Effect of Heat Input on Microstructure and Tensile Properties in Simulated CGHAZ of a V-Ti-N Microalloyed Weathering Steel

Abstract

The mechanical properties of a coarse-grained heat-affected zone (CGHAZ) are affected by welding thermal cycling with varied heat input (E_j), but its effect on tensile properties is rarely studied. In the present work, E_j = 15, 35, 55, 75 kJ/cm CGHAZ samples were prepared via Gleeble^TM (St. Paul, MN, USA) for a novel V-Ti-N microalloyed weathering steel. The tensile properties of CGHAZ with varied E_j were evaluated. The results indicated that mixed microstructures dominated by lath bainitic ferrite (LBF) and granular bainitic ferrite (GBF) were obtained at E_j = 15 and 35 kJ/cm, respectively, while a mixed microstructure composed of GBF, intragranular acicular ferrite (IGAF), and polygon ferrite (PF) formed at E_j = 55 and 75 kJ/cm, apart from martensite/austenite (M/A) constituents in each E_j condition. The above variation tendency in the microstructure with the increase in E_j led to coarsening of low-angle grain boundaries (LAGBs) and a decrease in dislocation density, which in turn resulted in a yield strength (YS) decrease from 480 MPa to 416 MPa. The mean equivalent diameter (MED), defined by the misorientation tolerance angles (MTAs) ranging from 2–6°, had the strongest contribution to YS due to their higher fitting coefficient of the Hall–Petch relationship. In addition, the increase in the average size (d_M/A) of M/A constituents from 0.98 μm to 1.81 μm and in their area fraction (f_M/A) from 3.11% to 4.42% enhanced the strain-hardening stress. The yield strength ratio (YR) reduced as the E_j increased, and the lower density and more uniform dislocations inside the ferrite led to a uniform elongation (uE) increase from 9.5% to 18.6%.

1. Introduction

Weathering steel (WS) is, essentially, an atmospheric-corrosion-resistant steel, usually containing no more than 0.20 wt % C, 0.25–0.55 wt % Si, and 0.8–1.5 wt % Mn, as well as one or more other alloy elements, such as Cu, Cr, Ni, with a total alloy content of 1.00–5.00 wt % [1,2,3]. WS normally possesses a stable rust layer to prevent erosion, having many advantages over laborious, costly, and high-emission coatings, and it has been increasingly used for important applications, such as uncoated bridges, buildings, and transmission towers, through the traditional arc welding process [4,5,6]. However, during the welding of WS, due to the influence of thermal cycling, the strength and toughness reduction in CGHAZ near the fusion line cannot be ignored [7,8]. Since weathering steel contains more Si [9], Cr [4,10], and other alloy elements, after subjection to the welding thermal cycle with a peak temperature higher than 1300 °C, hard brittle phase GBF and M/A constituents are enriched in CGHAZ [11,12]. When the E_j increases, the above-mentioned hard and brittle phases will coarsen and increase in quantity [13,14], further increase the reduction in toughness and strength.

Extensive investigations [15,16] have reported that the nucleation of heteronucleated ferrites on inclusions or precipitated particles can reduce the content of GBF, ferrite side plate, etc., refining the CGHAZ microstructure and improving its mechanical properties. V-Ti-N microalloyed steels with high ferritic heterogeneous nucleation capacity have received extensive attention. These utilize Ti (C, N) particles with high thermal stability to pin prior austenite grain boundaries (PAGBs) and refine them. VN and/or V (C, N) particles from the solid γ-phase reduce the mismatch with ferrite and promote the heterogeneous nucleation of ferrites [17,18]. These studies provide good microalloyed methods to improve the CGHAZ mechanical properties of weathering steel. Zhang and Hu et al. [19,20] studied the effect of E_j on the CGHAZ microstructure and impact properties of V-Ti-N microalloyed C-Mn steel; the results showed that with the increase in E_j, the ability of V-rich particles to promote heterogeneous nucleation of ferrite was enhanced, and the corresponding impact energy increased. Shi et al. [21] also obtained similar research conclusions: as the time of T_8/5 is increased from 45 s to 285 s, the CGHAZ toughness of V-Ti-N microalloyed steel significantly improved; however, the YS of CGHAZ decreased from 385 MPa to 335 MPa due to the formation of a large amount of intragranular polygon ferrite (IGPF) with the extension of T_8/5. Moreover, Wang et al. [22] investigated the CGHAZ tensile properties of V-Ti-N microalloyed steel containing 0.28 wt % Mo under varied E_j, during the process of increasing E_j from 35 kJ/cm to 65 kJ/cm and 120 kJ/cm; the YS of CGHAZ showed a trend of first increasing from 545 MPa to 610 MPa, and then decreasing to 394 MPa, respectively, the main reason being that the CGHAZ main microstructure transformed from GBF to IGAF and IGPF, respectively.

