Mixing Design for ATIG Morphology and Microstructure Study of 316 L Stainless Steel

This work is a study of the effects of oxides combination on the morphology of 316L stainless steel welds. A series of thirteen weld lines were carried out using thirteen different oxides. Based on the depth and ratio D/W results, three candidate oxides were selected: Ti2O, Mn2O3, and SiO2. Mixing method available in Minitab 17 software is the most appropriate method to find the optimal combinations to get the best depth and D/W ratio. According to simplex lattice degree four, nineteen combinations of these oxides were prepared. The results show that the optimal composition of flux was: 66%SiO2-34% Mn2O3. The depth and D/W ratio increased to 8.85mm and 0.98 respectively for optimal ATIG, whereas for the conventional TIG welding, the depth and the ratio D/W didn't exceed 1.65mm and 0.17 respectively. For TIG weld joint the hardness is about 47 HRA and it increases to 77 HRA for the optimal ATIG weld joint. The absorbed energies in Charpy impact test are 146 and 138kJ in the weld zone and in the heat affected zone respectively for the TIG welding and they dropped to 111 and 74kJ for the optimal ATIG welding. The fracture surface examined by scanning electron microscope (SEM) shows a ductile fracture for TIG weld with small dimples but ductile-brittle fracture for ATIG weld. Energy dispersive spectroscopy (EDS/SEM) analysis shows the formation of FeS2 and SiO2 in the weld zone causing low absorbed energy for ATIG weld. Keywords-ATIG; 316L austenitic stainless steel; ATIG welding mixing method; ternary flux


INTRODUCTION
TIG welding process is quite widespread.However, weld penetration varies from cast to cast and the maximum weld depth in a single pass does not exceed 3mm.Activated TIG (ATIG) welding proposed by Paton Institute-Kiev [1,2] seems an appropriate process to overcome these weaknesses.ATIG welding uses the same equipment as TIG except a prior weld operation layer of activated flux paste is deposited on the workpiece.Hypotheses have been made to explain the phenomenon favoring a high weld depth.Authors in [3][4][5][6] proposed that the Marangoni convection is the most important factor in determining weld shape.Surfactant elements as S, Se, Te, O in weld pool affect the convection of molten metal leading to centripetal circulation of weld metal getting a depth of weld.Authors in [7,8] applied the principle of electron absorption [9], electronegative elements as halides (fluorine) migrate to arc weld and capture outer arc electrons which contribute to contract the arc welding.This phenomenon increases the density of arc welding and the heat provided to weld pool.Authors in [10][11][12] proposed that high electrical resistance of flux components contributes to arc constriction.Many works were subsequently carried out using oxides like SiO2, TiO2, Cr2O3, Fe2O3, CaO, ZrO2, MgO [13][14][15][16][17][18][19][20][21][22][23][24], chlorides [25][26][27], and fluorides [28,29] to improve weld depth.These studies showed an increasing in the weld depth without affecting the mechanical properties of the specimens [30,31].Other studies were dedicated to microstructure and mechanical properties [32][33].In this work, 13 oxides were investigated on 316L austenitic stainless steel.Design of experiments was used to predict the best combination of the tested oxides.Mixing method based on a simplex lattice degree four was the key tool to optimize the flux combination in order to enhance the depth and D/W ratio.

A. Materials
The material used was the 316L stainless steel whose chemical composition is shown in Table I.Experiments consisted of welding a 20cm line on a rectangular plate of 6mm thickness.Before welding, the plates were cleaned with acetone.Flux powders were dried in furnace to eliminate humidity, and then mixed with methanol in 1:1 ratio.

B. Welding Procedure
A Tungsten Inert Gas welding machine was used.A watercooled torch with a standard 2% thoriated tungsten electrode rod having diameter of 3.2mm has been used for the experiments.The torch was mounted on a motorized carriage as shown in Figure 1.A series of tests were carried out with 150A welding current and 15cm/min welding speed leading to a provided energy of 540kJ/m.The welding parameters for the TIG welding are listed in Table II.

