Regular Article
Influence of Cr on Weld Solidification Cracking in Fe-15Mn-0.5C-3.5Al-xCr Alloys
2015 Volume 55 Issue 1 Pages 257-263
Details
2015 Volume 55 Issue 1 Pages 257-263
The solidification cracking sensitivity of Fe-15Mn-0.5C-3.5Al-xCr (x= 5, 12 mass%) alloys was evaluated using a longitudinal Varestraint test and compared with that of austenitic Fe-18Mn-0.6C alloy. Weld microstructures of Low Cr (5 mass%) and High Cr (12 mass%) alloys revealed duplex structures with a mixture of austenite and δ ferrite. The amounts of residual δ ferrite and (Cr, Fe, Mn)23C6 carbide remarkably increased with increasing Cr content. A small amount of Cr addition (5 mass%) provided negligible influence on the solidification cracking susceptibility. However, the welds of high Cr alloys demonstrated excellent resistance to solidification cracking due to healing by the eutectic liquid. The addition of Cr enhanced the formation of a low melting point (γ + (Cr, Fe,Mn)7C3) eutectic during solidification by increasing the eutectic formation temperature and simultaneously decreasing the eutectic C content.
Recently, austenitic high Mn steels have received much attention due to their exceptional combination of high strength and high ductility.1,2,3) The research activity regarding high Mn steel has drastically increased over the past few years. Scientific interests have predominantly focused on the mechanical properties and phase transformation during deformation.4,5,6) However, austenitic high Mn steels have not yet been evaluated in regard to their weldability.
It has been reported that the austenitic high Mn TWIP steels produce heat affected zone (HAZ) liquation cracks during resistance spot welding.7) They also exhibited a high sensitivity to liquation cracking in contrast to Cr–Ni steels.8) In addition, the welds of Fe-18Mn-0.6C steels were susceptible to solidification cracking due to the formation of a low melting γ + (Fe,Mn)3C eutectic during solidification.9) It has been generally accepted that C has a strong tendency to segregate at the grain boundary and tends to form the low melting point eutectic during solidification. However, it is difficult to control the formation of an eutectic in austenitic high-Mn steels, because a large amount of C should be present for stabilizing the austenite structure at room temperature. In the case of solidification mode, the welds solidified as primary ferrite are reported to be more resistant to solidification cracking than fully austenitic welds due to beneficial effects of δ ferrite at high temperature.10,11) It is believed that the increased grain boundary area in γ + δ dual phase welds reduces the chances of forming continuous low melting liquid along grain boundaries during solidification.12) Further, many researchers have suggested the proper range of residual ferrite content for optimum resistance to solidification cracking in Fe–Cr–Ni13) and Fe–Mn–Al–C14) alloys.
In the present study, Fe–Mn–C–Al–Cr alloys were designated for reducing the susceptibility of solidification cracking of Fe-18Mn-0.6C alloys. Both Al and Cr, which stabilize the ferrite phase, were added into the Fe–Mn–C system to control the solidification mode. The solidification cracking susceptibilities of Fe–Mn–C–Al–Cr alloys were evaluated using a longitudinal Varestraint and discussed with regard to the solidification behavior and microstructural evolution during solidification.
The present alloys were prepared in a vacuum induction furnace. The ingots were homogenized at 1150°C for 2 h and hot-rolled into a final thickness of 5 mm. The plates were finally air-cooled to room temperature. The chemical compositions are listed in Table 1.
| C | Mn | Cr | Al | P | S | |
|---|---|---|---|---|---|---|
| No Cr | 0.63 | 17.92 | – | – | 0.0040 | 0.0030 |
| Low Cr | 0.52 | 15.31 | 4.96 | 3.53 | 0.0038 | 0.0045 |
| High Cr | 0.52 | 15.25 | 12.00 | 3.72 | 0.0041 | 0.0042 |
A longitudinal Varestraint test was utilized to evaluate the solidification cracking sensitivity of the alloys. Specimens were cut into dimensions of 127 × 25.4 × 3 mm3 (Length × Width × Thickness). Autogenous GTA bead-on-plate welding was carried out for the Varestraint test, and the welding parameters are listed in Table 2. The augmented strain was in the range of 1% and 5%. The lengths of solidification cracks were examined and measured on the as-welded surface at 25× magnification.