Considering that the influences of heat inputs on CGHAZ microstructure and the mechanical properties of V-Ti-N microalloyed steels with varied compositions are different, for the V-Ti-N microalloyed weathering steel, there may still be complex connections between the heat inputs and CGHAZ microstructure and mechanical properties. The effect of heat inputs on their CGHAZ impact property was reported by the author [23], because the change in CGHAZ strength can reflect its tendency towards softening and hardening, while also affecting the fatigue performance of the joint [24]. It is necessary to further evaluate the strength of CGHAZ with different E_j for V-Ti-N microalloyed weathering steel, and to the authors’ knowledge, research on heat inputs and CGHAZ strength has been scarcely reported previously.

In addition, during the continuous cooling process of the welding thermal cycle, bainitic ferrite, IGAF, IGPF, and other mixed microstructures are produced in the CGHAZ of V-Ti-N microalloyed weathering steel. It is difficult to establish relationship between complex microstructure and tensile properties [25,26]. In recent years, electron backscatter diffraction (EBSD) technology has been developed to analyze the influence of complex microstructure on tensile properties based on the variation of MTAs of different grain boundaries [25,27]. Fan et al. [28] showed that grain boundaries with MTAs ranging from 2–6° are the main boundaries controlling the tensile properties for low-alloy steel, while Zhu et al. [29] pointed out that the MED defined by MTAs smaller than 10° effectively governed the tensile properties. Olasolo et al. [30] indicated that MED defined by LAGBs less than 4° strongly contributes to tensile properties. Further, Li et al. [31] found the closest Hall–Petch relationship between the YS and the ferrite grains defined by LAGBs ranging from 2–12° in C-Mn steel. Hence, complex CGHAZ microstructure and corresponding tensile properties need further study. Moreover, the evolution of E_j will lead to differences in M/A constituents, dislocation structure, and density in CGHAZ microstructure, and its influences on strain-hardening behavior and elongation need to be elaborated.

In the present work, a novel V-Ti-N microalloyed weathering steel was designed and prepared. The steel was subjected to a CGHAZ welding thermal cycle with E_j = 15, 35, 55, 75 kJ/cm on a Gleeble-3500, and then the CGHAZ microstructure under each heat input condition was characterized and its tensile properties were tested. Moreover, the CGHAZ microstructure was further analyzed and quantitatively characterized by BESD and TEM, the relationship between the complex CGHAZ microstructure and tensile properties was evaluated, and the most effective structural unit for controlling YS was determined. In addition, the influence of CGHAZ microstructure on strain-hardening ability and elongation under different heat input conditions had also been studied. The research in this article will further clarify the influence of heat input on the microstructure and tensile properties of the welding heat-affected zone and understand the tensile fracture behavior with multiphase microstructure.

2. Materials and Experimental Method

The steel used in this article is a novel V-Ti-N microalloyed weathering steel. The specific chemical composition is shown in

Table 1. Chemical composition of experimental steel (wt. %) [23].

C	Si	Mn	P	S	Cr	Cu	Ni	Mo	Ti	V	N	Fe
0.046	0.19	1.30	0.009	0.005	0.45	0.32	0.37	0.05	0.015	0.024	0.0073	Bal.

, and the mechanical properties of the test steel are shown in

Table 2. Mechanical properties of experimental steel [23].

YS ReL/MPa	TS Rm/MPa	YR ReL/Rm	Elongation A/%	Energy Absorbed at −40 °C AkV/J	Micro-Hardness /HV10
489	594	0.82	26.5	238.4	226.0

; this indicates that this novel weathering steel has the advantages of high strength, high toughness, and low yield ratio.