C. Mechanical Testing
After welding, the samples were cut as shown in Figure 2 to acquire test samples for mechanical testing and weld morphology.The Rockwell hardness measurement was performed with a Digital Hardness Tester Model HRS-150.The hardness tests were conducted in the bulk of weld bead, and the results were the average of four indentations as shown in Figure 3.  tested.Among these oxides, 3 gave the best depth and D/W ratio and were selected to be used in the mixing design method.
Based on the simplex lattice degree four designs, 19 combinations have been selected.Nineteen weld lines were executed and samples for morphology, hardness and impact tests were cut.The optimizer module available in mini tab 17 was used to get the optimal formulation of flux which is expected to give the best depth and D/W ratio of weld.Finally, using the optimal formulation, ATIG weld line has been carried out along with another weld line without flux (TIG).

A. Depth D and D/W Ratio
Figure 5 shows the weld depth results of ATIG weldment of the tested preliminary 14 oxides.The highest value of depth is 6.04mm which is obtained for the sample welded with Mn2O3 flux, followed by the sample welded with SiO2 flux (5.29mm), and then by the sample welded with TiO2 (4.81mm).So the oxides Mn2O3, SiO2, and TiO2 were selected.The results of the depth and D/W ratio for different proportions of the 3 selected oxides are shown in Table III.

C. Confirmation Test
The confirmation test is the last step in the experimental process.Confirmation tests were conducted based on the optimum flux composition obtained in the mixing design method of Minitab.The experimental values of depth and D/W ratio of the optimal flux ATIG weld are reported in Table VI.

www.etasr.com Touileb et al.: Mixing Design for ATIG Morphology and Microstructure Study of 316L Stainless Steel
The results show the depth of optimized flux is greater than TIG weld by 5 times.The ratio D/W improved 6 times.The experimental values obtained for depth in ATIG and TIG welds are shown in Figure 8.The penetration weld of optimized flux (8.85mm) is greater than the maximum value predicted by the software (7.01mm).Furthermore, the D/W value obtained for ATIG weld (0.98) is very close to the maximum predicted value of 0.96 as shown in Figure 9. TIG molten metal behaves as pure metal with centrifugal convection leading to wide and sallow weld bead contrariwise.ATIG weld metal has a centripetal movement related to oxygen liberated from oxides during welding operation.The quantity of liquid melted in ATIG (51.51mm 2 ) is higher than in TIG weld bead (14.2mm 2 ) as shown in Figures10-11.

D. Hardness Test
The experimental values obtained for hardness are shown in Figure 12.ATIG hardness number of HRA=77 is greater than TIG hardness number of HRA=47.

E. Impact Test
The experimental values obtained from the impact test are shown in Figures 13-14   The energies absorbed either in the weld zones or in HAZ of TIG welds are higher than the ones of ATIG welds.Figure 15 shows the images of fractograph of the "V" notch Charpy impact test conducted on scanning electronic microscope (SEM).The images show the formation of micro-voids, with multiple dimples which demonstrate that the fracture is in ductile mode in the cases of TIG weld zone (a) and in TIG heat affected zone (c).Combined ductile (small colonies with finer dimples) and brittle fracture mode formation characterizes the ATIG weld zone (b) and the ATIG affected zone (d).The combined ductile and brittle fracture mode leads to poor resistance to sudden impact loads.

F. Microstructure Study
Microstructure study was performed in order to view the material properties after the welding in both TIG and ATIG weldings.

1) EDS/SEM Analysis
The results of EDS/SEM analysis in Figures 16 and 17 show the presence of silicon, sulphur and oxygen in ATIG weld bead and in the heat affected zone, that lead to formations of inclusions as FeS2.Moreover, oxygen initiates microvoids and propagates by micro void coalescence, which facilitate the fracture phenomenon.These results can explain the low impact resistance in ATIG weld.The results of EDS/SEM analysis in Figures 18 and 19 show the presence of silicon in TIG weld zone in quantities close to the quantity in the received plate.The weld zone and the HAZ are free from oxygen and sulphur which can explain the withstanding of TIG weld to sudden impact.The results obtained with EDS/SEM regarding TIG weld may explain the high resistance to impact test.The results of EDS/SEM analysis in Figures 18 and 19 show the presence of silicon in TIG weld zone.The weld zone and the HAZ are free from oxygen and sulphur which can explain the widstand of TIG weld to sudden impact.The results obtained with EDS/SEM regarding TIG weld may explain the high resistance to impact test.

