The Role of La–Ti–Al–O Complex Inclusions in Microstructure Refinement and Toughness Enhancement of the Coarse-Grained Heat-Affected Zone in High-Heat-Input Welding

Abstract

The low-temperature impact properties of high-heat-input steels, particularly low-carbon Nb–Ti steel, are significantly influenced by the coarse-grained heat-affected zone (CGHAZ) in welded joints. The microstructure predominantly consists of granular bainitic ferrite (GBF), ferrite side plate (FSP), degenerate pearlite (DP), coarse plate-like ferrite (PF), and limited acicular ferrite (AF). This study investigates the effect of lanthanum (La) addition to Nb–Ti steel, leading to the formation of composite inclusions with a LaAlO₃·TiN core surrounded by MnS/MnC precipitates. Unlike conventional Al₂O₃·MnS inclusions in Nb–Ti steel, these La-modified inclusions promote enhanced AF nucleation. This not only refines prior austenite grains but also reduces detrimental microstructural constituents such as GBF and FSP. As a result, the impact energy at −40 °C significantly improves from 23 J (Nb–Ti steel) to 137 J (Nb–Ti–La steel). Moreover, the inclusions exhibit an increase in size but a decrease in number density. The Nb–Ti–La variant demonstrates a higher AF volume fraction and increased AF density within the CGHAZ. The refined grain structure, along with an increased proportion of high-angle grain boundaries, effectively impedes secondary crack propagation. These microstructural modifications contribute to a substantial improvement in the low-temperature impact toughness of welded joints.

1. Introduction

Conventional low-carbon Nb–Ti steels subjected to high-heat-input welding (E_j > 50 kJ/cm) often exhibit significant challenges, including abnormal austenite grain coarsening and excessive formation of detrimental microstructures such as granular bainitic ferrite (GBF) and ferrite side plate (FSP) in the coarse-grained heat-affected zone (CGHAZ). These microstructural changes lead to a drastic reduction in the weld joints’ low-temperature impact toughness [1,2,3,4,5]. In response, recent studies have focused on optimizing alloy composition to mitigate these adverse effects.

Non-metallic inclusions are inevitable by-products formed during steel production [6,7,8,9]. These inclusions exhibit size-dependent effects on mechanical properties: larger inclusions, owing to their substantial hardness, frequently act as stress concentrators that initiate secondary cracks, thereby significantly deteriorating impact toughness. In contrast, smaller inclusions exert minimal disruption to microstructural continuity while simultaneously serving as effective nucleation sites for high-angle grain boundary structures, including acicular ferrite (AF) and polygonal ferrite (PF). This dual functionality makes controlled dispersion of fine inclusions a critical factor in enhancing impact toughness.

Extensive research has been conducted from the perspective of composition control to determine the optimal approach for generating numerous finely dispersed non-metallic inclusions in structural steels [10,11,12,13,14,15]. Chai et al. [15] investigated the effect of Mg addition to Ti-killed steel, revealing that as Mg content increased from 0.0023% to 0.006%, the inclusions transformed from Ti₂O₃ + MgTiO₃ complex oxides to MgTiO₃ + MgO composite inclusions and single-phase MgO. Notably, the MgTiO₃ composite oxides were found to effectively promote acicular ferrite (AF) nucleation, simultaneously refining Ti-containing inclusions and improving the impact toughness of the coarse-grained heat-affected zone (CGHAZ). Yang et al. [12] examined the influence of Nb on inclusions in low-carbon Nb/Ti micro-alloyed steel. Their results demonstrated that while the prior austenite grain size remained unchanged at 0.045% Nb addition, the ability of micron-scale inclusions to induce AF nucleation was suppressed. Additionally, Nb exhibited a tendency to segregate at the interface between inclusions and the matrix. The study also observed an increase in nanoscale precipitates under Nb’s influence, leading to enhanced pinning effects and increased CGHAZ hardness, though the impact on toughness was not explicitly addressed. Zhao et al. [13] introduced Ce into low-alloy cast steel and conducted morphological analyses, which revealed significant spheroidization of elongated inclusions (reduced from 14.7 μm to 5.7 μm), thereby mitigating their detrimental effects on the steel matrix. The inclusion composition underwent the following transformation upon Ce addition: 5MnS + Al₂O₃ + 8Ce + 2O → Ce₂O₂S ⋅ Ce₂S₃ ⋅ MnS + CeAlO₃ ⋅ Ce₂S₃ ⋅ MnS + 4Mn. Furthermore, austenite grain growth was effectively suppressed, with the average grain size decreasing from 19 ± 11 μm to 15 ± 7 μm.

