Effect of strain rate on the work-hardening rate in high-Mn steel

Introduction

The austenitic wear-resistant high manganese steel, containing 1.0–1.2 wt-% C and 12–14 wt-% Mn, was developed by Robert Hadfield [1]. It combines high toughness and ductility with high work-hardening capacity and excellent resistance to wear for that: strain-induced γ→α′ or γ→ transformation [2], mechanical twinning [3], dislocation hardening [4], etc. These outstanding properties make it used widely in the applications of metallurgy, mining, thermal power plant, construction and manufacturing. However, its high work-hardening properties want heavy stress or high load impact, or else, its work-hardening properties will be poor. What is more, its yield strength is low (Generally <300 MPa), and plastic flow is easy to occur. In order to tackle these problems in Hadfield manganese steel, the results of study from Wen et al. show that -martensite obtained in high manganese steel can improve the work-hardening ability and the hardness to enhance the wear resistance [5]. The strain-induced γ→α′ or γ→ transformation is based on the stacking-fault. The stacking-fault energy (SFE) of austenite can be calculated as follows [6]: (1) where ΔG^γ→ is the free molar enthalpy of the transformation γ→ , ρ is the molar surface density of atoms in the {111} plane and σ^γ/ is the energy per surface unit of a {111} interface between γ phase and -martensite, 8 mJ m⁻².

According to the previous simplified model for the Fe–Mn–C system [7], the free molar enthalpy of martensite formation can be shown as follows [6,8]: (2) Besides, knowledge is still limited concerning the effect of strain rate on SFE and mechanical properties [9]. Previous research confirmed that the decrease in elongation of TWIP steel at high strain rate was a direct result of adiabatic heating [10]. The adiabatic temperature rise is calculated by the following formula [11]: (3) where β is a coefficient related to the fraction of mechanical energy converted to heat energy, assumed to be 90%, C_p is the typical heat capacity of steels which is taken as 0.46 kJ (kg K)⁻¹, ρ is the density as 7.8 g cm⁻³. Plastic deformation at high strain rate generates a large amount of strain energy in the material. There is less time for the heat to dissipate to the surroundings, resulting a significant increment of temperature. The highest surface temperature was measured to be ∼120°C for the specimens deformed at 1 × 10¹ s⁻¹ but ∼25°C for the specimens deformed at 1 × 10⁻⁴ s⁻¹ [12]. The increase in temperature to 80°C causes an increase in the SFE of ∼4–5 mJ m⁻² for all studied TWIP grades [11]. What is more, the flow stress of Fe–Mn–C TWIP steels drops at high strain rate for its negative strain rate sensitivity (SRS) [10,13]. But, the yielding strength generally increases with strain rate, showing substantial positive SRS [14]. High strain rate leads to a reduction in the work-hardening rate [15,16].

Moreover, temperature changes also influence the dynamic strain aging (DSA) due to the temperature dependence of the reorientation of Mn–C point defect complexes close to the stacking-fault plane [17], so the effect of DSA can be observed at low strain rate deformation and may be suppressed at high strain rate. The DSA effect can increase the localised stress concentration to facilitate deformation twin nucleation [16,18]. While nucleating mechanism twin is the same to -martensite, the DSA effect can also facilitate -martensite nucleation.

In order to further study the mechanical properties and microstructure of high manganese austenitic steel, the tensile tests were performed at different strain rates of 10⁻¹ s⁻¹, 10⁻² s⁻¹ and 10⁻³ s⁻¹. Then the work-hardening rate of high manganese austenitic steel is discussed through the analysis of the microstructure and mechanical properties.

Experimental procedure

The chemical composition of steel investigated in the present study is: Fe–0.40C–0.24Si–18.5Mn–0.72Al–0.005P–0.007S–0.026Nb–0.047Ti–0.49V (in wt-%). After 50 kg vacuum melting, the ingot was forged and then homogenised at 1150°C for 6 h. The heat treatment processing regimes were as follows: samples were first heated at 1080°C for 3 h, then quenched by water.

The tensile specimens of 1.5 mm thickness were cut from the sheet parallel to the forging direction according to the geometry using a wire-cutting machine. The tensile tests were performed at room temperature on a SANS XYB605C universal testing machine with a load capacity of 100 kN and a maximum velocity of 500 mm min⁻¹ in tension. The tests were performed at different strain rates of 10⁻¹ s⁻¹, 10⁻² s⁻¹ and 10⁻³ s⁻¹.

