Received Date: April 12, 2017; Accepted Date: April 24, 2017; Published Date: May 04, 2017
Citation: Al-Moussawi M, Smith AJ, Faraji M (2017) Friction Stir Welding of DH36 Steel and Segregation of Mn, Si, Al, and Oxygen. J Material Sci Eng 6: 336. doi: 10.4172/2169-0022.1000336
Copyright: © 2017 Al-Moussawi M, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Journal of Material Sciences & Engineering
This work investigates the role of welding speed in elemental segregation of Mn, Si, Al, and oxygen during Friction Stir Welding (FSW) in DH36 steel. The experimental work undertaken showed that when the speed of the FSW process exceeds 500 RPM with a traverse speed of 400 mm/min then elemental segregation of Mn, Si, Al and O occurred. The mechanism of this segregation is not fully understood; additionally the presence of oxygen within these segregated elements needs investigation. This work examines the elemental segregation within DH36 steel by conducting heat treatment experiments on un-welded samples incrementally in the range of 1200-1500°C and at cooling rates similar to that in FSW process. The results of heat treatments were compared with samples welded under two extremes of weld tool speeds namely W1 low tool speeds (200 RPM with traverse speed of 100 mm/min) and W2 high tool speeds (550 RPM with traverse speed of 400 mm/min). The results from the heat treatment trials showed that segregation commences when the temperature exceeds 1400°C and Mn, Si, Al and oxygen segregation progress occurs at 1450°C and at a cooling rate associated with acicular ferrite formation. It was also found that high rotational speeds exceeding 500 RPM caused localised melting at the advancing-trailing side of the friction stir welded samples. The study aims to estimate peak temperature limits at which elemental segregation does not occur and hence prevent their occurrence in practice by applying the findings to the tool’s rotational and traverse speed that correspond to the defined temperature.
Friction stir welding; DH36 steel; Elemental segregation
Since FSW was invented in 1991 by The Welding Institute (TWI) much effort has been put into its commercialisation latterly for higher melting point materials such as steel. The High Integrity Low Distortion Assembly (HILDA) project is an example of this effort which is aiming to understand the process of FSW of a DH36 galvanised steel grade including microstructure evolution, modelling, mechanical properties and defect formation during FSW. In earlier work, Toumpis et al.  examined the effect of the FSW tool traverse and rotational speeds on the microstructure evolution and mechanical properties of 6 mm thick DH36 steel. It was found that the microstructure became heterogeneous and mechanical properties declined when high rotational and traverse speeds (exceeding 650 RPM and 500 mm/min, respectively) were applied. In another paper by the same authors  on mathematical modelling of the FSW process using Computational Fluid dynamics (CFD) technique they reported that a maximum temperature of 1250°C was computed when the rotational and traverse speeds were 500 RPM and 400 mm/min, respectively. A comparison between FSW and Submerged Arc Welding (SAW) of DH36 showed microstructure and mechanical properties improved as a result of applying the FSW technique [3,4]. Advantages of FSW of DH36 steel was also studied by relating welding parameters to mechanical properties. Defects associated with FSW of DH36 steel were reported by Stevenson et al.  and classified into five groups, namely lower embedded, upper embedded, incomplete fusion, connectivity flaw and root flaw. Most of these flaws were suggested to have resulted from a lack in material flow due to high traverse speeds. Thermal stresses as result of local heating during FSW of DH36 were also studied [6,7] and were calculated numerically from a thermo-mechanical model. They were found to be below the manufactured yield stress of the alloy. The maximum measured temperature was also below the alloy’s melting point. It is noticed that previous work on FSW of DH36 steel did not investigate the possibility of elemental segregation as a significant type of defect which could deteriorate the mechanical properties of the welded joints. Also there exists no previous numerical or experimental measurement of the peak temperature during the FSW process that reported localised melting. In the current work, for the first time, the mechanism responsible for elemental segregation in FSW of 6 mm DH36 steel is investigated by applying two FSW tool speeds, W1 (200 RPM, 100 mm/min) and W2 (550 RPM, 400 mm/min). Alongside FSW, un-welded samples of DH36 steel were heat treated in a temperature range of 1200-1500°C and then subjected to cooling rates similar to that in FSW. This is in an effort to produce very similar microstructure to FSW samples in a controlled environment to be able to determine the temperature and cooling rate which cause elemental segregation. The heat treatments (1200°C-1500°C) were conducted with holding times between 1 and 10 minutes to understand how the segregation starts and advances. This work aims to determine the limits of FSW tool speed that cause elemental segregation and also to estimate the maximum peak temperature that the welded joint experienced during FSW.
The as-received chemical composition of the 6 mm thick DH36 steel is shown in Table 1. The top surface of plate is covered by a galvanised layer of Zn. The plates were friction stir welded at TWI (Yorkshire) using the Powerstir machine. A hybrid PCBN tool was used with a shoulder of 25 mm diameter and probe of 5.7 mm length. Samples were produced using low (200 RPM, 100 mm/min) and high (550RPM, 400 mm/min) speeds which named as W1 and W2 respectively from now on. Thermocouples were embedded in positions down the length of the weld in an attempt to establish the peak temperature achieved during welding. It is worth to mention that fixing thermocouples in the tool/work piece contact region or in the stirred zone can lead to errors in the temperature readings as thermocouples may move before recording the maximum temperature [TWI experiments].