In recent years, an increasing number of studies have focused on rare earth (RE)-treated steels, providing new insights for the application of high-heat-input welding technologies [16,17,18,19]. Garrison, W.M. and Maloney, J.L [20] used HY180 steel with a similar composition to AF1410 steel as the test steel and added 0.25% Mn, 0.06% La, and 0.015% La to steel numbered 180-1 to 180-3. The mechanical properties of the three steels were tested, and the strength did not change significantly. In terms of impact toughness, the impact value of 180-3 steel was the highest (213 J), and fine La-containing inclusions were found in the dimples of the impact fracture. The impact value of 180-2 steel was the lowest (127 J), and a large and fragmented La-containing inclusion was found in the dimple. Ning et al. [21] studied the effect of Lanthanum on inclusions in Q355B weathering steel in the laboratory and refined four test steels with different La contents (0, 0.0075%, 0.0184%, 0.0425%). The composition of inclusions mainly experienced three stages of metallurgical reactions before finally forming: I. 3Al₂O₃ ⋅ MnS + 2La → 2 LaAlO₃ ⋅ Al₂O₃ ⋅ MnS + (4/3) Al, II. 2LaAlO₃ ⋅ Al₂O₃ ⋅ MnS + 2La → 2LaAlO₃ ⋅ La₂O₃ ⋅ MnS + Al₂O₃ + (2/3) Al, and III. 2LaAlO₃ ⋅ La₂O₃ ⋅ MnS + MnS + 2S → 2La₂O₂S⋅La₂O₃ + 2Mn + Al₂O₃. When La = 0, 0.0075, and 0.0184%, the sizes of inclusions were 5.35 μm, 3.4 μm, and 2.48 μm, respectively, indicating that the inclusions were clearly refined when La content increased from 0 to 0.0184%. Peng et al. [22] studied the effects of adding La alone and composite Al–La on inclusions and microstructure in sulfur-containing pericrystalline steel. When La was not added (steel B), the inclusions in steel were mainly MnS. After adding La alone (steel L), the main inclusions changed to La₂O₂S + LaAlO₃–La₂O₂S, while after composite Al–La addition (steel AL), the main inclusions changed to LaAlO₃ + La₂S₃. The number density of inclusions decreased from 308 mm² (steel B) to 202 mm² (steel L) and 210 mm² (steel AL), and the size of inclusions decreased from 3 μm to 2.8 μm and 2.4 μm, indicating that both the addition of La alone and the composite addition of Al–La can achieve the effect of purifying the steel matrix and refining inclusions.

Rare earth elements cerium (Ce) and lanthanum (La) both belong to the light rare earth group, sharing similar chemical properties [23]. Compared with other rare earth elements in the same group, they offer distinct advantages in terms of resource abundance and cost-effectiveness [24], with La being particularly advantageous in this regard. However, research focusing specifically on the application of La in welding technology remains scarce in the published literature. In particular, studies investigating how La addition influences the microstructure and inclusions in the coarse-grained heat-affected zone (CGHAZ), and consequently how it affects impact toughness, have yet to establish a comprehensive theoretical framework. Considerable room for progress still exists in this research domain within the domestic academic community.

In this paper, low-carbon Nb–Ti steel was used as the base steel, and a trace amount of the rare earth element La was added to the Nb–Ti steel. A welding thermal cycle with a welding heat input of 100 kJ/cm was simulated on a Gleeble-3800. A series of characterizations were conducted on the microstructure and inclusions in the CGHAZ, explaining the influence of the rare earth element La on the welding performance under high heat input from a microscopic perspective. An oscillating impact test at −40 °C was conducted to characterize the fracture morphology and secondary cracks, further elucidating the effect of the rare earth element La on the impact fracture behavior of the CGHAZ. In addition, TEM samples were prepared using SEM-FIB for inclusions containing La, allowing for a deeper analysis of the composition distribution within the inclusions. The relationship between inclusions containing La and the ferrite matrix was also investigated, further explaining the promoting effect of inclusions containing La on AF nucleation and related mechanisms.