The specimens for observing fracture morphology were cut from the gauge area containing the fracture surface after the tensile tests, then observed using a field emission scanning electron microscope (FEI Company, Hillsboro, USA). For a more detailed analysis of the microstructures, the samples were characterised by a ZEISS ULTRA 55-type field emission scanning electron microscope equipped with a electron back-scattering diffraction (EBSD) unit. An accelerating voltage of 20 kV and a step size of 0.07 µm were used in EBSD scans.

Results

The calculated SFE and ΔG^γ→ of Fe–0.40C–0.24Si–18.5Mn–0.72Al at 293 K are 11.57 mJ m⁻² and −75.3 J mol⁻¹, implying that the austenite steel is highly metastable due to its low SFE. Low SFE (Γ_fcc ≤ 20 mJ m⁻²) favours the γ→ phase transformation, whereas high SFE (Γ_fcc > 20 mJ m⁻²) suppresses the phase transformation [19–21]. The mechanical properties of tested steel at room temperature are listed in Table 1. The result shows that yield strength increases with increasing of strain rate, but the tensile strength and elongation are opposite, which are shown in Figure 1(a). In addition, the higher the strain rate, the higher the product of strength and elongation, which indicated that the tested steel with low strain rate has a better combination of strength and ductility and more energy could be absorbed during damage. The SRS value m can be defined by [10]: (4) where σ and represent the ultimate stress and strain rate, respectively.

OPEN IN VIEWER

Figure 1

(a) The engineering stress–engineering strain curves and (b) The SRS value and (c) corresponding work-hardening rate-true curves of the tested steel at different strain rates.

OPEN IN VIEWER

Table 1

Mechanical properties of the tested steel at different strain rates.

Strain rate/s−1	σ0.2/MPa	σb/MPa	δ/%	σb × δ/(GPa·%)
10−3	265	805	58.50	47.09
10−2	275	760	57.07	43.37
10−1	311	727	53.01	38.54

As shown in Figure 1(b), the SRS value, m, is −0.2213. The negative value suggests that the mechanical property of the tested steel becomes worse during deformation with the increase of strain rate. This yields a softening mechanism, known as negative SRS [10].

The engineering stress–engineering strain curves of the tested steel at different strain rates are shown in Figure 1(a). Based on the data, their corresponding work-hardening rate versus true strain curves were calculated using the Considère's criterion [20]: (5) where σ denotes true stress and true strain. The results are shown in Figure 1(c) (both the elastic sections and the sections close to rupture are removed), which reveals that the effect of strain rate on work-hardening rate is significant. The work-hardening rates of all strain rates first decreased rapidly with the strain increasing up to 0.05, and then it decreased slowly with further increase of the strain until occurrence of fracture without necking, which is because γ→α′ or γ→ transformation occurs [22]. However, at a low strain rate of 10⁻³ s⁻¹, the steel exhibits highest work-hardening rate compared with other strain rates. The slope of work-hardening rate between 0.1 and 0.2 strain is almost constant, indicating the amount of transformation that occurs during the deformation process at a low strain rate of 10⁻³ s⁻¹. And then, work-hardening rate decreases slowly. At the strain rate of 10⁻² s⁻¹, there is a small platform with increasing strain. Then, the work-hardening rate decreases. At the highest strain rate of 10⁻¹ s⁻¹, there is no platform, and work-hardening rate decreases continuously. These also suggest that the volume fraction of -martensite decreases with the increase of strain rate for increasing SFE.

The relationship between the work-hardening rate sensitivity and SRS parameters can be written as follows [15]: (6) Note that σ > 0, and which implies that m′ is more negative than m. The expression also suggests that the ability of work-hardening rate becomes weakened with the increase of strain rate. The expression also indicates that work-hardening rate sensitivity is affected by which mechanisms. SRS m and ∂m/∂ depend on the DSA effect [15], and is affected on the twin and -martensite. Therefore, obvious work hardening occurs at a small strain rate.

Figure 2 shows SEM images of deformed samples under different strain rates of 10⁻¹ s⁻¹, 10⁻² s⁻¹ and 10⁻³ s⁻¹. At a low strain rate of 10⁻³ s⁻¹, the density of deformation band is the highest, the deformation band is the thinnest and the distance between two adjacent deformation bands is the smallest. With the increase of strain rate, the density of deformation band decreases, the deformation band becomes coarse and the distance between two adjacent deformation bands gets bigger. The reason for this phenomenon is that the nucleation of twin and -martensite is hindered at high strain rate. Moreover, high strain rate may improve the energy of the system to facilitate the growth of twin and -martensite [16], so the deformation band becomes coarse.

OPEN IN VIEWER

Figure 2

SEM-micrographs of deformed samples; (a) 10⁻³ s⁻¹, (b) 10⁻² s⁻¹, (c) 10⁻¹ s⁻¹.