Round bars with sizes φ15 × 80 (rolling direction) mm² were machined from the test steel. The simulated CGHAZ with E_j = 15, 35, 55, and 75 kJ/cm were conducted in a Gleeble-3500 system (Figure 1a). The samples were heated to 1320 °C at 100 °C/s and cooled based on the thermal cycle curve developed via the Rykalin-2D mode; the model formula is shown in Equation (1), the T_8/5 time corresponding to E_j = 15, 35, 55, and 75 kJ/cm is 18.9 s, 50.6 s, 112.3 s, and 232.5 s, respectively, and the thermal cycle curves are shown in Figure 1b. Taking samples from a 2 mm area on both sides of the thermocouple of the round bars, the CGHAZ microstructure and tensile properties of each E_j sample was detected, as shown in Figure 1c. Tensile properties are tested on an MTS servo hydraulic testing machine at room temperature with a strain of 0.25 mm/min, and the specimens obtained for each E_j were tested three times. The YS was determined by the 0.2% offset flow stress. In addition, a Vickers hardness tester (HV-10Z) was used to test the hardness of different heat input samples, and the average value was measured 5 times under each condition.

The microstructure was characterized via optical microscopy (OM; Axiover-200MA T, ZEISS, Jena, Germany), scanning electron microscopy (SEM; Hitachi-SU5000, Hitachi, Tokyo, Japan), and transmission electron microscopy (TEM; Talos-F200X, FEI, Hillsboro, USA) for each E_j sample. OM and SEM specimens were polished and etched with 4% nitrate alcohol. TEM samples for observing precipitated particles at different E_j were prepared by carbon extraction replication. The film samples of TEM were obtained by sanding to 30 μm and subsequently electrolytic polishing, and the polishing occurred in a solution of 5% perchloric acid in ethyl alcohol with v = 25 V, i = 55–65 mA, and T = 25 °C. The OM samples were re-polished and then etched with LePera’s reagent. The M/A constituents were corroded into bright white and subsequently observed by OM, then Image-Pro Plus (IPP) software (Media Cybernetics, Austin, USA) was used to evaluate the size and area fraction of M/A constituents for each E_j through more than 1000 M/A constituents in 15 fields of view. Samples from each E_j were ground and then electrolytically polished in 90% methanol solution of perchlorate with v = 18 V, I = 1 A, and t = 30 s, and evaluated by SEM equipped with a TSL electron backscatter diffraction (EBSD); the scanning step was 0.20 μm, and the MED of grains at different MTAs was determined by EBSD. In addition, an XRD (Rigaku D/max-2500 PC, Rigaku, Tokyo, Japan) with Cu K-α radiation was employed to determine the dislocation density for each E_j sample, and the specimens were scanned at a step width of 0.02° and counting time of 2 s over the range of 40°–105°, with three tests per sample.

$$E_j=\sqrt{\frac{4\mathsf{\pi }kc\rho \Delta \mathrm{t}}{1/{\left(T_2 -T_0\right)}^2 - 1/{\left(T_1 - T_0\right)}^2}}\times \mathsf{\delta }$$

(1)

where E_j is heat input; c is the specific heat capacity, 0.586 J/(g·°C); ρ is the density of iron, 7.800 g/cm³; k is the heat conductivity coefficient, 0.360 J/(cm·s·°C); δ is the thickness of plating, 0.24 cm; T₁ and T₂ are the defined cooling time temperatures T₁ = 800 °C, T₂ = 500 °C, respectively; T₀ is the preheating temperature, 25 °C; Δ_t is temperature cooling time from T₁ to T₂.

3. Experimental Result

3.1. Microstructural Characteristics

OM observations of simulated CGHAZ samples under different E_j are displayed in Figure 2; there are obvious differences in the morphology of the microstructures. When the E_j = 15 and 35 kJ/cm, the prior austenite grain boundaries (PAGBs) can be observed, and the PAGBs may be further divided by lath bainitic ferrite (LBF) and granular bainitic ferrite (GBF). However, the proportion of LBF is higher at E_j = 15 kJ/cm, while a microstructure of GBF is dominant at E_j = 35 kJ/cm. As the E_j further increased to 55 and 75 kJ/cm with two large heat inputs, PAGBs were almost covered by growth of ferrite and the CGHAZ microstructures were mainly intragranular acicular ferrite (IGAF) and polygon ferrite (PF). These types of microstructures occupy a higher proportion at E_j = 75 kJ/cm. In addition, developed GBF can also be found in CGHAZ microstructure of two large heat inputs.