| Process | Current (A) | Voltage (V) | Travel speed (mm/s) | Shielding gas & Flow rate (L/min) |
|---|---|---|---|---|
| Autogenous GTAW | 100 | 12 | 4 | Ar (20) |
The specimens were etched in a solution of 1:1:1 mixture of nitric acid, hydrochloric acid, and acetic acid. For transmission electron microscopy (TEM, JEM-2100F) examination, the samples were prepared using a twin-jet polisher (Tenupol-3) in a solution of hydrochloric acid (10%) + methanol (90%) at –40°C. The microstructures were examined using an optical microscope (OM) and field emission scanning electron microscope (FESEM, JSM-6700F). Elemental mapping was performed using electron probe microanalysis (EPMA, JXA-8100). Focused ion beam (FIB, NOVA200) technique was utilized to verify the eutectic phase along the grain boundary.
Figures 1(a) and 1(b) indicate typical micrographs of base metals in the low Cr alloy (5 mass%) and high Cr alloy (12 mass%), respectively. The microstructure of the low Cr alloy demonstrated a fully austenitic structure and no distinct carbides at grain boundaries. In contrast, increasing the Cr content to 12 mass% led to a mixture of austenite and a large amount of carbides at grain boundaries. It is clearly observed that precipitates are positioned along grain boundaries (shown as a dark etched region). Typical microstructures of weld metals formed by GTA bead-on-plate welding are shown in Figs. 1(c) and 1(d). The weld microstructure of the low Cr alloy consisted of a mixture of the austenitic dendrite and island shaped δ ferrite positioned at the dendrite core. In weld metal of the high Cr alloy, the volume fraction of δ ferrite noticeably increased and the morphology of δ ferrite apparently changed into a vermicular shape. The presence of δ ferrite at the dendrite core provides evidence regarding the solidification sequence, which indicates primary ferritic solidification (FA mode).
Optical micrographs of the base metals: (a) Low Cr, (b) High Cr; and the welds: (c) Low Cr, (d) High Cr.
Figures 2(a) and 2(b) compare the solidification cracking susceptibility of no Cr, low Cr, and high Cr alloys. Total crack length (TCL) and maximum crack length (MCL) as a function of augmented strain permit the relative comparison of cracking tendency and solidification cracking temperature range, respectively. In the previous study, it was clarified that Fe-18Mn-0.6C steel (No Cr alloy) welds are sensitive to solidification cracking due to high levels of C, which affect the formation of low melting γ + (Fe,Mn)3C eutectic at the terminal stage of solidification.9) Further, both TCL and MCL of the no Cr alloy monotonically increased until 4% augmented strain and then saturated at 5%. The low Cr alloy was less susceptible to solidification cracking than the no Cr alloy. Meanwhile, it is unclear if a low Cr content decreased the solidification cracking susceptibility of the Fe–Mn–C system due to the difference of initial chemical composition. Slightly reduced C content in comparison with the no Cr alloy may play a major role in reducing the susceptibility of solidification cracking. In addition, the change of solidification mode from fully austenitic (A) to ferritic-austenitic (FA) may dominantly contribute to improving the solidification cracking resistance. More addition of Cr into low Cr alloys obviously decreased the susceptibility of solidification cracking. The high Cr alloy was found to offer excellent resistance to solidification cracking over a strain range of 1% to 3%. There was no evidence of solidification cracking at 1%, and only a few cracks were detected at 2% and 3% applied strain. It is apparent that the threshold strain of high Cr alloys lies in a range of 1% to 2%, and is higher than those of the no Cr and low Cr alloys. However, the extent of solidification cracking of the high Cr alloy drastically increased from 4% applied strain. The extent of TCL at 4% applied strain was more than twice that at 3%. At 5% applied strain, the TCL of the high Cr alloy exhibited the highest value among other alloys. Some cracks in no Cr alloy and low Cr alloy were found the evidence of ductility dip cracks (DDC) at the tip of cracks over 4% augmented strain, but the length of DDC was negligible and positioned within the range of error bar.
The results of fusion zone cracking versus augmented strain for no Cr, low Cr, and high Cr alloys: (a) Total crack length (TCL) and (b) Maximum crack length (MCL).
