Effect of Nb–V microalloying on the hot deformation behavior of medium Mn steels —— Effect of Nb–V microalloying on the hot deformation behavior of medium Mn steels

1.   Introduction

For crash safety, a lightweight design is an effective way for the automotive industry to improve fuel efficiency and reduce CO₂ emissions. Advanced high-strength steels (AHSSs) emerge as a viable solution, catering to the industry’s trifecta of priorities, namely, economic viability, environmental sustainability, and safety enhancement. AHSSs have been developed into the third generation (3rd Gen) and have garnered considerable attention owing to their exceptional mechanical properties and heightened cost competitiveness [1–4]. Medium Mn steels (MMSs) with Mn concentrations in the range of 4wt%–10wt% have been widely regarded as promising candidates for 3rd Gen AHSSs. To date, the industrialized production of 3rd Gen AHSSs is still highly restricted by many processing difficulties [5–6]. In hot working, the metal experiences work hardening during plastic straining and softening due to dynamic recovery (DRV) and dynamic recrystallization (DRX). Cross-slip and climb of dislocations are usually difficult to achieve in the austenite of MMSs during hot deformation due to low stack fault energy. Hence, DRX is preferred over DRV. DRX can eliminate dislocations and even the cracks generated by work hardening and refine the prior austenite grains [7–10]. The optimization of hot deformation parameters through flow behavior modeling is of paramount significance in delineating the industrial processing parameters of MMSs. Sun et al. [11] established a finite element model for the hot compression of 0.15C–7Mn steel to derive the DRXed fraction and the average grain sizes under different hot deformation conditions using DEFORM software, and the findings are in good agreement with the experimental results. Liu et al. [12] established a critical strain model on the DRX of 0.22C–8.5Mn steel using the P–J double differential method and finally constructed the hot processing maps for determining the optimal processing window.

Nb and V are often added into MMSs as microalloying elements for grain refinement and precipitation strengthening [13–14]. Current studies mainly focus on the effects of Nb and V microalloying on mechanical properties and microstructure evolution. Luo et al. [15] found that 0.04wt%Nb microalloying significantly improved the hydrogen embrittlement resistance of hot-rolled 7Mn–3.5Ni–1.5Cu–0.08C steel. Nb addition resulted in a high C concentration at the prior austenite grain boundaries (PAGBs) and the formation of filmy retained austenite along the PAGBs, which served as a suitable hydrogen-trapping site that reduced hydrogen diffusion and thus inhibited the occurrence of hydrogen-induced cracks. By contrast, Varanasi et al. [16] reported that Nb addition was detrimental to the nucleation and growth of austenite during the intercritical annealing of cold-rolled 10Mn–0.05C–1.5Al steel, leading to reduced retained austenite fraction and deteriorated plasticity. Zhu et al. [17] showed that Nb and V addition to hot-rolled 0.3C–9Mn–2.5Al steel significantly increased the yield strength and had almost no detrimental effect on ductility.

The effects of Nb–V addition on the hot deformation behavior of MMSs remain largely unexplored. In this study, the hot deformation behaviors of MMSs with and without Nb–V were investigated systematically by hot compression tests. The constitutive equations were established for the two steels to describe the flow behavior during hot deformation, and the resultant processing maps and microstructures for the two steels were compared.

2.   Methods

The nominal chemical compositions of the two studied steels are listed in Table 1. Both contained approximately 0.1wt% C, 5wt% Mn, and 2wt% Al; one was microalloyed with Nb and V (termed 5MnNbV), and the other was not (5Mn).

Table  1.  Chemical composition of the studied MMSs wt%

MMS	C	Mn	P	S	N	Al	V	Nb
5Mn	0.11	4.89	0.005	0.002	0.002	1.96	—	—
5MnNbV	0.11	4.90	0.005	0.003	0.002	1.96	0.19	0.0079

 | Show Table

DownLoad: CSV

Thermodynamic calculations on the two steels were performed using Thermo-Calc software with the database of TCFE9. The calculation results in Fig. 1 indicated that Ae₃ temperatures, the ortho-equilibrium austenite-to-ferrite transformation temperatures, of 5Mn and 5MnNbV steels are 929°C and 972°C, respectively, and NbC and VC have completely dissolved at 1180°C and 848°C, respectively. In the temperature range of 1000–1150°C employed for hot compression, both steels are in the austenitic region, and V cannot precipitate as carbide in austenite. By contrast, NbC can precipitate and gradually dissolve with the increasing temperature.

Fig. 1.  (a) Temperature dependence of equilibrium volume fraction of different phases for the two steels calculated by Thermo-Calc software and (b) enlarged view on precipitations.