Moreover, Figure 3 characterizes the variation in M/A constituents of simulated CGHAZ under different E_j; the M/A constituents appear bright white after being corroded by LePera’s reagent. With the increase in E_j, the macroscopic morphology of M/A constituents gradually transforms from elongated to blocky. Meanwhile, according to the statistics of the IPP software, the average size (d_M/A) and area fraction (f_M/A) of M/A constituents increase with the increase in E_j. Specifically, d_M/A increases from 0.98 μm to 1.13 μm, 1.68 μm, and 1.81 μm, and f_M/A increases from 3.11% to 3.56%, 4.21%, and 4.42%, respectively.

The effect of E_j on CGHAZ microstructure is further characterized by TEM. With the increase in E_j from 15 kJ/cm to 75 kJ/cm, the ferrite morphology changes from parallel lath bundles to blocky, and the size of ferrite coarsens significantly, with the results shown in Figure 4a–d. The black island structures can be found in each E_j sample, and their size increases with the increase in E_j. The island structure is identified as M/A constituent by light and dark fields and selected area electron diffraction (SAED) pattern (Figure 4e–e2). Moreover, the dislocation structures inside the ferrite are also characterized; when the E_j = 15 and 35 kJ/cm, there are high-density dislocation structures in the ferrite. In the partial enlarged view of Figure 4a,b, dislocation tangles (DTs) occur around the elongated M/A constituents (Figure 4a1,b1). With a further increase in E_j to 55 kJ/cm, dislocation density decreases significantly compared to E_j = 15 and 35 kJ/cm, but compared to the relatively uniformly distributed dislocation lines (UDDLs) at E_j = 75 kJ/cm (Figure 4d1), the internal dislocation lines (DLs) are densely distributed (Figure 4c1).

In addition, XRD patterns of samples that are subjected to different E_j are displayed in Figure 5a. The dislocation density can be calculated by Equation (2) according to Ref. [32]; the relevant calculation results are shown in Figure 5b. With the increase in E_j from 15 kJ/cm to 75 kJ/cm, the dislocation density decreases from 2.20 × 10¹⁴ m⁻² to 0.73 × 10¹⁴ m⁻².

$$\rho =k{\epsilon }^2/{Fb}^2$$

(2)

where ρ is the dislocation density, k is the Burgers vector of dislocations, ɛ is microstrain, (b) along <111> for body-centered cubic metals; the value of F is assumed to be 1.

Moreover, TEM characterizations of CGHAZ carbon extraction replica samples are shown in Figure 6. When the E_j = 15 and 35 kJ/cm (Figure 6a,b), the precipitated particles are mainly square particles. When the E_j increases to 55 and 75 kJ/cm (Figure 6c,d), in addition to square particles, there are also round particles in the microstructure, and the density and size of precipitated particles is enhanced with the increase in E_j. Further, the typical precipitated particles under each E_j condition are analyzed by energy dispersive spectroscopy (EDS), as shown in Figure 6I–IV. The EDS results indicate that the square particles are Ti (C, N) particles at low E_j = 15 and 35 kJ/cm (Figure 6I,II). However, when the E_j increases to 55 and 75 kJ/cm, based on the results of EDS analysis, the circular precipitated particles appearing in the sample are V (C, N) particles (Figure 6III), and the chemical composition of the square particles transforms into V-rich (Ti, V) (C, N) particles (Figure 6IV); this is also consistent with the research results of Zhang and Hu et al. [19,33]. In addition, 20 square particles and round particles under each heat input condition are selected respectively, and their EDS data were further processed and the atomic fraction of elements were statistically analyzed; the results are shown in

Table 3. The atomic fraction of each chemical element in particles (wt %).

Ej kJ/cm	Type of Element in Square Particles					Type of Element in Circular Particles
	Ti	V	C	N	Fe	V	C	N	Fe
15	23.15%	0%	41.81%	24.15%	10.89%	/	/	/	/
35	23.19%	0.05%	42.43%	24.89%	10.44%	0.02%	60.15%	5.89%	33.94%
55	24.55%	0.95%	43.05%	26.15%	5.30%	0.88%	61.89%	10.25%	26.98%
75	25.11%	1.60%	43.10%	27.33%	2.86%	1.45%	62.82%	12.63%	23.10%

. The results indicate that with the increase in E_j, the atomic fraction of V in square and circular precipitates is increased.