Figure 3 shows the EBSD phase contrast maps with grain boundaries of tested steel under different strain rates. The results present that most of α′-martensite is surrounded by -martensite and the volume fraction of -martensite decreases with the rise of strain rate. The reason is that the deformation at high strain rate is close to a process of adiabatic heating which increases the SFE resulting in a weaker transformation induced plasticity (TRIP) effect. The results also indicate that -martensite and α′-martensite are satisfied with the Burgers relation {(110)_α′// ,[1-11]_α′//}. The α′-martensite and γ phases show the Kurdjumov–Sachs (K–S) orientation relation{(110)_α′//(111)_γ,[1-11]_α′//[1-10]_γ}. The -martensite and γ phases exist with the Shoji–Nishiyama (S–N) orientation relation{(111)_γ//, [1-10]_γ//} [23].

OPEN IN VIEWER

Figure 3

EBSD phase contrast and grain boundary map of tested steels under different strain rates (a) 10⁻³ s⁻¹, (b) 10⁻² s⁻¹, (c) 10⁻¹ s⁻¹ (blue: austenite, yellow: -martensite, red: α′ -martensite, white lines: the K–S orientation relationship, green lines: the S–N orientation relationship, light blue lines: the Burgers orientation relationship).

Dilation curves of specimens after tensile testing to failure were obtained during continuous heating with the heating rate of 5°C s⁻¹ and cooling process with the cooling rate of 20°C s⁻¹ and the results are shown in Figure 4. During the heating process, the temperature of transformation from to γ is different under different strain rates, while α′→γ transformation does not occur. The phase transformation temperatures are determined from these dilation curves. During the cooling process, there is no occurrence of phase transformation. The volume of -martensite decreases with the increase of strain rates. At the smallest strain rate, the increase of temperature produced by plastic deformation is negligible. However, at high strain rate, the temperature increases significantly [9]. Thus, SFE and ΔG^γ→ of tested steel increase, which causes the volume fraction of -martensite decrease at high strain rate. The γ→ transformation temperature of 10⁻¹ s⁻¹ specimen is the highest, because the -martensite is produced at high temperature which shows that the -martensite produced at high strain rate is more stable than at low strain rate.

OPEN IN VIEWER

Figure 4

Dilation curves; (a) 10⁻³ s⁻¹, (b) 10⁻² s⁻¹, (c) 10⁻¹ s⁻¹.

Based on the work reported by Yang et al. [24], the increased SFE reduces the fraction of twinned grains and concurrently increases the thickness of deformation twins. In the tested steel, it is -martensite which causes a decrease in the interfaces between martensite and the matrix, resulting in an overall reduced work-hardening ability.

Figure 5 shows the XRD data of the tested steel deformed at varying strain rates. It indicates that there is -martensite under any strain rate suggesting that the γ→ transformation occurs. There is no α′-martensite in Figure 5 because the volume fraction of α′-martensite is too low for being detected by X-ray diffractometer. Intensities and areas of (111)_γ and (200)_γ decrease, while there is no obvious change in other orientation austenitic grain. Therefore, the TRIP effect occurs mainly in (111)_γ and (200)_γ austenitic grain.

OPEN IN VIEWER

Figure 5

X-ray diffraction results of the tested steel under different strain rates.

In order to observe features presented in the damage zone and identify the main damage mode for all tested steels, the fracture morphologies were examined after tensile testing to failure. Figure 6 shows the fractured surfaces of the tested steels after the tensile test. All specimens exhibit many small dimples, a characteristic of ductile fracture, which is a mark of admirable fracture toughness. The results show that the number of dimples decreases, but the diameter and depth of the dimples become larger with the rise of strain rate. This might be the reason that elongation decreases. With the increase of strain rate, volume fraction of -martensite becomes low for the rising of SFE, so the fine-grain strengthening effect becomes weaker, resulting in a decreased ductility. What is more, cleavage fracture area increases with the increase of strain rate. Thus, the higher the strain rate, the worse the ductility.

OPEN IN VIEWER

Figure 6

SEM surface fractographs of the failure tensile specimens under different strain rates. (a) 10⁻³ s⁻¹, (b) 10⁻² s⁻¹, (c) 10⁻¹ s⁻¹.

Conclusions

SFE and of the tested material at 293K are calculated to be 11.57mJ m⁻² and −75.3J mol⁻¹, respectively, and the strain-induced γ→α′ and γ→ transformation both occur. However, SFE increases with increasing of strain rate. Yield strength is line up straight with the rise of strain rate, but the tensile strength and elongation are opposite. The tested steel has a negative SRS and work-hardening rate sensitivity.