DownLoad: Full-Size Img PowerPoint

The alloys were melted in a 50-kg vacuum-induced furnace and then hot forged into rectangular billets. The cylindrical specimens with a diameter of 8 mm and a height of 12 mm were machined for the hot compression experiments. Unidirectional compression tests were performed on a Gleeble 3800 thermo-mechanical simulator. Tantalum sheets with a thickness of 0.1 mm were glued with conductive adhesive between the specimen and the anvil to provide isolation and lubrication. Hot compression experiments were conducted at temperatures ranging from 1000 to 1150°C with an interval of 50°C and a strain rate range of 0.001–1 s⁻¹. The specimens were first heated to 1200°C with 10°C/s for 300 s, followed by cooling to the target temperature with 10°C/s for 60 s before compression. All the specimens were compressed to a true strain of 0.6 at the specified strain rate, followed by immediate water quenching to room temperature to preserve their microstructures.

Microstructures were examined under optical microscope and transmission electronic microscope (TEM) attached with energy dispersive spectrum (EDS) in the central section of the compressed specimen cut by a wire parallel to the compression direction after standard grinding, polishing, and etching in the saturated picric acid at 63°C. The statistic results on the prior austenite grain sizes (PAGSs) in the specimens were derived with ImageJ software.

3.   Results and discussion

3.1   Flow behavior during hot compression

Fig. 2 shows the flow stress curves of the two studied steels at different temperatures and strain rates ($\dot{\varepsilon}$). In all the curves, work hardening exists at the initial stage of deformation, implying significant dislocation multiplication. With further straining, the deformation storage energy increases gradually until it reaches a critical value to trigger DRX. Subsequently, the flow stress starts to decrease with strain due to the dynamic softening effect. Changes in the flow stress essentially reflect the competition between the propagation and annihilation of dislocations [18]. When the work hardening and softening mechanisms reach a dynamic equilibrium, the flow curves flatten out. Pronounced peak stresses are widely indicative of DRX. For the flow stress curve without a distinct peak, DRV is generally the main softening mechanism under the corresponding deformation conditions. Some studies claim that DRX may occur even in the absence of a significant peak in the flow curves [19–20]. Especially at high strain rates, dislocations proliferate rapidly, and the work hardening effect is enhanced to such an extent that the softening mechanism is difficult to overcome.

Fig. 2.  Flow stress–strain curves of the experimental MMSs at different deformation conditions: (a) $\dot{\boldsymbol{\varepsilon}}$ = 0.001 s⁻¹; (b) $\dot{\boldsymbol{\varepsilon}}$ = 0.01 s⁻¹; (c) $\dot{\boldsymbol{\varepsilon }}$ = 0.1 s⁻¹; (d) $\dot{\boldsymbol{\varepsilon}}$ = 1 s⁻¹.

DownLoad: Full-Size Img PowerPoint

When the strain rate is constant, the flow stress decreases with the increasing temperature because many dislocations are annihilated, and grains grow rapidly at high temperatures [11,21]. When the temperature is constant, the flow stress increases with the strain rate because dislocations build up rapidly, and stress concentrations cannot be released. A low strain rate can provide additional time for the rearrangement and annihilation of dislocations [22].

Fig. 3 illustrates the TEM and EDS results on the precipitates in 5MnNbV steel after solution treatment at 1200°C and hot compression at 1000°C with a strain rate of 0.001 s⁻¹. The nanosized particles precipitate at grain boundaries and dislocations and have a size of (107.5 ± 14.2) nm, as shown in Fig. 3(a). Energy dispersive X-ray spectroscopy (EDX) compositional examination in Fig. 3(b) indicates that the particles are mainly enriched with Nb, i.e., NbC.

Fig. 3.  TEM and EDS results of 5MnNbV after hot deformation at 1000°C with a strain rate of 0.001 s⁻¹: (a) TEM bright-field image of 5MnNbV; (b) EDX maps of the precipitates.

DownLoad: Full-Size Img PowerPoint

Fig. 4(a) shows the variation of the peak stress σ_p with deformation temperature and strain rate for both of the studied steels. When hot deformation is performed at the same temperature and strain rate, 5MnNbV exhibits higher σ_p than 5Mn. However, the difference between them decreases with the increase in deformation temperature until 1150°C because many NbC particles are dissolved and get coarse at high temperatures, leading to reduced inhibition on dislocation motion and grain boundary migration.

Fig. 4.  Temperature dependence of the peak stress σ_p (a) and critical strain ε_c (b) of the two studied MMSs at different strain rates.

