Influence of Metallic Oxide Nanoparticles on the Mechanical Properties of an A-TIG Welded 304L Austenitic Stainless Steel

Austenitic stainless steels represent a significant aerospace material, being used for various castings, structural components, landing gear components, afterburners, exhaust components, engine parts, and fuel tanks. The most common joining process is tungsten inert gas (TIG) welding, which possesses many advantages such as suitability to weld a wide range of ferrous and non-ferrous metals and alloys, providing high quality welds with good mechanical properties. Its major disadvantage is low productivity due to low penetration and welding speed. This can be overcome by introducing an activating flux before welding. The activating flux reverses the material flow of the weld pool, significantly increasing penetration. Therefore, shielding gas consumption is reduced and welding without a consumable is enabled. However, the consumable in conventional TIG also enables the conditioning of the mechanical properties of welds. In this study, Si and Ti metallic oxide nanoparticles were used to increase the weld penetration depth, while bend testing, tensile, and impact toughness were determined to evaluate the mechanical properties of welds. Furthermore, optical emission spectroscopy, light, and scanning electron microscope were used to determine the chemical compositions and microstructures of the welds. Chemical compositions and weld mechanical properties were similar in all specimens. The highest tensile and impact properties were obtained with the specimen welded with the flux containing 20% TiO2 and 80% SiO2 nanoparticles. Although lower than those of the base metal, they were well within the nominal base metal mechanical properties.


Introduction
Austenitic stainless steels are used in various applications due to their excellent corrosion resistance in atmospheric and pure water environments, marine environments as well as in acid and alkaline solutions. Their advantages also present their unique properties such as hardness, strength, density, malleability, ductility at lower temperatures, elasticity, conductivity, thermal expansion, good heat, and fire resistance. Austenitic stainless steels are also easily fabricated using forming, machining, cutting, and welding, which makes them good materials for producing components in the aerospace industry. Common applications of austenitic stainless steels in aerospace applications include various castings, structural components, landing gear components, afterburners, exhaust components, engine Instead of aiming at optimizing the flux to achieve the maximum penetration as in our previous study [12], in this study, the influence of nanoparticle-based fluxes on the mechanical properties of the weld was determined. The obtained mechanical properties were compared to the base metal and procedures for determining the weld quality, which may be the impeding factor of the wider acceptance of A-TIG technology in the industrial sector. The main novelty identified was finding the effect of nanoparticles in the flux and their effect on the weld shape and mechanical properties of welds.