2. Materials and Methods

2.1. Smelting and Composition of Experimental Steel

In a 100 kg vacuum induction furnace (Shenyang Hengrun Vacuum Technology Co., Ltd, Shenyang, China), two steel ingots (Nb–Ti/Nb–Ti–La) weighing 80 kg each were smelted. Through controlled rolling and controlled cooling processes, steel plates with a thickness of 18 mm were formed. During the alloying treatment, the oxygen (O) content in the molten steel was controlled by adding carbon (C) (O ≤ 0.003%); after alloying treatment, pure lanthanum (La) was directly added to the molten steel.

The chemical compositions measured in the laboratory are shown in

Table 1. Composition of test steel (wt.%).

Elements	C	Si	Mn	P	S	Nb	Ti	La	Al	O	N
Nb–Ti steel	0.065	0.21	1.55	0.006	0.002	0.026	0.012	0	0.020	0.0025	0.0028
Nb–Ti–La steel	0.068	0.19	1.57	0.005	0.002	0.024	0.010	0.0135	0.023	0.0029	0.0027

, where the La content is 0 and 0.0135%.

2.2. Welding Thermal Cycle Test

The following welding thermal cycle process was simulated on the Gleeble-3800 thermal simulation test machine (Dynamic Systems Inc., Albany, NY, USA) (Figure 1b): rapidly heating up to the peak temperature (1350 °C) at a heating rate of 100 °C/s, with a holding time of 1 s; then cooling down to 200 °C at a certain cooling rate, with a measured t_8/5 = 30 s and a corresponding heat input of 100 kJ/cm. The welding thermal cycle curve is shown in Figure 1b.

At the beginning of the experiment, 5 samples need to be taken from Nb–Ti steel/Nb–Ti–La steel plates: 4 square bars (3 of which can be used for subsequent oscilloscope impact tests and 1 can be used for characterization tests) and 1 round bar (to facilitate the installation of a thermal expansion meter during welding thermal simulation tests to measure the phase transition point). The detailed dimensions and sampling locations are shown in Figure 1a.

After completing the welding thermal simulation experiment, the specimens need to be reprocessed into Charpy pendulum impact specimens with dimensions of 10 × 10 × 55 mm³. The welding point connected to the thermocouple wire is located directly above the V-notch, as shown in Figure 1d. The low-temperature impact toughness of the simulated welding joint was tested on the NI300C Instrumented Impact Testing Machine, NCS Testing Technology CO., LTD., Beijing, China. The obtained specimens (including the fracture surface, e1) can be used for secondary crack EBSD analysis and main fracture SEM; small pieces with thicknesses of 0.5 mm (e2) and 3 mm (e3) were cut from the specimens that were not reprocessed and were used for TEM experiments and OM/SEM experiments, respectively.

Firstly, the OM sample that was not treated in the corrosive solution (obtained by cutting with molybdenum wire, grinding with sandpaper, and polishing with a polishing machine) was observed using an Axiover-200MAT optical electron microscope (ZEISS, Oberkochen, Germany) for statistical analysis of the number and size of inclusions (Image Pro Plus 6.0, MEDIA, Rockville, MD, USA). Then, the OM sample was corroded with 4% nitric acid alcohol, and the obtained sample was used for OM and SEM (SU5000, HITACHI, Tokyo, Japan) microstructure characterization. After the microstructure characterization experiment was completed, the characterization of inclusions was carried out, including morphology observation (SEM) and composition analysis (EDS-SE mode: convenient for observing the AF around inclusions). Finally, TEM experiments were conducted with the assistance of a JEM-F200 transmission electron microscope (HITACHI, Tokyo, Japan, accelerating voltage 200 kV, modes including TEM, STEM, SAED). The preparation of tissue characterization samples was carried out using an electrolytic double-spray method (corrosive solution: 7% perchlorate ethanol solution, voltage/temperature = 20 V/20 °C). In addition, electron backscatter diffraction (EBSD, ZEISS, Oberkochen, Germany) samples were prepared by electrolytic polishing in a 10% methanol perchlorate solution, and the crystal orientation and microcracks in the simulated Coarse grain heat affected zone (CGHAZ) were characterized by EBSD (SU5000, HITACHI, Tokyo, Japan) with a step size of 0.3 μm. The inclusion TEM characterization sample was prepared using SEM-FIB, and the deposited material was Pt with a thickness of about 1 μm. The ion beam voltage was 30 kV, and the obtained sample thickness was less than 100 nm.