DownLoad: Full-Size Img PowerPoint

Critical strain (ε_c) is the minimum strain required for DRX. Fig. 4(b) shows the critical strains for the two steels as calculated by the P–J double differential method [23] at the strain rates of 0.001–0.1 s⁻¹. At 1 s⁻¹, DRX occurs in both steels only at 1150°C, and the critical strains for 5MnNbV are higher than those for 5Mn, suggesting a pronouncedly delayed DRX due to Nb and V addition. In particular, the difference in ε_c for the two studied steels is large and gradually diminishes with the increase in temperature at the low strain rates of 0.001–0.01 s⁻¹, indicating the diminishing inhibition on DRX. Solute Nb and NbC particles can impede DRX. When the strain rate is low, Nb as a solid solution exerts less hindrance to DRX compared with its role as a precipitate within the temperature range considered in this study. This finding is consistent with previous research, stating that the hindering effect of NbC particles on grain boundary migration is usually pronounced in austenite [17].

At 0.1 s⁻¹, the difference in ε_c of the two steels increases with the increase of deformation temperature, which is different from the case at 0.01 and 0.001 s⁻¹. This phenomenon is due to the great difference in deformation time, which is 6, 60, and 600 s for the strain rate of 0.1, 0.01, and 0.001 s⁻¹, respectively. At 0.1 s⁻¹, the time for NbC to dynamically precipitate is insufficient. At low temperatures and high strain rates, the precipitation kinetics of NbC are considerably less favorable. Hence, the two studied steels exhibit comparable critical stresses of DRX during deformation at 1000–1050°C with a strain rate of 0.1 s⁻¹.

DRX leads to the softening of the hot deformed matrix, so the flow stress after DRX is lower than the peak stress. Therefore, the stress (σ) reduction after the peak stress reflects the softening effect as shown in Fig. 5. The 5Mn steel exhibits more pronounced softening at 1000–1150°C after the small peak strain than 5MnNbV, suggesting that Nb–V microalloying delays the onset of DRX and significantly increases the hot strength of the recrystallized matrix. The latter may result from the fine size of recrystallized grains at high temperatures or the low recrystallized extent at low temperatures resulting from Nb–V alloying, which will be discussed later.

Fig. 5.  Influence of Nb–V microalloying on the stress reduction, σ_p − σ, after the peak stress at the strain rate of 0.01 s⁻¹.

DownLoad: Full-Size Img PowerPoint

3.2   Constitutive equations

During hot working, the hot deformation behavior is influenced by strain, strain rate, deformation temperature, and microalloying elements [18] and is commonly expressed using the Arrhenius equation [24–25]:

where T, $\dot{\varepsilon}$, and σ are the temperature in K, strain rate in s⁻¹, and stress in MPa, respectively; R is the universal gas constant 8.314 in J·mol⁻¹·K⁻¹; Q represents the activation energy for hot deformation in kJ/mol.

The f(σ) in Eq. (1) can be described by the power function, the exponential function, and Sellars’ hyperbolic sine function as given below [26–27]:

where A₁, A₂, A, n₁, n, β, and α are material constants, and σ is usually taken as the peak stress σ_p.

For the constant T, n₁ and β can be derived by linear fitting of $\mathrm{ln}\,\sigma$ vs. $\mathrm{ln}\,\dot{\varepsilon}$ plots and $\sigma$ vs. $\mathrm{ln}\,\dot{\varepsilon}$ plots, respectively. The stress multiplier α ($\alpha \approx \beta /{n}_{1}$) is an adjustable constant that ensures that ασ is in the appropriate range, making the $\mathrm{ln}\,\dot{\varepsilon}$ vs. $\mathrm{ln} \left[\mathrm{sinh}\left(\alpha \sigma \right)\right]$ plots linear and parallel at constant temperature [28–29]. n is derived by the linear fitting of $\mathrm{ln} \left[\mathrm{sinh}\left(\alpha \sigma \right)\right]$ vs. $\mathrm{ln}\,\dot{\varepsilon}$ plots (Fig. 6(a) and (b)). Besides, according to Eq. (4), when the strain rate is constant, the mean value of Q can be derived by the linear fitting of $\mathrm{ln} \left[\mathrm{sinh}\left(\alpha \sigma \right)\right]$ vs. 1/T plots (Fig. 6(c) and (d)). The determined apparent activation energies are 377.775 and 415.722 kJ/mol for 5Mn and 5MnNbV, respectively. Both values are higher than the self-diffusion activation energy of Fe in γ-Fe (280 kJ/mol) [30] and the hot deformation activation energy for plain carbon steels (300 kJ/mol). The Q value of 5Mn steel is lower than those for some high Mn steels, e.g., 391 kJ/mol for Fe–23Mn–2Al–0.2C [31] and 387 kJ/mol for Fe–20Mn–3Si–3Al [32], suggesting that the high contents of Mn, Al, and C in a solution could raise the activation energy because these elements can restrict the movement of dislocations. Meanwhile, Nb and V exhibit stronger retardation than Mn, so the 5MnNbV steel has even higher deformation activation energy than high Mn steels.