Table 1: Chemical composition of as-received DH36 steel (wt%).
Furnace heat treatments of 24 samples of DH36 (6 mm cubic) were carried out between 1200°C and 1500°C at 50°C intervals and then quenched in either hot oil (150°C) or oil at ambient temperature (22°C) which produced cooling rates of (30°C/sec) and (50°C/sec) respectively. The surface temperature of the samples was checked by the attachment of W-Re thermocouples. Strain rate effect was not included in the heat treatments as the maximum temperature on Gleeble simulator hot compression test is not exceeding 1100°C and the current heat treatments are mainly in temperature ranges of 1400°C-1500°C. However, it is expected that strain rate will affect mainly the time of phase transformation and elemental precipitation/ segregation as proved by previous work.
Heat treated samples were sectioned, polished to 1 micron finish and etched in 2% Nital to reveal the microstructure. Samples were examined using both optical light microscopy and Scanning Electron Microscopy and Energy Dispersive X-Ray Analysis.
The SEM micrographs of the as-received base metal show a matrix of ferrite grains (7-9 μm) with bands of pearlite (Figure 1a), and nanoscale sized interstitial elements such as Ti and Nb (Figure 1b). Ca which has a strong affinity to form sulphides and oxides was found in the DH36 plate as a compound of calcium aluminates (CaO-Al2O3) and also calcium-manganese-sulphide (MnS-CaS) as shown in Figures 2 and 3 respectively. The melting point of (CaO-Al2O3) is in the range of 1450-1850°C depending on the amount of CaO compound with Al2O3 . Pure MnS melts at 1600°C but it is expected that the melting point is lower when it forms compounds with other elements . Hence, it is expected that some melting of the CaO-Al2O3 and also MnS-CaS occur when temperature exceeds 1400°C.
It is recognised that the samples heat treatments are nonequilibrium. The samples are heated between 1 to 10 min which is insufficient time for the whole sample to reach the same temperature. The heating temperature is designed to simulate the surface temperature that a FSW sample would experience.
Heat treatments up to a temperature of 1400°C followed by ambient and hot oil quenching showed significant segregation of MnS compound with Nb as shown in Figure 4 for quenching in ambient oil. The segregation of MnS and its detrimental effects on hot tensile fracture was studied by Santillana et al. . They found that segregation is occurring at a temperature range up to 1440°C which is close to the temperature in the current study. Figure 5 shows the microstructure of a sample heat treatment at 1450°C followed by hot oil quenching (at two different magnifications). It shows that significant segregation has occurred (Figure 5a) and was associated with acicular ferrite formation (Figure 5b). Comparing Figures 4 and 5 it is evident that the segregation of spinel elements of Mn, Si, Al and O only occurs at temperatures exceeding 1450°C with a cooling rate sufficient to cause the formation of an acicular ferrite microstructure. This phenomenon requires more in-depth investigation and the current work only focuses on the observation of this segregation. Figure 6 shows the microstructure of W1 and W2 respectively, both welds appear to have a microstructure of acicular ferrite and bainite, however, segregation of Mn, Si, Al, O elements appear only in the W2 joint stir zone. Figure 7 shows SEM-EDS of Mn, Si, Al, O, Zn segregation in W2. The Zn present in the 2nd phase particle shown is Figure 7 will have originated form the galvanised surface of the DH36 steel. Smith et al.  suggested from CFD modelling that the maximum peak temperature in FSW of 304 stainless steel is localised at the tool/work piece contact surface on the advancing trailing side and it can reach close to the melting point when the tool speed exceeds 400 RPM. The melting and boiling temperatures of Zn are 693 K (420°C) and 1180 K (907°C) respectively, welding peaks temperatures especially at the tool/work piece contact region is expected to be higher than the boiling temperature of Zn and thus the possibility of Zn presence in the stirred zone (SZ) should be low. In previous study [11,12] it is explained that the existence of Zn in the stirred zone as a result of a tool effect, whereby Zn will melt and evaporate during FSW but because there is no way for it to escape outside the tool/work piece contact region it will explode into the SZ. In this paper an interpretation to coalesce Zn with other elemental segregation in W2 FSW sample can be as follow: (a) The centrifugal forces from FSW tool are pushing the steel including the Zn coated layer at tool/work piece contact surface to the advancing trailing side during the FSW. (b) Heat generated during the FSW results in a peak temperature at the advancing-trailing side which exceeds 1400°C. (c) and at a cooling rate in the range of 10-30°C/sec, segregation is occur as shown previously by heat treatments for the peak temperature and cooling rate required for segregation. (d) The segregated elements are moved from the top surface of advancing-trailing side and distributed downward into the SZ under the effect of tool rotation and traverse speeds. This finding is in agreement with Arbegast et al.  who showed that material during FSW experiences upward flow into the shoulder and then downward flow into the extrusion zone around the tool. Thermocouples readings fixed at the bottom of the plates recorded peak temperatures of 850°C and 910°C for W1 and W2 respectively. These readings can give indicates that despite the high traverse speed of W2, temperature is higher than the W1, this mean that tool rotational speed is override the effect of traverse speed in term of heat generation. An interpolation of thermocouples reading into the stirred zone can give an error because of the difference between the heat transfer mechanism between the stirred zones in which convection is dominant and the region away in which the heat transfer is by conduction [13,14].