2) Ferrite Proportions
The measurement of ferrite content in the austenite matrix was conducted using area image processing software of Microvision Instruments.The weld zone appearance of ATIG and TIG was examined by SEM.The measurements were taken in 4 different locations in the weld zone and the results were the average of these measurements.Delta ferrite morphology has a vermicular network structure in both ATIG and TIG welds as shown in Figure 20.Results show ferrite content equal to 10.4% for ATIG.Instead, there is a decrease in ferrite content to 6.8% for TIG.This can be explained by the fact that in the case of ATIG operation, the weld bead is fully penetrated with a large weld bead area up to 55mm 2 .During ATIG weld operation, the activated flux has a beneficial effect in increasing the heat provided to weld pool and consequently the peak temperature in ATIG weld is greater comparatively to TIG weld leading to increase the proportion of delta ferrite retained in austenite matrix.On the other hand, in TIG welding, the bead is wide and shallow so the cooling rate is high.As the cooling rate is fast the transformation of delta ferrite to austenite is not completed, resulting on the presence of delta ferrite in the weld bead.The ferrite content in heat affected zone is up to 4.6% in ATIG welding and decreases at 1.4% in TIG heat affected zone.In TIG, the delta ferrite is randomly distributed in globular form or in small slats as shown in Figure 21(b).However, in ATIG heat affected zone, delta ferrite looks-like elongated inclined stringers as shown in Figure 21(a  IV.CONCLUSION This study was conducted on ATIG welding of austenitic 316L.The results are summarized as follows: • Among 13 oxides tested, SiO2, TiO2 and Mn2O3 present the highest depth penetration.
• The size of ATIG weld bead is bigger than the one of TIG weld bead.The activated flux enhances the heat provided to the weld pool leading to high proportion of retained delta ferrite (10.4%).However, TIG weld is wider and shallower so the molten metal cools rapidly in comparison with ATIG molten metal.In the case of TIG weld the transformation of delta ferrite to austenite is stopped earlier due to the fast cooling giving delta ferrite proportions up to 6.8%.
• The hardness HRA property of ATIG weld is improved comparatively to TIG weld.This result is related with the high delta ferrite content in ATIG weld zone.
• EDS/SEM analysis reveals the presence of FeS2, silicon, and oxygen in ATIG welded zone which could be the reason of the decrease of sudden impact load in comparison with TIG weld.
• Mixing design method is the more appropriate tool to get optimal flux composition when depth and D/W ratio improvements are desirable.Based on experiments' results, one can say that Minitab software can predict accurately the responses.

Fig. 4 .
Fig. 4. Test specimen for impact test (a) weld zone, (b) HAZD.Experiment by Mixing Method Design (DOE)For the DOE, mixing method in mini tab 17 was applied and consequently the number of experiments to be conducted was reduced.In order to compare the effect of the flux on the tensile strength, 13 oxides (SiO2, TiO2, Fe2O3, Cr2O3, ZnO, ZrO2, CaO, Mn2O3, V2O5, MoO3, SrO2, CoO3, and MgO) were

Fig. 18 .
Fig. 18.EDS/SEM spectrum analysis of fracture face of TIG welded zone

TABLE III .
DIFFERENT COMPOSITIONS OF FLUXES, DEPTH AND D/W Exp.No. SiO2 % TiO2 % Mn2O3 % Depth (mm) D/W B. Mixture Contour Plot Mixture contour plots for depth D and D/W were generated using Minitab software and are shown in

TABLE IV .
PREDICTED RESPONSES (UTS, HRA) OF WELD EXECUTED WITH OPTIMAL FLUX

TABLE V .
OPTIMAL COMPOSITION OF FLUX.

TABLE VI .
MORPHOLOGY OF TIG AND ATIG WELDS.

TABLE IX .
ELEMENTS PRESENT IN FRACTURE FACE OF TIG WELDED ZONE Fig. 19.EDS/SEM spectrum analysis of fracture face of TIG HAZ