Fig. 6.  Relationships curves of $\mathbf{ln}\,\dot{\boldsymbol{\varepsilon}}$ vs. ln[sinh(ασ)] and 1/T vs. ln[sinh(ασ)] at the strain of 0.6 for (a, c) 5Mn steel and (b, d) 5MnNbV steel.

DownLoad: Full-Size Img PowerPoint

The Zener–Hollomon parametric equation is widely used to analyze the effect of deformation temperature and strain rate on the hot deformation behavior of metallic materials [24,33]:

Combination of Eqs. (4) and (5) leads to:

The linear relationship of $\mathrm{ln}\,Z$ vs. $\mathrm{ln} \left[\mathrm{sinh}\left(\alpha \sigma \right)\right]$ for the two steels is confirmed, as shown in Fig. 7. A can be determined from the intercept of the linear fittings.

Fig. 7.  Relationships between ln Z and ln[sinh(ασ)]: (a) 5Mn and (b) 5MnNbV. r is Pearson’s correlation coefficient

DownLoad: Full-Size Img PowerPoint

Finally, the constitutive equations are derived by substituting the above-derived parameters into Eq. (1) for the two steels:

3.3   Activation energy maps

In the foregoing analysis, the average deformation activation energy has been determined for both studied steels. However, the activation energy is actually the resistance to deformation, which depends on the capacity of dislocations to surmount obstacles during plastic deformation and should be strongly influenced by hot deformation parameters, including temperature and strain rate [34]. In this case, a map of activation energy is required to describe the variation of Q with deformation temperature and strain rate. Q can be expressed as:

The values of n and S are obtained from the slopes of the $\mathrm{l}\mathrm{n}\,\dot{\varepsilon}\;\mathrm{v}\mathrm{s}.\;\mathrm{ }\mathrm{l}\mathrm{n} \left[\mathrm{sinh} \left(\alpha \sigma \right)\right]$ plots at different temperatures and the slopes of the $\mathrm{l}\mathrm{n} \left[\mathrm{sinh} \left(\alpha \sigma \right)\right]$ vs. 1/T plots at different strain rates, respectively. The Q at each temperature and strain rate can be derived by fitting a polynomial to the n–T and S–$\dot{\varepsilon}$ plots and substituting the results into Eq. (9). Fig. 8 shows the variation of calculated Q values with temperature and strain rate for both steels at the strain of 0.6. At 1 s⁻¹, both steels exhibit high activation energy. This phenomenon arises from the reduced time available for dynamic softening and enhanced dislocation multiplication due to work hardening, both increasing the energy requirement for dislocation motion [35].

Fig. 8.  Hot deformation activation energy (Q) maps for two studied MMSs at the strain of 0.6: (a) 5Mn and (b) 5MnNbV.

DownLoad: Full-Size Img PowerPoint

Q serves as an indicator of the deformation mechanism. A relatively stable and low Q region generally corresponds to the feasible processing window [36], and a region with rapidly changing Q represents the unfavored window. Under the experimental conditions, the Q values of 5Mn and 5MnNbV are in the ranges of 261–498 kJ/mol and 336–499 kJ/mol, respectively. The former shows a large variation in the region of high temperature and high strain rate, which may result in flow instability [37]. Meanwhile, the Q values for 5MnNbV remain notably stable, suggesting its superior hot workability. In the region of low temperature and low strain rate, the Q of 5MnNbV is significantly higher than that of 5Mn because many NbC particles can dynamically precipitate during deformation in this region. In conclusion, Nb–V microalloying results in the overall increase in Q and the reduced the sensitivity of Q to temperature and deformation rate, the latter improving the hot working performance. In addition to the activation energy map, the processing maps based on microstructural analysis are required to determine the optimal processing window.

3.4   Processing maps

Processing maps are an effective way to evaluate the processing performance of materials and are often used to predict and analyze the characteristics and mechanisms of material deformation under different deformation conditions. They are used to optimize the manufacturing process and avoid hot processing defects.

