Figure 2 clearly shows the characteristics of the base material microstructures of 304 stainless steel and Q235 steel. For 304 stainless steel (
Figure 2a), the matrix microstructure is mainly composed of austenite; this austenite has relatively coarse grains, the grain microstructure is compact and uniform in size, and there are no precipitates at the grain boundaries. Notably, many twin structures exist within the austenitic grains, which are arranged along the rolling direction, and a small amount of residual high-temperature δ-ferrite is also visible, which helps increase the stability of the austenitic steel. The microstructure of the Q235 steel is mainly composed of ferrite and pearlite, as shown in
Figure 2b. Ferrite appears as a gray polyhedral microstructure, whereas pearlite appears as a black polyhedron, which is relatively uniformly distributed in the ferrite grain boundaries and their interiors. Further observation of
Figure 2c reveals that the pearlite has a typical lamellar morphology with lamellar spacings between 0.8 and 1.1 μm, in which the carburite is intermittently rod-like or granular.
3.1. Q235/Weld Interface Microstructure Analysis
Figure 3 shows the metallographic microstructure of the Q235/weld near the fusion line obtained under different bias arcs.
Figure 3a shows that the location of the fusion line between the Q235 base metal and the weld metal is more pronounced and that there is a clear heat-affected zone. The heat-affected zone and the Q235 base metal are on the left side of the fusion line, and the weld seam is on the right side. Since the magnetic permeability of mild steel is greater than that of stainless steel, the arc will be biased more toward mild steel. During the welding process, even if the tungsten electrode is located directly above the weld seam, the arc will be more biased toward the Q235 side, the Q235 base metal will melt more than 304 stainless steel, and the heat-affected zone will be larger at the same time. The width of the heat-affected zone is approximately 2 mm for sample 1# and 1.7 mm for sample 2#. When the arc is biased to the Q235 side, the heat input to the Q235 base material is greater, the heat-affected zone is further expanded, as shown in
Figure 3c, and the width of the heat-affected zone is greater than 2 mm.
The OM map reveals that the bias of the arc does not lead to changes in the microstructure of the heat-affected zone, which is mainly composed of ferrite (F) and pearlite (P), and this compositional ratio changes significantly with decreasing distance from the fusion line, as evidenced by a gradual decrease in the content of pearlite. Immediately to the left of the fusion line, there is a region of significant superheating, where carbides precipitate in an ordered manner within the austenite grains and tend to extend in a needle-like fashion; ultimately, these structures progressively develop into coarse Weiss tissue with a complex lattice or feathery texture. The evolution of the microstructure exhibits continuous characteristics, as evidenced by a clear downward trend in the amount of pearlite as the fusion line approaches. When the arc deflects to the 304 side, the heat-affected zone on the Q235 base metal side is small, and the proportion of pearlite in the heat-affected zone is larger and closer to the base metal than that of sample 1#. The edge of the molten pool is near the base metal because the liquid metal residence time is shorter, and the mechanical stirring effect is weakened; so, the composition of the weld in this region is closer to that of the base metal.
Figure 4 shows the elemental-scanning results of the fusion line position on the Q235 side. The elemental-line-scanning analysis of the interface between the Q235 and the weld clearly reveals that key elements, such as Cr and Ni, exhibit significant concentration gradient changes in the region adjacent to the fusion line of the joint. Specifically, in the narrow region adjacent to the fusion line on the Q235 side of the joint, the fluctuation in the elemental concentration is relatively smooth, indicating a relatively uniform elemental distribution. However, once close to the fusion line on the weld side, the change in the elemental concentration becomes particularly drastic: the content of Cr and Ni increases dramatically, and this transition constitutes a significant concentration gradient interface.
In
Figure 4a, we observe a significant increase in the content of Cr- and Ni-alloying elements as the scanning direction moves toward the center of the weld, especially for Cr, which significantly increases in content close to the fusion line. In contrast, the change in the content of Ni is slightly greater. This phenomenon is explained by the fact that the high Cr and Ni contents in the 304 stainless steel are diluted into the weld. Moreover, there was a slight decrease in the Fe content, which was caused by the increase in the Cr and Ni contents.
