.

It is observed that Al₂O₃-n are the ones that induce a greater growth in long and wide AF needles. From these results, it can be established that the type of added oxide-NP influences the morphological characteristics of AF needles.

SEM metallographic images (right side in figure 1) show that most of the inclusions, whose morphology is spheroidal, are found in regions with high surface density of AF needles; they can be classified into two types: Internal that are located inside of above regions, and the External located mainly in regions with AL-Fe and that are outside of the AF colonies. Figure 2 shows the size distribution of Internal and External inclusions in the WM for different submerged arc welds.

Zoom In Zoom Out Reset image size

As can be seen in table 7, these ranges vary from 0.91 to 1.35 μm and from 2.56 to 4.37 μm for the Internal and External inclusions, respectively.

In particular, the internal inclusions diameters are within the range reported by other authors [7, 8], who conclude that they promote the growth of the AF phase, acting as nucleation sites. To visualize this, in figure 3 the values of the surface densities (table 5) were plotted as a function of the internal inclusions diameter (table 7). It can be observed that the surface density of the AF needles is greater as the average diameter of the internal inclusions decreases.

Zoom In Zoom Out Reset image size

The chemical compositions of the internal inclusions in the weld metal were punctually determined by SEM-EDS on dozens of inclusions and the average chemical compositions are shown in table 8. Of course, when a particle analysis is carried out in a MEB, the uncertainty always prevails as to whether the surrounding material, in this case the weld metal, is reached by the incident electron beam and the x-rays generated contribute to the analysis of inclusions.

In order to rule this out, calculations of the electron beam penetration range at 20 kV were made using the Kanaya-Okayama equation [22], obtaining average values ​​of 0.9 and 0.4 μm for inclusions and weld metal, respectively. In this way, the calculated penetration was always less than or equal to the average diameters of the internal inclusions, which indicates a minimal interaction of the electron beam with the weld metal below the inclusions. For reference, the average chemical composition of the surrounding weld metal (WM, table 8) to the inclusions was also determined. It is observed that most of the elements have higher concentrations in the inclusions than in the weld metal.

In particular Ti, Si, Al and Mn, important elements in the inclusions formation, are detected with substantially higher contents than in the Matrix, which gives certainty that they are indeed part of them. On the other hand, the inclusions of the submerged arc welds with added oxide-NPs showed a higher percentage content by weight of Ti and Mn compared to WB-0, so it can be established as a direct influence of the addition of oxide-NPs. Furthermore, when SiO₂-n is added to the submerged arc welds, the highest Si values are obtained in the internal inclusions. Adding Mn₂O₃-n to the submerged arc weld (WB-M) yields one of the highest Mn values, second only to the WB-S by 2.5 percentage points. If Al₂O₃-n is added to the submerged arc weld (WB-A), a high content of Al is obtained in the Internal inclusions of the WB-A, only surpassed by the WB-M in 0.7 percentage points.

This allows to establish an influence on the type of oxide added as nanoparticle, and the content of the internal inclusions. In this way, it is possible to control the formation of the active inclusions of complex oxides that favors growth of AF and avoid the formation of those that inhibit it. For example, complex oxides of Al₂O₃ [23, 24] have been reported to inhibit the growth of the AF phase, which may be occurring in WB-0 with a high content of Al and O in its inclusions (table 8) and that at the same time presents the lowest percentage of surface density of AF (table 5).

The results of the CVN impact test for weld beads without and with added oxide-NPs are shown in table 9.

It can be seen that the CVN impact energy is always higher for the submerged arc welds with added oxide-NPs compared to the WB-0. The percentages increased 30.6, 83.9, 50.9 and 60% for WB-T, WB-S, WB-A and WB-M, respectively. This means a clear increase in toughness in the submerged arc weld, by adding oxide-NPs, especially SiO₂-n, where the impact energy is higher than obtained for WB-0.

The loading angles (α) and the shear fracture percent (% F_s ) of the tested CVN impact specimens were determined following the methodology indicated in Annex A6 of the Standard Methods for Impact Tests on Bars with Notched Metallic Materials (ASTM E23). The loading angles were determined for each of the tested CVN impact specimens, measuring the angle α resulting from joining the broken parts, as shown in figure 4(a).