Low magnification micrographs of the low Cr and high Cr alloys are shown in Fig. 3. In the low Cr alloy, the solidification cracks were clearly detected at a given augmented strain range from 3% to 5%. In the high Cr alloy, the cracks are distinctly observed at 4% augmented strain. However, the majority of solidification cracks were observed far from the weld pool boundary and disconnected adjacent to the weld pool. At 5% augmented strain, the cracks were generated and inter-connected between the weld pool line and crack tips. In addition, the number of solidification cracks of the high Cr alloy noticeably increased compared to that of the low Cr alloy at the same strain levels of 4% and 5%.
Low magnified optical micrographs of as-welded surfaces in a Varestraint tested specimen as a function of augmented strain (low Cr: (a) 3%, (b) 4%, (c) 5%; high Cr: (d) 3%, (e) 4%, (f) 5%).
Solidification cracking generally occurred along the grain boundaries during solidification when the applied strain exceeded the limited level. Representative as-welded solidification cracks are shown in Figs. 4(a) and 4(b). The solidification cracks of the low Cr alloy initiated and propagated along the grain boundary in Fig. 4(a), which indicates that the region between solidification grain boundaries was completely solidified at the moment of crack formation. On the other hand, the solidification sub-grain boundaries of the high Cr alloy exhibited a dendritic cell morphology and several small fissures were detected along the sub-grain boundaries, as shown in Fig. 4(b). This finding suggests that some solidification cracks were generated and interconnected along the solidification sub-grain boundaries, which resulted in the presence of liquid at the sub-grain boundaries upon crack formation. Figures 4(c) and 4(d) exhibit solidification cracks after polishing and etching. Coarse and straight solidification cracks are observed in the low Cr alloy, as shown in Fig. 4(c). The fissure morphology of the low Cr alloy was similar to that of a weld solidified as fully austenitic (A).9) The majority of solidification cracks in Fe-18Mn-0.6C alloy (No Cr) also exhibit coarse and smooth morphologies due to a large amount of liquid present at the moment of crack formation. However, most solidification cracks in the high Cr alloy were relatively short and narrow, and discontinuously propagated along the grain or sub-grain boundaries, as shown in Fig. 4(d). Grain size, which was refined by changing the solidification sequence from A to FA mode, served as an effective barrier for crack propagation. The difference of solidification cracking behavior between low Cr and high Cr alloys is possibly attributed to the δ ferrite content. Though the welds of the low Cr alloy solidified as a primary ferrite, a relatively low percentage of δ ferrite transformed during solidification may not affect the grain refinement as much as the high Cr alloy.
Solidification cracks in the as-welded surface of (a) low Cr and (b) high Cr; in the as-etched surface of (c) low Cr, and (d) high Cr.
It is well-known that the welds solidified in a non-equilibrium state due to fast cooling have inevitable chemical inhomogeneity during solidification. It is important to predict or reproduce the solidification process in order to understand microstructural evolution during welding. Equilibrium pseudo-phase diagrams of the low Cr and high Cr alloys are shown in Figs. 5(a) and 5(b), respectively. As mentioned above, the welds of the no Cr alloy were susceptible to solidification cracking due to the low melting γ + (Fe,Mn)3C eutectic at 1090°C.9) In the case of the low Cr alloy, the increased Liquid + BCC_A2 area is noticeable following the addition of Al and Cr corresponding to ferrite stabilizing elements, as shown in Fig. 5(a). An initial phase during solidification is anticipated as the ferrite at a given chemical composition. In addition, the eutectic reaction is changed from γ + (Fe,Mn)3C to γ + (Cr, Fe,Mn)7C3, which is attributed to the enlarged M7C3 area from the right side by the addition of Cr. Such thermodynamic behavior is prominent in the phase diagram of the high Cr alloy, as shown in Fig. 5(b). Moreover, increased BCC_A2 and M7C3 areas are clearly noted in comparison with the low Cr alloy. Specifically, the liquid + M7C3 region expanding from the right side leads to the decrease of eutectic C content, and the increase of eutectic formation temperature. Corresponding values are approximately 4.9 mass% C at 1202°C in the low Cr alloy and 3.6 mass% C at 1269°C in the high Cr alloy.
Equilibrium pseudo-binary phase diagrams of (a) low Cr and (b) high Cr alloys.