Materials and Methods
In this study, the base metal used was AISI 304L austenitic stainless steel. Its chemical composition is presented in Table 1, while its mechanical properties are shown in Table 2. The base material size was cut from the 10 mm plate to 200 mm × 50 mm dimensions by waterjet, and then butt welds were done without a gap. Prior to welding, several nanoparticle-based fluxes in acetone solvents were prepared. The active element of the fluxes was 20 nm TiO 2 and SiO 2 nanoparticles, weighed by a Tehtnica Type 2615 (Zelezniki, Slovenia) analytic balance, mixed with the solvent by means of a Tehtnica mm530 (Zelezniki, Slovenia) magnetic stirrer for 10 min. After mixing, the fluxes were smeared over the surface of the specimens to be welded by a brush, having the width of 20 mm at both base metal sides and were 10 µm thick, measured immediately after application. Subsequently, the welding was performed by using an EWM Tetrix300 TIG welding machine (EWM, Mündersbach, Germany). Welding current was 250 A and the WTh2 electrode tip was sharpened to 90 • . The distance from the tip to the base metal was 2 mm, while the welding speed was set and kept constant at 40 mm/min. All specimens were welded without a consumable and with Ar shielding gas at a 14 L/min flow rate as well as a ceramic backing plate. The welded specimens were designated in relation to the flux applied during welding: 5Ti (100% TiO 2 ), 4Ti1Si (80% TiO 2 ; 20% SiO 2 ), 3Ti2Si (60% TiO 2 ; 40% SiO 2 ), 2Ti3Si (40% TiO 2 ; 60% SiO 2 ), 1Ti4Si (20% TiO 2 ; 80% SiO 2 ), and 5Si (100% SiO 2 ).
Weld characterization was done in macro imagery, microstructures, microhardness and tensile, bend, and impact testing. Tensile testing was conducted in accordance to EN ISO 4136:2012 standard, impact energy in accordance to EN ISO 9016:2012, bend testing with EN ISO5173:2009, and microhardness with EN ISO 6507:2018. Specimens were cut by a waterjet, in accordance with Figure 1. In Figure 1, the weld direction is indicated. Metallographic preparation was done on Struers equipment (Bellerup, Denmark) and consisted of cutting, mounting, abrasive paper grinding (150 to 2000 grit), polishing with diamond suspensions (6 to 0.25 µm particle size), and etching by using aqua regia (three parts 37% HCl; 1 part 70% HNO 3 ). Examinations were done with a Leitz Orthoplan light microscope (Oberkochen, Germany) and JEOL JSM6460LV (Tokyo, Japan) scanning electron microscope (SEM). Microhardness was measured by a Wilson Tukon 1102 (Uzwil, Switzerland) device with 100 gf loading in the base metal and weld metal, as shown in Figure 2. The distance between the indentations was 0.5 mm. Weld metal chemical analysis was done by an ARL ISpark 8860 (Waltham, MA, USA) optical emission spectrometer (OES). Tensile and bend testing were done by the Schenck Hydropuls PSB 250 (Darmstadt, Germany) universal hydraulic testing machine, while impact testing was performed on a Testing Equipment IE JWT-450 (Jinan, China) instrumented Charpy pendulum tester. Image analyses to obtain the δ-ferrite content was done by ImageJ software (version 1.51).
Materials 2020, 13, x FOR PEER REVIEW 4 of 11 device with 100 gf loading in the base metal and weld metal, as shown in Figure 2. The distance between the indentations was 0.5 mm. Weld metal chemical analysis was done by an ARL ISpark 8860 (Waltham, MA, USA) optical emission spectrometer (OES). Tensile and bend testing were done by the Schenck Hydropuls PSB 250 (Darmstadt, Germany) universal hydraulic testing machine, while impact testing was performed on a Testing Equipment IE JWT-450 (Jinan, China) instrumented Charpy pendulum tester. Image analyses to obtain the δ-ferrite content was done by ImageJ software (version 1.51).

Metallographic Testing and Chemical Compositions
Macro-sections of the welded specimens are shown in Figure 3. Although similar in shape and columnar morphology, typical for A-TIG weldments, there was a marked difference in width. In all specimens, the root was significantly wider than the face, as was obtained in [12], which was the result of the heat accumulation and the ceramic backing plate, which limited excessive penetration. This indicates that the maximum penetration with applied parameters can be higher, that is, the present parameters can be applied to weld a thicker base metal of the same type. Alternatively, this means that the same thickness as applied in this experiment could be welded with a lower welding current, or a higher welding speed. The widest weld was obtained in the 1Ti4Si specimen. However, it is difficult to assess the maximum penetration of tested fluxes with the parameters used in this experiment just by evaluating the weld metal shape from the macro-sections. Moreover, it is difficult to compare the present macro-sections to the macro-sections obtained by the same fluxes, because in [12], different welding parameters were used. At higher currents, larger particles may be optimal for achieving high penetration, but this remains to be proven in future studies.  device with 100 gf loading in the base metal and weld metal, as shown in Figure 2. The distance between the indentations was 0.5 mm. Weld metal chemical analysis was done by an ARL ISpark 8860 (Waltham, MA, USA) optical emission spectrometer (OES). Tensile and bend testing were done by the Schenck Hydropuls PSB 250 (Darmstadt, Germany) universal hydraulic testing machine, while impact testing was performed on a Testing Equipment IE JWT-450 (Jinan, China) instrumented Charpy pendulum tester. Image analyses to obtain the δ-ferrite content was done by ImageJ software (version 1.51).