3. Results

3.1. Inclusions

3.1.1. OM Observation and Statistical Analysis of Inclusions

The number and distribution of inclusions in the Nb–Ti steel and Nb–Ti–La steel CGHAZ were observed under an optical electron microscope with a magnification of 200×. A total of 30 fields of view were selected for each and IPP was used to statistically analyze the inclusions (Figure 2). The results show that the average size of inclusions in Nb–Ti steel was 1.76 μm, with inclusions with a size ≤ 2 μm accounting for the highest proportion (65.4%), inclusions with a size of 2–3 μm accounting for 18.1%, inclusions with a size of 3–4 μm accounting for F7.9%, and inclusions with a size > 4 μm accounting for 2.1%. No inclusions with a size > 6 μm were found in the field of view. The average size of inclusions in Nb–Ti–La steel is 2.63 μm, with a decrease in the proportion of inclusions between 1 and 2 μm (48.7%), a slight increase in the proportion of inclusions between 2 and 3 μm (26.1%), a decrease in the proportion of inclusions between 3 and 4 μm (15.2%), a decrease in the proportion of inclusions between 4 and 6 μm (8.2%), and a decrease in the proportion of inclusions larger than 6 μm (1.8%).

3.1.2. SEM Observation and EDX Analysis of Inclusions

As shown in Figure 3 and Figure 4, the morphology and composition of inclusions in the CGHAZ microstructure of Nb–Ti/Nb–Ti–La steel were observed and analyzed. The results show that the main component of inclusions in Nb–Ti steel is Al₂O₃ · MnS, which is located at the grain boundary and the presence of AF can be directly observed around the inclusions. The composition of inclusions in Nb–Ti–La steel is La–Ti–Al–O · MnS, and typical ferrite structures can usually be found around these inclusions. Therefore, this type of inclusion has a significant promoting effect on ferrite nucleation. A simple schematic diagram was drawn based on the distribution of components in the inclusions, and it can be seen that La–Ti–Al–O is the main body that forms inclusions, while MnS is attached to the outside of the inclusions.

3.2. Microstructures

Figure 5 demonstrates the OM and SEM microstructures of CGHAZs in Nb–Ti and Nb–Ti–La steels. The microstructure composition of the CGHAZs of the two types of steel was statistically analyzed (Figure 6), and it can be seen that both types of steel CGHAZ contain granular bainitic ferrite (GBF), acicular ferrite (AF), polygonal ferrite (PF), martensitic/austenitic components (M/A), and degenerate pearlite (DP). The difference is that the proportion of GBF in the Nb–Ti–La steel CGHAZ is significantly smaller than that in the Nb–Ti steel CGHAZ, while the content of AF and PF has significantly increased. In addition, M/A and DP are not the main microstructure morphology, being slightly reduced. Figure 7 reflects the microstructure observed by TEM in the CGHAZ region of Nb–Ti/Nb–Ti–La steel. The microstructures of both steels are composed of a large number of plate structures, including ferrite flat noodles and black linear dislocation tangles. In addition, DP with a black short linear and layered structure was also found.

Figure 8 displays EBSD maps of the simulated CGHAZs for Nb–Ti and Nb–Ti–La steels, featuring inverse pole figures (IPFs), image quality (IQ) maps with grain boundary distributions, and kernel average misorientation (KAM) maps. As demonstrated in Figure 8g, the mean equivalent diameter (MED) of the Nb–Ti–La steel CGHAZ is significantly smaller than that of the Nb–Ti steel CGHAZ. At a 15° misorientation angle (MTA) threshold, La addition reduced the MED from 8.48 μm to 6.82 μm, indicating grain refinement. As illustrated in Figure 8b,e, distributions of low-angle grain boundary (LAGB, MTA: 2~15°) and high-angle grain boundary (HAGB, MTA > 15°) are demonstrated. Figure 8h provides grain boundary statistics as a function of misorientation angle, revealing a higher HAGB but lower LAGB fraction in the Nb–Ti–La steel CGHAZ compared to the Nb–Ti steel CGHAZ. The KAM value reflected the microstrain level in the alloy materials, with the average KAM value of the Nb–Ti–La steel CGHAZ (0.79°) being measurably higher than that of the Nb–Ti steel CGHAZ (0.65°).

3.3. Impact Properties

Table 2. Statistical results of low-temperature impact test.