Based on a thermodynamic database (Thermo-Calc, TCFE7), the solidification path is described by Scheil simulation, as shown in Fig. 6. Both the low Cr and high Cr alloys solidify as a primary ferrite and finish the solidification process by the γ + (Cr, Fe,Mn)7C3 eutectic reaction. The noticeable difference is the volume fraction of γ + (Cr, Fe,Mn)7C3 eutectic. The amount of liquid at the terminal stage of solidification provides important evidence to explain the solidification cracking sensitivity. The eutectic fraction obtained from Scheil simulation is 3% from the low Cr alloy and 10% in the high Cr alloy. The increased eutectic fraction is strongly related to the eutectic C content and its formation temperature. A reduction in eutectic C content leads to easy formation of eutectic, because C segregation during solidification enables easy access to the eutectic C content. In addition, an increase in eutectic starting temperature results in the formation of eutectic at an earlier stage of solidification.
Solid fraction versus temperature curve representing a Scheil model: (a) low Cr and (b) high Cr.
Thermodynamically calculated results cannot perfectly reflect the actual welding process. However, the solidification behaviors of both alloys are in good agreement with the above microstructural characteristics and Varestraint results. Excellent resistance to solidification cracking of the high Cr alloy at an augmented strain level below 3% corresponds to the liquid present at the final stage of solidification. The eutectic liquid between grains and sub-grains may be enough to recover the cracks formed at a low strain level. In addition, a large amount of δ ferrite can also serve to accommodate the stress, because δ ferrite generally has a higher ductility than austenite at high temperature.10) However, when the applied strain exceeds the criterion for healing incipient cracks, the solidification grain and sub-grain boundaries covered eutectic liquid act as initiation and propagation sites of solidification cracking. It is closely related to the drastic increment in TCL of the high Cr alloy over 4% augmented strain.
4.2. Formation of δ FerriteIn austenitic stainless steels, the formation of δ ferrite is dominantly dependent on the Cr content because the major alloy components of stainless steels are Cr and Ni. The control of ferrite content using Cr is effective in low C alloys. In austenitic high Mn steels, high levels of C content should be included to stabilize the austenite at ambient temperature. Therefore, the addition of approximately 3.5 mass% Al was required to change the solidification mode because Al acted as s strong ferrite stabilizer in Fe-18Mn-0.6C alloys.15) Above mentioned in Figs. 1(c) and 1(d), the typical morphology of δ ferrite was prominently different between the low Cr and high Cr alloys. Individual δ ferrite of the low Cr alloy was positioned solely within interior grains and rarely detectable, however, that of the high Cr alloy revealed a well-interconnected network and was uniformly distributed. Ferrite contents measured by Feritscope (Fischer MP30) were 0.19% in the low Cr alloy and 12.9% in the high Cr alloy. Many previous studies have investigated systematically the role of δ ferrite within solidification cracking.10,16) The major factors can be summarized as follows: 1) high solubility of impurities such as P and S, 2) high ductility at high temperature, 3) lower thermal expansion coefficient compared to austenite, and 4) increased boundary area by grain refinement. However, the formation of an adequate δ ferrite fraction should be accompanied by an effective reduction of the solidification cracking susceptibility. From the solidification crack micrograph of the low Cr alloy in Fig. 4(c), the small amount of δ ferrite provided negligible effect on solidification cracking even though the welds of the low Cr alloy solidified as a primary ferrite.
The typical TEM micrograph of δ ferrite containing the concentration profile in high Cr alloy is shown in Fig. 7. The result shows that the δ ferrite was formed on the basis of Al and Cr. The variation from the nominal composition was pronounced. Further, a drastic increment of Cr in the γ/δ interphase boundary was attributed to the presence of (Cr, Fe, Mn)23C6, that formed during cooling and was extensively distributed along the grain and sub-grain boundaries.
(a) TEM micrograph of residual δ ferrite in high Cr alloy and (b) elemental profile across δ ferrite outlined by dotted line from (a).