3.2. Weld/304Interface Microstructure Analysis
On the 304 stainless steel side shown in
Figure 5, the location of the fusion line can be clearly observed via the OM and SEM map analysis. This region demonstrates that the heat-affected zone is mainly composed of austenitic grains, but the grain size distribution is not uniform. This nonuniformity stems from the differences in the degree of mixing of the weld metal with the base metal in the molten pool and at the edges when welding Q235 to 304 stainless steel. In particular, close to the edge of 304 stainless steel, the lower temperature, restricted metal mobility, and short liquid-residence time result in a unique fusion zone where the molten metal and base metal fail to fully fuse at the interface. As shown in the figure, the fusion zone is formed on the 304 side. The fusion zone is the weakest part of the welded joint and is prone to hot and cold cracks. In actual industrial production, because the main weak link of the joint is generally the fusion zone, it is necessary to strictly control the width of the fusion zone. During low-carbon steel welding, the welding-process-induced fusion zone is heated to the highest temperature, reaching 1400~1539 degrees. Celsius metal partially melts; after crystallization, the material does not melt but rather experiences heated growth of coarse grains and new crystallization of the cast.
Figure 5 shows that the widths of the joints welded in joints No. 1 and No. 2 in the fusion zone are between 200 and 300 μm. The welded joints in the fusion zone of mild steel are very narrow, generally approximately 0.1~1 mm, but they largely determine the mechanical properties of the welded joints.
At the junction of 304 stainless steel and the weld, along the boundaries of the austenitic grain boundaries, there is a fine distribution and a discontinuous state of worm-like ferrite. Further observation of the heat-affected zone reveals that the microstructure consists mainly of austenitic equiaxial crystals, and banded ferrite is scattered on these grains. The size of these grains does not change significantly compared with that of the parent material, but the amount of banded ferrite in the heat-affected zone is significantly greater than that of the parent material.
When 304 stainless steel is welded with Q235 carbon steel, the enrichment of the weld Cr reduces the carbon activity, resulting in the tendency of carbon to diffuse from the Q235 to the weld. The SEM observations revealed the distribution of ferrite on the austenitic matrix near the fusion line. Ferrite can eliminate single-phase austenite directionality, refine grains, and reduce intergranular segregation, thus enhancing the crack resistance and intergranular corrosion resistance of welded joints, which is a key factor in improving the welding quality.
Figure 6, on the other hand, shows the line scan direction opposite to the Q235 side, i.e., from the center of the weld toward the 304 stainless steel side. Along this path, the content of Cr- and Ni-alloying elements gradually increases, whereas the content of Fe decreases accordingly. This reflects the dilution of Cr- and Ni-alloying elements in 304 stainless steel by Q235 carbon steel during the dissimilar-steel-welding process.
3.3. Weld Metal Microstructure Analysis
According to the results of the elemental content test of the energy spectrometer, the Cr contents in the scanning regions of the 1#, 2#, and 3# welds were 3.37, 7.81, and 3.26, respectively, and the Ni contents were 1.41, 3.33, and 1.36, respectively, as shown in
Figure 7. Combined with Formulas (1) and (2) of Cr-equivalent (Creq) and Ni-equivalent (Nieq) in Schaeffler’s phase diagram, the Cr equivalents in the weld regions of samples 1#, 2#, and 3# were 3.7, 8.2, and 3.57, respectively. The Nieq value was calculated to be a maximum of 9.93 on the basis of the maximum carbon content of 0.22 wt% in the Q235 plate. The results of the Creq and Nieq calculations can be combined with the Schaeffler diagram [
20] (
Figure 8) to determine that the microstructure within the welded joint is martensite.
The Q235 and 304 dissimilarity welding weld is a low-carbon steel; its weld-lath martensite formation originates from the specificity of the welding thermal cycle, and rapid cooling leads to nonequilibrium solidification. The rapid cooling process of steel in the austenitic state of diffusion-type decomposition is inhibited, a diffusion-free phase transition process occurs, and martensite microstructure is eventually formed at room temperature. The weld-center microstructure verified the theoretical predictions. However, stainless steel welding is prone to produce harmful precipitates, such as the σ-phase and M
23C
6 phases, at 427~800 °C, which originates from the sensitization phenomenon of chromium-nickel austenitic stainless steels, leading to chromium depletion at the grain boundaries and a reduction in corrosion resistance. Compared with the 1# sample, the 3# sample with the arc bias toward the Q235 side exhibited fine martensite bars of approximately the same size oriented in parallel, forming large martensite bundles, i.e., lath martensite. The 2# sample with the arc bias toward the 304 side formed a fine, needle-shaped, and lath martensite microstructure, as shown in
Figure 9. The fine lath martensite can increase the surface area of the material, which is conducive to the formation of more corrosion protection layers, thus improving the corrosion resistance to a certain extent. At the same time, it can also reduce the stress concentration within the material, reducing the emergence of corrosion cracks and the expansion rate. Notably, the alloying elements away from the fusion zone were uniformly distributed and did not appear to be biased, which had a positive impact on the overall performance of the welded joint. Therefore, controlling the welding thermal cycle and temperature range is essential for preventing the precipitation of harmful phases and safeguarding the performance of welded joints.