The EBSD inverse pole figure maps and the corresponding kernel average misorientation (KAM) maps of the welding thermal cycle at E_j = 15–75 kJ/cm are shown in Figure 7a–d and Figure 7e–h, respectively. The high-angle grain boundaries (HAGBs) are defined by misorientation tolerance angles (MTAs) of boundaries higher than 15° and marked by black lines, while the low-angle grain boundaries (LAGBs) are marked by white lines, which are defined by MTAs of boundaries ranging from 2°–15° [34]. The mean equivalent diameter (MED) of grains with varied MTA is evaluated, with the results counted in

Table 4. MED of grains with boundaries at MTAs from 2° to 15°.

Ej kJ/cm	MEDMTA≥2° /μm	MEDMTA≥4° /μm	MEDMTA≥6° /μm	MEDMTA≥8° /μm	MEDMTA≥10° /μm	MEDMTA≥12° /μm	MEDMTA≥15° /μm
15	2.75	3.15	3.45	4.15	5.21	5.47	6.36
35	3.16	3.6	4.13	4.96	5.62	5.76	6.73
55	4.03	4.54	5.12	5.67	5.79	5.84	6.15
75	4.69	5.56	5.68	5.76	5.87	5.91	5.96

MED—Mean equivalent diameter; MTAs—Misorientation tolerance angles; MED_MTA≥θ—The MED of grains with MTAs higher than θ.

. With the increase in E_j, coarsening occurs in the MED defined by MTAs of 2°, 4°, 6°, 8°, 10° and 12°. Moreover, the KAM maps can be employed to estimate the degree of plastic deformation and the density of geometrically necessary dislocations (GNDs) [35,36]. The average KAM value (KAM_ave) decreases from 0.83° to 0.55° with the increase in E_j; this trend is also consistent with the results of TEM and XRD observations, where an increase in E_j leads to a decrease in dislocation density.

3.2. Mechanical Properties of CGHAZ

The tensile properties and Vickers hardness of CGHAZ samples with different welding heat inputs were tested, and the relevant results are shown in

Table 5. Summary of the tensile results at different E_j.

Heat Input/kJ·cm−1	YS/MPa	TS/MPa	YR	uE/%	Micro-Hardness /HV10
15	480 ± 5	615 ± 4	0.78	9.5 ± 0.5	238.4 ± 3.5
35	461 ± 6	605 ± 3	0.76	13.2 ± 0.7	222 ± 5.2
55	437 ± 6	585 ± 3	0.75	16.4 ± 0.7	210 ± 4.8
75	416 ± 5	574 ± 4	0.72	18.6 ± 0.6	201 ± 3.5

, and the typical CGHAZ stress-strain curves and detailed tensile properties of each E_j sample are shown in Figure 8a and Figure 8b, respectively. With the increase in E_j, the hardness of the CGHAZ reduced from 238.4 HV10 to 201HV10, and the YS and TS decreased from 480 MPa to 416 MPa and from 615 MPa to 574 MPa, respectively, while the YR decreased monotonically from 0.78 to 0.72. In addition, the uE increased significantly from 9.5% to 20.1%.

4. Discussion

4.1. Effect of E_j on CGHAZ Microstructure

With the increase in E_j from 15 kJ/cm to 35 kJ/cm and 55/75 kJ/cm, the main CGHAZ microstructure changes from LBF to GBF, and further changes to a mixture of PF and IGAF. The evolutions in microstructure are related to the difference in the cooling process caused by varied E_j. When the E_j = 15 kJ/cm to 35 kJ/cm, the high-temperature residence time is shorter and the cooling rate is faster, which inhibits the diffusion of C atoms, and the supercooling degree is larger and hence promotes the transformation of bainitic ferrite [37]. The PAGBs are known to be the nucleation sites of bainitic ferrite due to the irregular atomic arrangement and lattice distortion [11]. The γ (austenite)→LBF transition tends to require a greater degree of supercooling than the γ→GBF transition [8]; therefore, with the increase in E_j, the nucleation of LBF is inhibited, resulting in a large amount of GBF at E_j = 35 kJ/cm, while the microstructure is mainly LBF at E_j = 15 kJ/cm.