Figure 8(a) presents the SEM micrograph of a backfilled crack in the high Cr alloy. Numerous γ + (Cr, Fe,Mn)7C3 eutectics were detected along the grain boundaries in the vicinity of the backfilled crack. They provide strong evidence that the chemical composition in liquid satisfied the eutectic concentration at the final stage of solidification. A portion of the eutectics were chemically analyzed by EPMA, as shown in Fig. 8(b). Severe co-segregations of C, Cr, and Mn were found in the interior of eutectics. Moreover, a noticeable depletion of Al was prominent. The formation of γ + (Cr, Fe,Mn)7C3 eutectic was dominantly attributed to the elemental segregation during solidification, especially for C and Cr. There were a few eutectics detected adjacent to backfilled cracks in the low Cr alloy, because the relatively high eutectic C content of 4.9 mass% made the formation of γ + (Cr, Fe,Mn)7C3 eutectics difficult.
(a) SEM micrographs showing γ+(Cr,Fe,Mn)7C3 eutectic along the grain boundary adjacent to backfilled crack in the high Cr alloy and (b) corresponding EPMA elemental mapping in the backfilled region indicated by dotted square from (a).
Representative micrographs of eutectic morphology are shown in Fig. 9. The γ + (Cr, Fe,Mn)7C3 eutectic of the low Cr alloy revealed a feathery-shaped morphology, while conversely, that of the high Cr alloy dominantly showed a typical lamellar structure. The γ + (Cr, Fe,Mn)7C3 eutectic size of the high Cr alloy was shown to be much larger than that of the low Cr alloy. It is apparent that a large amount of eutectic liquid at the final stage of solidification contributed to the increase of eutectic size and fraction.
Representative γ + (Cr,Fe,Mn)7C3 eutectic morphology of (a) low Cr and (b) high Cr alloys.
Vertical cross-section TEM micrographs of γ + (Cr, Fe,Mn)7C3 eutectic from the high Cr alloy are displayed in Figs. 10(a) and 10(b). The γ + (Cr, Fe,Mn)7C3 eutectic is noticeably divided into two distinct regions, which correspond to the Cr enriched and Fe enriched areas, as shown in Fig. 10(c). Figure 10(d) shows the high-resolution TEM image of the interphase between γ matrix and (Cr, Fe,Mn)7C3 carbide. Corresponding FFT patterns were taken from the A and B regions, as shown in Figs. 10(e) and 10(f), respectively. Region A denotes the FCC structure of austenite with a zone axis of [001] and region B shows the orthorhombic structure of M7C3 type carbide with a zone axis of [112].
(a) Vertical cross-section TEM micrograph obtained by focused ion beam (FIB) from γ + (Cr,Fe,Mn)7C3 eutectic in the high Cr alloy, (b) magnified TEM micrograph of the region indicated by dotted square, (c) corresponding elemental maps (yellow: Fe, wine: Cr) obtained from (b), (d) highresolution TEM of γ + (Cr,Fe,Mn)7C3 eutectic obtained from dotted circle region in (b), FFT patterns obtained from (e) region A and (f) region B.
Fractographic analysis of the solidification cracks specimen was conducted to verify the type of cracking. The cracks were artificially opened by fracturing. The fracture appearance of Varestraint samples tested at 5% applied strain are shown in Fig. 11. The surface of cracks in both the low Cr and high Cr alloys exhibit a smooth cellular dendritic structure, which provides strong evidence of the presence of liquid films at the moment of cracking. In addition, a magnified fracture surface of the high Cr alloy revealed that the solidification cracks were generated by the rupture of eutectic liquid film, which correspond to a γ + (Cr, Fe,Mn)7C3 eutectic.
Fractographs representing solidification crack surface of the low Cr alloy: (a) low magnification, (b) high magnification; and the high Cr alloy: (c) low magnification, (d) high magnification.
The solidification cracking susceptibility of the low Cr (5 mass% Cr) and high Cr (12 mass% Cr) alloys were evaluated using a longitudinal Varestraint. The addition of Cr enhanced the formation of γ + (Cr, Fe,Mn)7C3 eutectic by increasing the eutectic formation temperature and decreasing the eutectic C content. A low amount of Cr addition provided negligible effect on the solidification cracking susceptibility despite of changing solidification sequence from fully austenitic (A) to ferritc-austenitic (FA) mode. The weld of the high Cr alloy demonstrated excellent resistance to solidification cracking in a strain range of 1% to 3% compared to that of the no Cr and low Cr alloys due to healing by the eutectic liquid. However, the solidification cracking sensitivity of the high Cr alloy drastically increased over 4% applied strain due to the relatively large initiation and propagation sites of cracking.
The authors would like to thank the POSCO Technical Research Laboratory for financial support of this research.