The parameter η is introduced to represent the energy dissipation efficiency due to microstructure evolution and can be expressed as [27]:

where J represents the energy used for microstructure evolution. Its value reaches the maximum (J_max) when the material is in an ideal linear dissipative state (m = 1). m represents the strain rate sensitivity parameter, which is defined as the ratio of the energy used for microstructural evolution (J) to the energy dissipated by plastic deformation (G):

In addition to the favored deformation mechanisms such as DRX for eliminating the original defects in a material, the presence of detrimental mechanisms such as dynamic strain failure, void formation, and adiabatic shear bands can contribute to an increase in the power dissipation rate η. Some supplementary criteria are necessary to assess the processing instability of the material, enabling a discerning judgment and the avoidance of defects during hot processing.

The instability criterion proposed by Prasad and Seshacharyulu [38] suggests that material processing enters the flow instability region when Eq. (12) is satisfied:

where $\xi$ is the instability parameter. When $\xi < 0$, flow instability, such as the appearance of localized flow regions, shear zones, and microcracks, may occur [39]. The power dissipation efficiency and instability parameter are calculated from Eqs. (10) and (12), both are functions of temperature and strain rate. Their results at the same strain are superimposed to obtain the hot processing maps of the two studied steels as shown in Fig. 9. The gray area is the flow instability region, and the numbers on the contours represent the energy dissipation efficiencies. In general, high dissipation efficiencies indicate that the material is suitable for processing under these conditions. The flow instability regions of both studied steels are mainly concentrated in the high-strain-rate region, and the instability region for 5MnNbV is significantly smaller than that for 5Mn. These findings suggest that the Nb–V alloyed steel has a wide processing window for hot working without generating defects.

Fig. 9.  Processing maps of experimental MMSs at a strain of 0.6: (a) 5Mn and (b) 5MnNbV. The gray area is the region of flow instability, and the numbers on the contours represent the energy dissipation efficiencies.

DownLoad: Full-Size Img PowerPoint

3.5   Microstructural analysis

The microstructures formed after hot compression at 1100°C with a strain rate of 0.1 s⁻¹ are given in Fig. 10(a) and (b). The average grain sizes of 5Mn and 5MnNbV are 20.04 and 15.28 μm, respectively. Fully recrystallized grains are observed in both of them. Some grains in 5Mn have become coarse (Fig. 10(a) and (c)), suggesting that the unfavored mixture of coarse and fine sizes is formed under this processing condition. In 5MnNbV, uniformly distributed fine PAGS is found due to the presence of either solute Nb or fine NbC. This result is consistent with those observed under deformation in the unstable region in Fig. 9(a) and the stable region in Fig. 9(b). The microstructures formed after hot compression at 1100°C with 1 s⁻¹, i.e., the unstable region in Fig. 9 for the two studied steels, are given in Fig. 10(c) and (d). The microstructures in both specimens are a mixture of fine DRXed and coarse deformed grains, indicating an unstable state during hot working. Therefore, the microstructural examinations confirm the validity of the derived process maps in Fig. 9 and the significant improvement in the hot workability of MMS by Nb–V alloying.

Fig. 10.  Optical images of the microstructures in 5Mn (a, c) and 5MnNbV (b, d) after hot deformation (a, b) at 1100°C with strain rate of 0.1 s⁻¹ and (c, d) at 1100°C with strain rate of 1 s⁻¹.

DownLoad: Full-Size Img PowerPoint

4.   Conclusions

Hot compression tests were performed on Nb–V microalloyed and free MMSs at temperatures of 1000–1150°C with strain rates of 0.001–1 s⁻¹. The flow curves of the two steels were simulated by the established constitutive equations, from which the deformation activation energy and processing maps were constructed to identify the unstable region. Combining the above with the microstructural examination results, the following conclusions are drawn.

(1) For the two studied MMS, most flow curves exhibit distinct peaks and a subsequent flattening, indicating the occurrence of DRX. Moreover, Nb–V microalloying strongly retards DRX during the hot deformation of MMS.

(2) Under the same hot deformation condition, Nb–V microalloying leads to high peak stress. The difference in critical strain between 5MnNbV and 5Mn steels decreases with the increase in deformation temperature, suggesting that NbC precipitates could retard DRX more efficiently than solute Nb in the studied temperature range.

(3) Nb–V microalloying also results in a significant increase in the deformation activation energy of 5Mn from ~378 to ~416 kJ/mol. Since the reported activation energy for some high Mn steels is in between these two values, we propose that Nb–V alloying could produce stronger retardation on DRX than high Mn content.

(4) Comparisons of hot deformation activation energy and processing maps show that Nb–V microalloying results in the reduced dependence of Q value on temperature and strain rate and a large region of stability. As a consequence, the hot workability of MMS is improved by narrowing the hot processing window for the unstable flow stress, in which fine recrystallized and coarse unrecrystallized grains are present.