Metallographic Testing and Chemical Compositions
Macro-sections of the welded specimens are shown in Figure 3. Although similar in shape and columnar morphology, typical for A-TIG weldments, there was a marked difference in width. In all specimens, the root was significantly wider than the face, as was obtained in [12], which was the result of the heat accumulation and the ceramic backing plate, which limited excessive penetration. This indicates that the maximum penetration with applied parameters can be higher, that is, the present parameters can be applied to weld a thicker base metal of the same type. Alternatively, this means that the same thickness as applied in this experiment could be welded with a lower welding current, or a higher welding speed. The widest weld was obtained in the 1Ti4Si specimen. However, it is difficult to assess the maximum penetration of tested fluxes with the parameters used in this experiment just by evaluating the weld metal shape from the macro-sections. Moreover, it is difficult to compare the present macro-sections to the macro-sections obtained by the same fluxes, because in [12], different welding parameters were used. At higher currents, larger particles may be optimal for achieving high penetration, but this remains to be proven in future studies.

Metallographic Testing and Chemical Compositions
Macro-sections of the welded specimens are shown in Figure 3. Although similar in shape and columnar morphology, typical for A-TIG weldments, there was a marked difference in width. In all specimens, the root was significantly wider than the face, as was obtained in [12], which was the result of the heat accumulation and the ceramic backing plate, which limited excessive penetration. This indicates that the maximum penetration with applied parameters can be higher, that is, the present parameters can be applied to weld a thicker base metal of the same type. Alternatively, this means that the same thickness as applied in this experiment could be welded with a lower welding current, or a higher welding speed. The widest weld was obtained in the 1Ti4Si specimen. However, it is difficult to assess the maximum penetration of tested fluxes with the parameters used in this experiment just by evaluating the weld metal shape from the macro-sections. Moreover, it is difficult to compare the present macro-sections to the macro-sections obtained by the same fluxes, because in [12], different welding parameters were used. At higher currents, larger particles may be optimal for achieving high penetration, but this remains to be proven in future studies.
where typical dendritic weld metal and transition weld interface microstructures were obtained. The SEM microstructure revealed the typical δ-ferrite morphology that occurs between austenitic dendrites. The SEM micrograph is shown in Figure 5, presenting δ-ferrite in the austenite matrix of the weld metal. There were no differences between the microstructures obtained with different experimental conditions. The δ-ferrite content in the base metal was 6%, while in the weld metal of all specimens, it was 7.63% to 8.01%.    In Table 3, the widths of the weld metals are shown. The addition of SiO 2 nanoparticles makes the weld face narrower due to the arc constriction effect of the Si, with the only exception being 1Ti4Si. This is in contrast to the results shown in [12], where the 3Ti2Si specimen proved to have the highest penetration by using the same fluxes. However, it must be noted that the current used in this study was 300 A, which is considerably higher than that in [12], where it was 200 A. Obviously, 1Ti4Si represents the optimal mixture from the point of view of the welding parameters, most notably the welding current. The microstructures obtained with the light microscope of specimen 5Ti is shown in Figure 4, where typical dendritic weld metal and transition weld interface microstructures were obtained. The SEM microstructure revealed the typical δ-ferrite morphology that occurs between austenitic dendrites. The SEM micrograph is shown in Figure 5, presenting δ-ferrite in the austenite matrix of the weld metal. There were no differences between the microstructures obtained with different experimental conditions. The δ-ferrite content in the base metal was 6%, while in the weld metal of all specimens, it was 7.63% to 8.01%.
Materials 2020, 13, x FOR PEER REVIEW 5 of 11 In Table 3, the widths of the weld metals are shown. The addition of SiO2 nanoparticles makes the weld face narrower due to the arc constriction effect of the Si, with the only exception being 1Ti4Si. This is in contrast to the results shown in [12], where the 3Ti2Si specimen proved to have the highest penetration by using the same fluxes. However, it must be noted that the current used in this study was 300 A, which is considerably higher than that in [12], where it was 200 A. Obviously, 1Ti4Si represents the optimal mixture from the point of view of the welding parameters, most notably the welding current.
The microstructures obtained with the light microscope of specimen 5Ti is shown in Figure 4, where typical dendritic weld metal and transition weld interface microstructures were obtained. The SEM microstructure revealed the typical δ-ferrite morphology that occurs between austenitic dendrites. The SEM micrograph is shown in Figure 5, presenting δ-ferrite in the austenite matrix of the weld metal. There were no differences between the microstructures obtained with different experimental conditions. The δ-ferrite content in the base metal was 6%, while in the weld metal of all specimens, it was 7.63% to 8.01%.    Chemical compositions of weld metals obtained by OES as well as the 304L steel nominal chemical composition are shown in Table 4. It can be seen that all obtained chemical compositions of weld metals corresponded well to the nominal chemical composition of 304L austenitic stainless steel. Additionally, the weld metal chemical compositions corresponded well to the chemical composition of the base metal shown in Table 1, indicating that the flux does not influence changes in the chemical composition of the weld metal.