	Total Impact Absorbed Energy (Et/J)	Crack Initiation Energy (Ei/J)	Crack Propagation Energy (Ep/J)
Nb–Ti CGHAZ	23 ± 5	19 ± 3	4 ± 2
Nb–Ti–La CGHAZ	137 ± 15	37 ± 9	100 ± 7

presents the instrumented impact test results for the simulated CGHAZ samples at −20 °C, with typical deflection-load curves shown in Figure 9. The incorporation of La increased the total impact absorption energy from 23 J to 137 J. Among them, the crack initiation energy increased from 19 J to 37 J, while the crack propagation function significantly increased from 4 J to 100 J. Therefore, the addition of La improved the low-temperature toughness of CGHAZ samples at 100 kJ/cm.

Figure 10 presents typical impact fracture surface morphologies of the CGHAZ samples. As shown in Figure 10a,e, the conventional impact fracture morphology consists of three parts (from right to left as shown in the figure): I V-shaped notch; II Fiber zone; III Radiation zone. The fibrous zones in the Nb–Ti and Nb–Ti–La steel CGHAZ display equiaxial dimples (Figure 10b,e). Furthermore, inclusions of considerable size are observed at the base of the equiaxial dimples (Figure 10g). The radical zones of both CGHAZs exhibit the river pattern, with tear ridges present in the Nb–Ti–La steel CGHAZ. Additionally, the cleavage facet area of the Nb–Ti–La steel CGHAZ is considerably smaller than that of the Nb–Ti steel CGHAZ. As shown in Figure 10d,h, the inclusions exerted negligible influence on river pattern propagation in radical zones. The impact fracture surface morphologies of the CGHAZ samples demonstrate high consistency with the impact absorbed energy results.

4. Discussion

4.1. A Study on the Influence of Rare Earth La on the Microstructure of CGHAZs

In order to detect the phase transformation of CGHAZ microstructure during simulated high-heat-input welding, a thermal expansion instrument was installed on a round rod specimen with a diameter of φ6 × 80 mm², as described in Section 2.2. After processing by Origin, the curve obtained reflects the relationship between temperature and expansion, as illustrated in Figure 11. The phase transformation start temperature (Ac3) increased from 702 °C to 758 °C, while the finish temperature (Ac1) decreased from 508 °C to 490 °C. The expansion of the γ → α transformation temperature range, coupled with an enhanced degree of phase transformation, resulted in a significant increase in ferrite content within the CGHAZ, particularly in the amount of AF (this aspect will be further discussed in Section 4.2). Consequently, the originally coarse austenite grains in the CGHAZ were effectively refined into smaller grains. Furthermore, the increased ferrite content also led to a reduction in the GBF. It can be confirmed that the thermal expansion curves of two different samples explain the significant influence of the rare earth element La on the phase transition of CGHAZ microstructure [25,26,27].

4.2. A Study on the Composition Distribution of La–Ti–Al–O Composite Inclusions and Their Effect on Acicular Ferrite Nucleation

As shown in Figure 12, the TEM-EDS analysis results of the La–Ti–Al–O composite inclusion FIB sample are presented. There is a small amount of MnS around the inclusions, and the core of the inclusions is mainly composed of La–Ti–Al–O composite inclusions.

To understand the nucleation effect of LaTi₂Al₉O₁₉ inclusions on AF, the LaTi₂Al₉O₁₉ inclusion and induced AF nucleation were observed by means of FIB-TEM, as illustrated in Figure 13. The bright-field image (Figure 13a) and high-angle annular dark-field image (Figure 13b) show the detailed TEM morphology of the LaTi₂Al₉O₁₉ inclusion and its induced AF. As shown in Figure 13d, SAED calibration revealed a crystallographic parallel relationship of ${\left(40\stackrel{\mathrm{-}}{1}\right)}_{\mathrm{La}{\mathrm{Ti}}_2{\mathrm{Al}}_9{\mathrm{O}}_{19}}\parallel {\left(110\right)}_{\mathsf{\alpha }–\mathrm{Fe}}$.