In an effort to understand the formation of the complex oxides which have been identified in the FSW joints a series of heat treatments were carried at 1450°C with holding times of between 1 min and 10 minutes. The temperature of 1450°C was chosen because it showed the limit of elemental segregation and below this limit no segregation is likely to occur as mentioned in the heat treatments. The purpose of varying the soak time in the furnace was to establish knowledge of segregation progress with increasing time. Following heat treatment all the samples were quenched in oil at 22°C which gives a cooling rate of approximately 30°C/sec. Following oil quenching, all samples were sectioned, polished and etched in preparation for microscopic examination of the microstructure. The results of the microstructural examination are presented in Figures 8a-8g. It can be seen in Figure 8a (1 min soaking time) that segregation appears to be starting with the joining of MnS-CaS with CaO-Al2O3 to form a complex compound as in Figure 8b which is an example of the chemical composition of the predominate second phase particles in the matrix after 3 minutes soak time. The segregation appears to be advancing with time by more elements joining the complex compound such as Si, Nb and P as shown in Figure 8c (5 min soaking time). Then a separation of sulphides from the main compound occurs including the formation of 2nd phase particles which are rich in calcium sulphide phosphorous and niobium sulphide as shown in Figures 8d and 8e respectively (7 min soaking time). The final compound with increasing holding time to 10 min elementally consists of Mn, Si, Al and O as shown in Figures 8f and 8g. This elemental segregation is expected to have a detrimental effects on the final welding joint mechanical properties as these elements can be a source of micro cracks and residual stresses and also the main matrix of steel is lost the benefit of uniform elemental distribution which were mainly added to steel during manufacturing in order to improve physical properties.
Figure 8: Steps of elemental segregation when heating at 1450°C and changing the soaking time (a) MnS-CaS+Al2O3-CaO (1 min soaking time); (b) forming one compound of MnS-CaS+Al2O3-CaO (3 min soaking time); (c) Si, Nb and P is incorporated in the compound (5 min soaking time); (d) Ca Sulphide; (e) Nb P Sulphide; (f) Mn, Si, Al and O(10 min soaking time); (g) Mn, Si, Al and O(10 min soaking time).
Figure 9 shows the elemental segregation of Mn, Si, Al, and O along a distance of 1 mm from the edges of the heat treated sample at 1500°C for holding time of 1 min. The middle of the sample did not show elemental segregation and this suggests that peak temperature away from the sample surface did not reach the maximum because of the limited holding time and also low thermal conductivity of steel.
It is worth highlighting that holding time for 1 min at 1500°C was enough for Mn, Al, Si and O elemental segregation at the sample surface and thus it is expected that the time required for segregation is decrease with increasing temperature. In FSW it is also expected that peak temperature at the advancing-trailing side is high enough for elemental segregation in a time of few seconds. It is also expected that the high strain rate which exist in the FSW will also reduce the time required for elemental precipitation/segregation as proved by previous work [15-19].
• SEM observations showed 6 mm-thick DH36 steel can be friction stir welded with a rotational speed limit up to 500 RPM and traverse speed of 400 mm/min without occurrence of any additional FSW related elemental segregation.
• From heat treatments experiments of 6-mm cubic DH36 steel, the peak temperature at which elemental segregation starts was determined as 1450°C with a cooling rate 30-50°C/s.
• SEM observations showed that MnS-CaS appear to coalesce around existing CaO-Al2O3 in DH36 microstructure. It appear that this evolution of the newly formed precipitates advanced with increasing soaking time by further joining of other elements such as Si, Nb and P. Finally, with extended time at 1450°C the Sulphur is separating from other sulphide compound such as phosphor sulphide, niobium phosphor sulphide and calcium sulphide leaving the final segregated compound composed of Mn, Si, Al and oxygen.
• Increasing the peak temperature of heat treatment to 1500°C reduced the holding time required to observe elemental segregation.
• The segregation of elements in the welded joints is expected to reduce mechanical properties which can be avoided by choosing the suitable tool rotational/traverse speed.
• Despite the fact that heat treatments did not include strain rate, they were useful to determine the limit of peak temperature at which elemental segregation of Mn, Si, Al and oxygen occur. Strain rate was approved by previous work to effect mainly the incubation time of precipitation/segregation and phase transformation.
• To carry out mechanical tests including tensile and fatigue in order to study the effect of elemental segregation on the FSW joints strength.
• To carry out heat treatments at higher temperatures and including the strain rate effect.
The Authors would like to thank TWI Company especially Steve Cater for providing data and samples.