Zoom In Zoom Out Reset image size

On the other hand, the shear fracture percent was determined by measuring the length ($a$) and width ($b$) over the flat fracture region of each specimen cross section, shown in figure 4(b). Once these values were measured, the flat fracture area (A_f ) was calculated, multiplying the length and width. Then, taking into account that the average specimen cross-section area before being tested (A_i ) is 8 × 10 mm = 80 mm², and that the notch depth is 2 mm. Then, the percentage difference between A_i and A_p is calculated as indicated by the following formula:

$$\begin{eqnarray} &&\% {F}_{s}=\displaystyle \frac{{A}_{i}-{A}_{f}}{{A}_{i}}\,* \,100\end{eqnarray} \tag{ 1 }$$

The average values of α, $a,$ $b,$ A_f and %F_s are shown in table 10.

In figure 5, it is observed that as the loading angle decreases or as the shear fracture percent of the tested CVN impact specimens for each submerged arc weld increases, the CVN impact energy increases. Higher loading angle means that the fracture of the specimen took place with less deformation, and therefore required less CVN impact energy. Higher shear fracture percent means that the region that defines resistance to the separation of the specimen into two pieces, was greater and with it also the required CVN impact energy.

Zoom In Zoom Out Reset image size

On the other hand, ductile fracture regions, such as those shown in figures 6(a)–(e), form the shear lips, that is, the edge contiguous to the notch and the edge opposite to the final fracture region. In contrast, brittle fracture regions, such as those shown in figures 7(a)–(e), were located only in the center of the flat fracture, delimited by $a$ and $b$ (figure 4(b)). In this region, some small islands of ductile fracture are found, such as the one shown in figure 7(c).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

These results are in accordance with the ASTM E23 standard, where near the V-notch, a plastic deformation of the grains occurs, leading to fibrous breakage, that is, a ductile type fracture. As the fracture moves away from the notch, the crack propagation becomes unstable, whereby the cleavage mechanism changes and complete grain separations occur, i.e. a brittle intergranular fracture. These cracks move radially, from the center of the sample to the edges. The last section of separation of the specimens takes place again ductility on the edge and opposite side of the notch, thus forming the ductile fracture cutting lips that delimit the brittle fracture region in the center of the cutting surface. In ductile fracture, excision occurs after plastic stretching of the material, resulting in the formation of intragranular microcavities or dimples; the greater this stretch, the deeper the cavity will be. In table 11 the average diameters of these dimples are framed and in figure 8 they are plotted against the CVN impact energy.

Zoom In Zoom Out Reset image size

An increase in the impact energy is observed as the diameter decreases, this is explained because the smaller the microcavity, there will be a greater number of walls resisting stretching, which will require more CVN impact energy to cause the brittle fracture. This is in agreement with Zhou and Chew [25], who report an increase in CVN impact energy associated with a decrease in the dimple diameter in the ductile fracture regions.

On the contrary, in brittle fracture the cleavage occurs by a sudden separation of the grains, so that a minimum consumption of Impact energy is required. This intergranular fracture reveals the facets of the grains (figure 7), whose measured average widths are framed in table 10. The sizes of these cleavage sheets agree with the grain sizes observed in the SEM metallographies of figure 7, corresponding to the WM, as expected as there was no grain deformation in the brittle fracture region. This criterion of CVN impact energy consumption by the different types of fracture coherently explains the inverse behavior of CVN impact energy with the brittle fracture area, shown in table 10. As the CVN impact energy increases, the brittle fracture area decreases, or equivalently the brittle fracture area increases, which will require higher CVN impact energy consumption.

From the values ​​reported in table 11, it is observed that both the dimples and the cleavage sheets are smaller in the submerged arc welds with oxide-NPs than in the WB-0. Compared to WB-0, the average dimples diameters are 6.1, 24.2, 18.2 and 15.2% higher in the welds WB-T, WB-S, WB-A and WB-M, respectively; while the thicknesses of the cleavage sheets are 41.6, 62.7, 66.9 and 21.1% less and in the same order already mentioned, respectively. Therefore, in figure 9 it can be seen that as the width of the cleavage sheets decreases (table 11), the superficial density of AF increases (table 6), which also agrees with other investigations [19, 23, 26], where the decrease is associated with AF formation and grain size refinement resulting in smaller cleavage sheets in the ductile fracture region.

Zoom In Zoom Out Reset image size

In this way, both the increase in CVN impact energy and the AF-needles surface density can be correlated with the decrease in the average dimples diameter and the cleavage sheets width, respectively. This allows to conclude that the toughness of the submerged arc welds can be improved by adding oxide-NPs.