Bramfitt [28] developed the two-dimensional lattice mismatch (δ) theory for heterogeneous nucleation, which can be formalized as follows:

$${\mathsf{\delta }}_{{\left(\mathrm{hkl}\right)}_{\mathrm{n}}}^{{(\mathrm{hkl})}_{\mathrm{s}}} = \sum _{\mathrm{i} = 1}^3{\displaystyle \frac{\frac{\left|\left({\mathrm{d}\left[\mathrm{uvw}\right]}_{\mathrm{s}}^{\mathrm{i}}\cos \mathsf{\theta }\right)-{\mathrm{d}\left[\mathrm{uvw}\right]}_{\mathrm{n}}^{\mathrm{i}}\right|}{{\mathrm{d}\left[\mathrm{uvw}\right]}_{\mathrm{n}}^{\mathrm{i}}}}{3}} \times 100\%$$

where (hkl)s and (hkl)n represent low-index planes in the matrix and nucleated phase, [uvw]s and [uvw]n denote low-index directions on these planes, and d[uvw]s/d[uvw]n represent atomic spacings along the respective directions. The angle between [uvw]s and [uvw]n is denoted by θ. In general, when the two-dimensional lattice mismatch (δ) exceeds 12%, the nucleation effect can be essentially negligible; conversely, when δ is less than 12%, the heterogeneous phase nucleates evidently [29,30,31,32]. Figure 14 demonstrates both crystal structures and two-dimensional lattice mismatch calculations. For the crystallographic parallel relationship of ${\left(40\stackrel{\mathrm{-}}{1}\right)}_{\mathrm{La}{\mathrm{Ti}}_2{\mathrm{Al}}_9{\mathrm{O}}_{19}}\parallel {\left(110\right)}_{\mathsf{\alpha }–\mathrm{Fe}}$, δ = 1.68% indicates that the LaTi₂Al₉O₁₉ inclusion effectively promoted AF nucleation.

4.3. A Study on the Influence of La on Impact Toughness

Figure 15 exhibits the SEM morphologies of microvoids and microcracks near impact fracture surfaces in Nb–Ti and Nb–Ti–La steel CGHAZs. Inclusion-matrix debonding under transient impact produces large microvoids, while θ-particle separation from DPF generates smaller microvoids. Since Nb–Ti–La steel CGHAZ inclusions exceeded Nb–Ti steel CGHAZ inclusions in average size (Figure 2), they formed larger microvoids.

Figure 15a–c,f show microcracks deflecting or stopping at grain boundaries. Nb–Ti steel CGHAZ microcracks exhibited longer propagation distances than Nb–Ti–La steel CGHAZ microcracks. To elucidate grain boundary types, microcracks were characterized using EBSD, as shown in Figure 16. Microcracks deflected or arrested at HAGBs, but propagated straight through LAGBs. HAGBs significantly hindered microcrack propagation, while LAGBs exhibited no effect. Since AF grain boundaries were HAGBs, the HAGB proportion in the Nb–Ti–La steel CGHAZ exceeded that in the Nb–Ti steel CGHAZ (Figure 8h). Moreover, the MED of the Nb–Ti–La steel CGHAZ was lower than that of the Nb–Ti steel CGHAZ (Figure 8g). Consequently, the hindering effect on microcrack extension was more pronounced in the Nb–Ti–La steel CGHAZ. The incorporation of La into the CGHAZ sample substantially increased crack propagation energy from 4 J to 100 J. Consequently, the incorporation of trace La into the Nb–Ti steel CGHAZ led to an enhancement in its low-temperature impact toughness.

5. Conclusions

(1)

The 0.0135 wt.% La created LaTi₂Al₉O₁₉ · MnS complex inclusions in the Nb–Ti–La steel CGHAZ, contrasting with Al₂O₃ · MnS complex inclusions in the La-free Nb–Ti steel CGHAZ, and the 1.68% two-dimensional lattice mismatch for ${\left(40\stackrel{\mathrm{-}}{1}\right)}_{\mathrm{La}{\mathrm{Ti}}_2{\mathrm{Al}}_9{\mathrm{O}}_{19}}\parallel {\left(110\right)}_{\mathsf{\alpha }–\mathrm{Fe}}$ enabled effective AF plate nucleation.

(2)

La, as the main component of LaTi₂Al₉O₁₉ · MnS core inclusions, has a strong adsorption capacity for small-sized non-metallic inclusions in steel, thereby increasing the average size of inclusions (increased from 1.76 μm to 2.63 μm). The precipitation position of La-containing composite inclusions gradually approached the interior of the crystal, creating conditions for AF/PF nucleation while significantly increasing the proportion of AF/PF in the microstructure of the CGHAZ (AF: from 5.3% to 29.8%; PF: from 11.9% to 32.1%).

(3)

La addition increased the AF fraction and HAGB proportion and reduced the MED, enhancing crack propagation energy (4 J to 100 J). The 0.0135 wt.% La addition thus enhanced CGHAZ low-temperature toughness for 100 kJ/cm high-heat-input welding.