3.4. Corrosion Resistance of the Welded Joints
Figure 10 shows the kinetic potential polarization curves of each sample in a 3.5 wt.% NaCl solution and their Tafel-zone-fitting data. The interval of ±60~120 mV from the open-circuit potential was regarded as the strong polarization zone, in which the corrosion current density and corrosion potential satisfy the Tafel relationship. The corresponding self-corrosion current (J
corr) and self-corrosion potential (E
corr) were obtained by using the Tafel extrapolation method for each sample, and the fitting data are shown in
Table 2. The over-passivation potential was represented as E
tp, i.e., the breakdown potential, which is the potential at which the corrosion current density increased rapidly with increasing anodic polarization potential. As shown in the dynamic potential polarization curves, the 304 stainless steel base material had no tendency to form a passivation film during electrochemical corrosion, whereas the other samples did. Unfortunately, the limited stable passivation zone of the curves suggested that other samples formed a passivation film during the process of electrochemical polarization but then quickly decomposed, and the film did not play a better role in corrosion resistance. From the fitted data in
Table 2, it was evident that the self-corrosion current density of the 304 stainless steel was the smallest, at only 5.78 × 10
−7 A·cm
−2, whereas the Q235 base material presented the largest self-corrosion current density of 4.63 × 10
−6 A·cm
−2, which demonstrates that the 304 stainless steel had the best corrosion resistance, whereas the Q235 base material had the poorest corrosion resistance. Among the welded samples, the self-corrosion current density of 3# was 1.04 times greater than that of 1# and 1.28 times greater than that of 2#, which illustrated that the corrosion resistance of all the samples was 304 stainless steel > 2# > 1# > 3# > #Q235 base material, in descending order from the strongest to the weakest. Although the stable passivation zone of the samples during electrochemical polarization was minimal, the over-passivation potential could still be considered a reference for the corrosion resistance of the samples in this experiment. The breakdown potential of 2# was the highest, at −0.544 V, and that of the Q235 was the lowest, at −0.650 V, which indicates that the corrosion resistance of 1# was the best, whereas that of the Q235 base material was the worst. From the perspective of breakdown potential, the corrosion resistance was 2# > 1# > 3# > Q235 base material, in descending order from the strongest to the weakest (304 stainless steel did not form a significant passivation film, so there was no breakdown potential value). The Cr content in a grain is crucial to its corrosion resistance, and in principle, the higher the Cr content is, the greater the degree of corrosion resistance it possesses [
21,
22]. As indicated by Equations (1) and (2), the Cr equivalents in the weld region of samples 1#, 2#, and 3# were 3.7, 8.2, and 3.57, respectively. This suggested that 2# exhibited superior corrosion resistance compared with 1# and 3#, followed by 1#, which demonstrated an intermediate level of corrosion resistance, and 3# presented the poorest resistance. These observations aligned with the data of the kinetic potential polarization curves.
To further investigate the corrosion resistance of the samples, an electrochemical impedance spectroscopy (EIS) test was conducted on each sample in a 3.5 wt.% NaCl solution. The data were fitted via ZSimpWin 3.60, and the results of the test and the corresponding equivalent circuit diagrams are presented in
Figure 11.
Figure 11a shows the Nyquist diagram, which allows the corrosion resistance of a material to be characterized by the magnitude of the capacitive loop radius, which is directly proportional to the aforementioned radius. It can be observed from the illustration that the 304 stainless steel sample corresponded to a considerably larger capacitive arc radius than the remaining samples did, but the Q235 base material exhibited the smallest one. Therefore, the corrosion resistance of the samples was 304 stainless steel > 2# > 1# > 3# > Q235 base material, in the order of strongest to weakest.