Mechanical Properties of Welds
The tensile properties of the welds are shown in Figure 6, while the typical fracture mode is shown in Figure 7. Tensile properties in all specimens were lower compared to the base metal, which could be the result of the replacement of polygonal austenitic by a cast dendritic microstructure, that is, a polygonal with dendritic austenite. Tensile strengths and reductions of area were well within the limits set up by ASTM A240/A240M relating to 304L austenitic stainless steel. The highest tensile strength and reduction in area of all welded specimens were obtained in the 1Ti4Si specimen, as the result of the widest weld metal and the widest space for deformation. This is also important from the point of view of penetration potential, since this mixture influences the largest weld metal. Fracture mode is fully ductile, with a notable necking, as shown in Figure 7. This was proven by the SEM image analysis (Figure 8). A dimpled specimen fracture surface was obtained, typical of ductile fracture. This is the result of microvoid formation around non-metallic inclusions, microvoid growth, and coalescence [24]. At the bottom of the dimples, circular phases can be observed, which can be morphologically attributed to non-metallic inclusions (Figure 8c,d). Chemical compositions of weld metals obtained by OES as well as the 304L steel nominal chemical composition are shown in Table 4. It can be seen that all obtained chemical compositions of weld metals corresponded well to the nominal chemical composition of 304L austenitic stainless steel. Additionally, the weld metal chemical compositions corresponded well to the chemical composition of the base metal shown in Table 1, indicating that the flux does not influence changes in the chemical composition of the weld metal.