In addition, it is noteworthy that uniform nucleation of ferrite and heterogeneous nucleation of ferrite are very difficult to detect under small heat inputs, which is related to limited carbon diffusion and high undercooling in favor of bainite ferrite transformation, as well as to heteronuclear particles. Although the nucleation temperatures of GBF and IGAF are both between 400–600 °C according to Ref. [38], at small heat inputs, the precipitated particles that reach the critical nucleation size (particles with an equivalent circular diameter higher than 100 nm have nucleation ability according to Refs. [38,39]) are Ti (C, N) particles (Figure 6a,b). Previous studies [40,41] have reported that the ability of Ti (C, N) particles as heterogeneous nucleation sites is weak; as shown in Figure 9a, Ti (C, N) particles fail to promote heterogeneous nucleation of ferrite. The competitive nucleation and growth of GBF and IGAF under small heat inputs are mainly dominated by GBF, while IGAF is inhibited. Furthermore, increasing the E_j to 55/75 kJ/cm, the high-temperature residence time is extended, the corresponding undercooling decreases, and the diffusion of carbon atoms is promoted, all of which are beneficial for γ→α (ferrite) phase transition to occur at a higher temperature. Meanwhile, the precipitated particles transform from Ti (C, N) to V-rich (Ti, V) (C, N) particles (Figure 6c,d). Compared with Ti (C, N) particles, the mismatch between V-rich (Ti, V) (C,N) particles and ferrite is lower, which can significantly promote the heterogeneous nucleation of ferrite [42], as shown in Figure 9b; the heterogeneous nucleation of the ferrite on the V-rich (Ti, V) (C, N) particles is evident. Accordingly, the nucleation at PAGBs is mainly GBPF, while the heterogeneous nucleation of IGAF and IGPF occurs in the grain; thereby high proportions of PF and IGAF are formed in the CGHAZ when E_j = 55 and 75 kJ/cm.

In fact, the γ→α phase transformation is accompanied by the diffusion of C atoms from α to γ due to the higher solubility of C atoms at γ, and the C-rich retained austenite (γ’) forms during this period. γ’ can be stable at relatively high temperature because of the high C content, and then transform into M/A constituents in the air cooling process [43]. With the increase in E_j from 15 kJ/cm to 75 kJ/cm, the high-temperature residence time gradually prolongs, and the diffusion of C becomes more sufficient. Therefore, with the increase in E_j, the mean size and area fraction of M/A constituents increase.

4.2. Effect of E_j on CGHAZ Tensile Properties

With the increase in E_j, the grain size, the dislocation density, and the size and volume fraction of precipitated particles in CGHAZ will all change, and these will have an effect on the strength. According to previous research [28,44], the YS of low-alloy steel is generally determined by the cumulative contributions of various strengthening mechanisms, and the main strengthening modes are shown in Equation (3):

$${\mathsf{\sigma }}_{\mathrm{y}}={\mathsf{\sigma }}_0+{\sigma }_{ss}+{\sigma }_d+{\sigma }_{\rho }+{\sigma }_{pp}+{\sigma }_{M/A}$$

(3)

Here, σ_y is the YS, σ₀ represents the Peierls–Nabarro stress or lattice friction stress, σ_ss represents the solid solution strengthening, σ_d represents the boundary strengthening, σ_ρ represents the dislocation strengthening, σ_pp represents the precipitation strengthening, and σ_M/A represents the M/A constituent strengthening. Among these strengthening mechanisms, the boundary strengthening exerted by LAGBs and the dislocation strengthening exerted by dislocation structures in the microstructure contribute the greatest [28,45]. With the increase in E_j, the MED defined by LAGBs are coarsened (Figure 7 and Table 4), while the dislocation density decreases (Figure 4 and Figure 5), resulting in a monotonic decrease in YS of CGHAZ.

Typically, the LAGBs defined by MTAs ranging from 2° to 15° have the most effective contribution to boundary strengthening [22], the HAGBs (θ ≥ 15°) have the most significant contribution to blocking crack propagation [21,46], and the remaining boundaries (θ < 2°) contribute to dislocation strengthening [47]. YS and grain size can be described by the Hall–Petch equation, as shown in Equation (4).

$${\sigma }_y={\sigma }_0*+{ \mathrm{k}}_{\mathrm{hp}}{d}^{-1/2}$$

(4)