Figure 11b shows a representation of the Bode impedance modulus plot. In general, materials with higher impedance moduli at low frequencies demonstrated superior corrosion resistance. The impedance modulus of 304 stainless steel was markedly different from that of the other samples, whereas that of the Q235 base material was the lowest but exhibited a slight proximity to that of the welded samples. This further substantiated the assertion that 304 stainless steel exhibited the most exemplary corrosion resistance, whereas the Q235 base material presented the least.
Figure 11c shows the Bode phase angle plot. This demonstrated that with increasing frequency, the phase angle of the electrochemical polarization reaction exhibited a distinctive trend, initially increasing to a peak and then decreasing. This phenomenon was evident in all the samples, indicating that the passivation film produced on the sample surface during the polarization process was unstable, which is consistent with the results of the limited stable passivation zone exhibited by the kinetic potential polarization curves in
Figure 10. The peaks of the curves in the middle- and low-frequency regions were distributed between 64.4° and 81.7°, all of which were less than 90°. This indicated that although the electrochemical reaction interface was still far from the state of ideal capacitance, the system still exhibited high impedance. (The value of the system impedance was determined by the combination of the Faraday impedance and the non-Faraday impedance.) The single peak of the curves in
Figure 11c pointed to the existence of a single-time constant in the electrochemical reaction system. This result implied that the electrochemical reaction interface was characterized by the presence of only the charge-transfer process. Furthermore, the unstable passivation film, which corresponds to a brief interval of stable passivation on the kinetic potential polarization curves, exerted a negligible influence at the electrochemical reaction interface. Additionally, the electrochemical polarization process was not influenced by other processes, such as the adsorption of intermediates. Accordingly, the electrochemical impedance parameters could be fitted by the equivalent circuit illustrated in
Figure 11d.
The principal electrochemical component parameters derived from the equivalent circuit (EEC) fitting are presented in
Table 3. The fitting errors of all the electrochemical component parameter values listed in the table were less than 3%, and those of the majority of the parameters were controlled to within 1%. As shown in
Figure 11a, the Nyquist diagram demonstrated that the equivalent circuit was accurately represented. Furthermore, the chi-square (χ
2) analysis indicated a high degree of correlation between the fitted circuit and the experimental data. The aforementioned fitting parameters suggested that the selected equivalent circuit was reasonable and could be used to explain the electrochemical reaction process. R
s represents the solution resistance from the working electrode to the reference electrode, which corresponds to the starting position of the capacitive arc in
Figure 11a. The resistance to charge transfer (R
ct) is a measure of the difficulty of charge transfer across the electrode and electrolyte solution interface during the electrode reaction process. A smaller value of R
ct was indicative of more facile charge transfer and a higher corrosion rate. The data in
Table 3 indicate that the charge-transfer resistance of 304 stainless steel was significantly greater than that of the other samples, reaching 95690 Ω·cm
2. Conversely, the resistance of the Q235 base material was the lowest. The value of R
ct could be utilized to determine the order of corrosion resistance of the samples, with 304 stainless steel exhibiting the highest resistance, followed by the 2#, 1#, 3#, and Q235 base materials, in that order. The constant phase element (CPE) was employed to describe the double-layer capacitance (CPE
dl), with n
dl representing the fitting parameter of the double-layer capacitance (0 < n < 1). When the electrodynamic potential was perturbed in the electrode system, a portion of the current was utilized to charge the double-layer capacitor (CPE
dl), as previously described, and was designated the non-Faradaic current. The remaining portion of the current was employed directly for the electrode reaction, traversing the Faraday impedance (Z
F) and adhering to Faraday’s law, and was classified as the Faraday current. The term R
p denotes the polarization resistance, which can be defined as the sum of all the impedances involved in the electrochemical polarization (activated polarization) process. Importantly, this value and the diffusion resistance of the substance (Z
W) produced by the concentration polarization process together constitute the Faraday impedance (Z
F). When calculating the polarization resistance, it was essential to consider all the impedances present in the activated polarization process. As evidenced by the preceding analysis, the electrochemical polarization system in this study encompassed solely the charge-transfer process. Consequently, the polarization resistance R
p was identical to the charge-transfer resistance R
ct. In general, a higher value of polarization resistance indicates superior corrosion resistance. Therefore, the polarization resistance could be utilized to characterize the corrosion resistance of the sample, and the results were in accordance with those obtained from the kinetic potential polarization curve, i.e., the corrosion resistance of the sample was in descending order, beginning with 304 stainless steel, followed by 2#, 1#, and 3#, and conclusions were drawn with the Q235 base material.