Mechanical Properties of Welds
The tensile properties of the welds are shown in Figure 6, while the typical fracture mode is shown in Figure 7. Tensile properties in all specimens were lower compared to the base metal, which could be the result of the replacement of polygonal austenitic by a cast dendritic microstructure, that is, a polygonal with dendritic austenite. Tensile strengths and reductions of area were well within the limits set up by ASTM A240/A240M relating to 304L austenitic stainless steel. The highest tensile strength and reduction in area of all welded specimens were obtained in the 1Ti4Si specimen, as the result of the widest weld metal and the widest space for deformation. This is also important from the point of view of penetration potential, since this mixture influences the largest weld metal. Fracture mode is fully ductile, with a notable necking, as shown in Figure 7. This was proven by the SEM image analysis (Figure 8). A dimpled specimen fracture surface was obtained, typical of ductile fracture. This is the result of microvoid formation around non-metallic inclusions, microvoid growth, and coalescence [24]. At the bottom of the dimples, circular phases can be observed, which can be morphologically attributed to non-metallic inclusions (Figure 8c,d).          The highest tensile strength is the result of the highest absolute amount of δ-ferrite in the austenitic matrix, which is proportional to the weld metal width, most particularly the weld metal width, as shown in Table 3. The body-centered cubic crystal lattice (BCC) had higher mechanical properties compared to the austenitic face-centered cubic lattice. This is similar to the findings of Hdhibi et al. [22], who also stressed the significance of δ-ferrite in obtaining the convenient tensile properties of 316L austenitic stainless steel A-TIG remelting. Furthermore, the tensile strengths obtained in this study were very similar to those obtained in Hdhibi et al. [22], who used 316L austenitic stainless steel welded/remelted by A-TIG with the flux based on SiO 2 and TiO 2 . However, in this study, a positive correlation between SiO 2 nanoparticle content and weld metal root width and tensile strength could be observed. Additionally, when single component fluxes are considered, it was found that the specimen welded with the flux containing SiO 2 nanoparticles exhibited a higher tensile strength and a lower reduction in area compared to the specimen welded with the flux containing TiO 2 nanoparticles. Impact energies obtained by the instrumented Charpy pendulum tester, with V-notch situated in the center of the weld metal are shown in Figure 9. As in base metal (Table 2), crack propagation values were higher compared to the crack initiation energy values. However, all impact energy values were lower than those of the base metal. Values for overall impact energies were between 80% and 95% of the base metal. The highest overall impact energies were obtained in specimens 2Ti3Si and 1Ti4Si. The presence of the less ductile δ-ferrite might explain why weld metals have a slightly lower impact energies compared to the base metal. Figure 9a,b depicts the typical force-time curves from the instrumented Charpy pendulum tester of the base metal and 3Ti2Si specimen. It can be seen that the curve shape was very similar, however, the overall impact energy of the base metal specimen was higher (blue accumulative curve).
Materials 2020, 13, x FOR PEER REVIEW 8 of 11 The highest tensile strength is the result of the highest absolute amount of δ-ferrite in the austenitic matrix, which is proportional to the weld metal width, most particularly the weld metal width, as shown in Table 3. The body-centered cubic crystal lattice (BCC) had higher mechanical properties compared to the austenitic face-centered cubic lattice. This is similar to the findings of Hdhibi et al. [22], who also stressed the significance of δ-ferrite in obtaining the convenient tensile properties of 316L austenitic stainless steel A-TIG remelting. Furthermore, the tensile strengths obtained in this study were very similar to those obtained in Hdhibi et al. [22], who used 316L austenitic stainless steel welded/remelted by A-TIG with the flux based on SiO2 and TiO2. However, in this study, a positive correlation between SiO2 nanoparticle content and weld metal root width and tensile strength could be observed. Additionally, when single component fluxes are considered, it was found that the specimen welded with the flux containing SiO2 nanoparticles exhibited a higher tensile strength and a lower reduction in area compared to the specimen welded with the flux containing TiO2 nanoparticles.
Impact energies obtained by the instrumented Charpy pendulum tester, with V-notch situated in the center of the weld metal are shown in Figure 9. As in base metal (Table 2), crack propagation values were higher compared to the crack initiation energy values. However, all impact energy values were lower than those of the base metal. Values for overall impact energies were between 80% and 95% of the base metal. The highest overall impact energies were obtained in specimens 2Ti3Si and 1Ti4Si. The presence of the less ductile δ-ferrite might explain why weld metals have a slightly lower impact energies compared to the base metal. Figure 9a,b depicts the typical force-time curves from the instrumented Charpy pendulum tester of the base metal and 3Ti2Si specimen. It can be seen that the curve shape was very similar, however, the overall impact energy of the base metal specimen was higher (blue accumulative curve). Although impact energies (crack initiation, propagation, and overall impact energies) were slightly lower in the welded specimens compared to