Here, σ_y is the YS, σ₀* is the is the sum of other strengthening methods except the boundary strengthening, k_hp is the structural constant, and d is the grain size corresponding to MED defined by varied MTAs. It is worth noting that the microstructure is complex; under different E_j conditions, there are obvious differences in the size, morphology and distribution of CGHAZ microstructures. In this paper, with the change in E_j, the YS have been plotted as a function of the MED^−1/2 with MTAs ranging from 2° to 12°, with the results shown in Figure 10. The results indicate that the MED_{MTA≥2–6°}^−1/2 defined by MTAs ranging from 2° to 6° showed a fairly strong linear fit with YS, with the correlation factor of 0.99–0.96 (Figure 10a–c). However, Figure 10 d–f show the correlation factor was reduced to 0.82–0.76 for the YS and MED_{MTA≥8–12°} defined by MTAs ranging from 8° to 12°. Accordingly, the MED^−1/2 defined by MTAs ranging from 2° to 6° had a higher correlation factor compared with MED^−1/2 defined by MTAs ranging from 8° to 12°. This result made it clear that that MED defined by MTA ranging from 2° to 6°of LAGBs is the most efficient microstructural unit to controlling YS; this result is also consistent with that reported in related literature [28].

Furthermore, with the increase in E_j from 15 kJ/cm to 75 kJ/cm, the YR monotonically decreases, which means that an increase in E_j leads to an increase in strain-hardening ability. This phenomenon is mainly related to the area fraction and size of the M/A constituents. M/A constituents exist in ferrite matrixes as hard phases, and their hardnesses are higher than those of ferrite matrixes, and M/A constituents can be plastically deformed by two-stage or three-stage work hardening. This deformation mechanism is mainly controlled by high-strength local plastic deformation; the local internal stress applied from the M-A constituent hard phase acts on the ductile ferrite phase, thereby improving the strain-hardening ability [48,49]. Accordingly, with the increase in E_j from 15 to 75 kJ/cm, the d_M/A increases from 0.98 μm to 1.81 μm, and f_M/A increases from 3.11% to 4.42%, leading to an increase in the strain-hardening ability, and the YR monotonically decreases.

Furthermore, the uE increases significantly with the increase in E_j from 15 to 75 kJ/cm. The uE is closely related to slip of dislocation: once dislocations slip over long distances without encountering barriers, high uE is generated [50,51]. When the E_j = 15 and 35 kJ/cm, there are many DTs and ferrite lath boundaries in the microstructure, and the dislocation density is high (Figure 4a,b and Figure 7e,f). The dislocations will be significantly hindered after short-distance movement during the stretching process. While E_j = 55 and 75 kJ/cm, the proportion of PF in microstructure increases, relatively uniform dislocations are distributed in the microstructure, and the dislocation density is low (Figure 4c,d and Figure 7g,h), these are beneficial for long-distance slip of dislocations and obtaining higher uE. In addition, the corresponding characterization results of tensile fracture are shown in Figure 11. The tensile fractures under low E_j and large E_j are composed of two parts: fiber zone and shear-lip zone: the fiber zone is mainly composed of dimples, but deeper and smaller dimples appear at E_j = 75 kJ/cm. Further, although microvoids can be found in all fiber regions, the number and the size of microvoids are larger at E_j = 15 kJ/cm. Compared to deeper dimples, the appearance of large-sized microvoids indicates that microcracks are more likely to occur during the stretching process [52], which is not conducive to obtaining high uE. Accordingly, the characterizations of tensile fracture further demonstrate that increasing E_j can improve uE.

5. Conclusions

The effects of heat input on CGHAZ microstructure and tensile properties were studied for a novel V-N-Ti microalloyed weathering steel, and the following main conclusions can be drawn:

• With the increase in E_j from 15 kJ/cm to 35 kJ/cm, and 55/75 kJ/cm, the dominant CGHAZ microstructure transforms from LBF to GBF, and then to a mixture of IGAF+PF. The size and area fraction of M/A constituents increase monotonically with the increase in E_j, the d_M/A increases from 0.98 μm to 1.81 μm, and the f_M/A increases from 3.11% to 4.42%.
• With the increase in E_j, the increase in MED defined by the LAGBs and a reduction in dislocation density led to the decrease in YS. And the MED with MTAs ranging from 2°–6° strongly controlled the YS due to their higher fitting coefficient of the Hall–Petch relationship.
• With the increase in E_j, the decrease in YR was mainly due to the increase in the size and area fraction of M/A constituents, which improved the strain-hardening ability. Meanwhile, the lower density and more uniform dislocation structures improved the uE.