the base metal, bend testing did not result in Although impact energies (crack initiation, propagation, and overall impact energies) were slightly lower in the welded specimens compared to the base metal, bend testing did not result in any cracks, on either side of any specimen tested ( Figure 10). This is in accordance with the relatively high values in the reduction of area (Z) obtained in tensile testing ( Figure 6). The deformed area of the weld metal, despite the fact that the surface was ground prior to testing, showed signs of columnar morphology, which is in accordance with the macro images ( Figure 2). The values obtained in all specimens, with crack propagation energies higher than crack initiation energies, proved that the suitability of welds to pressure vessel application is maintained. Namely, a higher crack propagation energy ensures a minimized sudden failure of the weld in the case of overpressure occurrence.
Materials 2020, 13, x FOR PEER REVIEW 9 of 11 any cracks, on either side of any specimen tested ( Figure 10). This is in accordance with the relatively high values in the reduction of area (Z) obtained in tensile testing ( Figure 6). The deformed area of the weld metal, despite the fact that the surface was ground prior to testing, showed signs of columnar morphology, which is in accordance with the macro images ( Figure 2). The values obtained in all specimens, with crack propagation energies higher than crack initiation energies, proved that the suitability of welds to pressure vessel application is maintained. Namely, a higher crack propagation energy ensures a minimized sudden failure of the weld in the case of overpressure occurrence. The microhardness results are shown in Figure 11. In the majority of specimens, the hardness of the weld metal was higher compared to the hardness of fusion zones, probably due to a higher content of δ-ferrite and a lower distance between δ-ferrite phases: grain size in base metal 20-100 μm and interdendritic distance in weld metal 10-40 μm. This is depicted in Figure 4a,b. Although the differences were small, the microhardness of specimens 2Ti3Si and 1Ti4Si were the lowest (Figure 11c), which is interesting, since the tensile properties and impact energies of these specimens were the highest. Apparently, there was an optimal δ-ferrite content, which provided the highest mechanical properties.  The microhardness results are shown in Figure 11. In the majority of specimens, the hardness of the weld metal was higher compared to the hardness of fusion zones, probably due to a higher content of δ-ferrite and a lower distance between δ-ferrite phases: grain size in base metal 20-100 µm and interdendritic distance in weld metal 10-40 µm. This is depicted in Figure 4a,b. Although the differences were small, the microhardness of specimens 2Ti3Si and 1Ti4Si were the lowest (Figure 11c), which is interesting, since the tensile properties and impact energies of these specimens were the highest. Apparently, there was an optimal δ-ferrite content, which provided the highest mechanical properties.
Materials 2020, 13, x FOR PEER REVIEW 9 of 11 any cracks, on either side of any specimen tested ( Figure 10). This is in accordance with the relatively high values in the reduction of area (Z) obtained in tensile testing ( Figure 6). The deformed area of the weld metal, despite the fact that the surface was ground prior to testing, showed signs of columnar morphology, which is in accordance with the macro images ( Figure 2). The values obtained in all specimens, with crack propagation energies higher than crack initiation energies, proved that the suitability of welds to pressure vessel application is maintained. Namely, a higher crack propagation energy ensures a minimized sudden failure of the weld in the case of overpressure occurrence. The microhardness results are shown in Figure 11. In the majority of specimens, the hardness of the weld metal was higher compared to the hardness of fusion zones, probably due to a higher content of δ-ferrite and a lower distance between δ-ferrite phases: grain size in base metal 20-100 μm and interdendritic distance in weld metal 10-40 μm. This is depicted in Figure 4a,b. Although the differences were small, the microhardness of specimens 2Ti3Si and 1Ti4Si were the lowest (Figure 11c), which is interesting, since the tensile properties and impact energies of these specimens were the highest. Apparently, there was an optimal δ-ferrite content, which provided the highest mechanical properties.

Conclusions
According to the presented results and within the limitations of this study, the following conclusions can be drawn: • Chemical compositions of the weld metals are not dependent on the composition of the flux. • Different SiO 2 to TiO 2 ratios result in obtaining similarly shaped weld metals, with the same columnar structure, but with different widths.

•
The widest weld resulted in the highest absolute amount of δ-ferrite, the widest space for deformation, resulting in the highest strength and ductility.

•
Tensile properties and impact toughness of weldments were marginally lower compared to the base metal, due to the replacement of strain hardened polygonal austenite by the mixture of cast dendritic austenite.

•
Tensile strength and reduction of area as well as impact toughness of the optimal specimen and the majority of other specimens, were within the limits set by the base metal standard limits.

•
Bend testing did not result in the occurrence of any cracks.

•
Microhardness values in the weld metal in the majority of specimens were higher compared to the fusion zones. Additionally, specimens that had the lowest average microhardness had the highest tensile and impact toughness.