Effects of Cu addition on formability and surface delamination phenomenon in high-strength high-Mn steels

1. Introduction

Automotive steels have required an enhancement of safety and crashworthiness quality as well as an excellent combination of ductility and strength. In particular, structural reinforcement components, such as B-pillar parts, require high strength with sufficient stiffness and formability. Austenitic high-Mn steels are considered as strong candidates to satisfy these requirements due to powerful strengthening mechanisms of transformation induced plasticity (TRIP) and twinning induced plasticity (TWIP) [[1], [2], [3], [4], [5], [6], [7]]. The TWIP steels generally consist of 0.5-1.2 wt.% and 12-30 wt.% of C and Mn contents, respectively [[8], [9], [10], [11], [12]]. Those steels show over 1 GPa tensile strength with excellent ductility (40-60 %). The properties result from deformation twins which hinder the movement of dislocations, so called ‘dynamic Hall-Petch effect’, and suppress the necking due to the high strain-hardening rate. The reduction of C or Mn content from the TWIP steels turns the major deformation mechanism into the TRIP [13,14]. Although limited studies of high-Mn TRIP steels are available yet, they have a great potential to show an excellent strain hardening capacity due to the transformation to hard martensite during the deformation [15].

Cu has been known as an effective element to control the stacking fault energy (SFE) and stability of austenite in the high-Mn steels [[16], [17], [18]]. Lee et al. [16] studied the effect of Cu on mechanical behavior of high-Mn TWIP steels. They reported that kinetics of twin formation was suppressed. Also, the critical strain related to the evident change of work hardening rate was delayed by increasing the Cu content, even though a small Cu addition has only a small influence on the increase in SFE. Peng et al. [17] observed that the occurrence of mechanical twinning was suppressed or delayed with increasing Cu addition in Fe-1.3C-20 Mn TWIP steels. The higher SFE causes the higher critical strain for serrated flows in tensile curves.

In a previous paper regarding the high-Mn TRIP steels, an Fe-0.4C-15Mn-Cu system was suggested to achieve the excellent combination of strength and elongation [18]. In these high-Mn TRIP steels, the addition of 1-2 wt.% Cu increased the stacking fault energy (SFE), and led to the transition of deformation mechanism from TRIP to TWIP. In particular, the 1-wt.%-Cu-containing steel exhibits an apparent improvement of tensile properties due to synergic effects of TRIP and TWIP. The preceding researches regarding effects of Cu addition in high-Mn steels are mainly focused on deformation behavior and tensile properties [[16], [17], [18]]. However, Cu addition effects on formability have not been sufficiently reported, although the formability of high-Mn steels is a critical issue in automotive applications under non-uniform deformation environments caused by dynamic strain aging [19,20].

In the high-Mn TWIP steels, the strain rate sensitivity (m-value) and normal anisotropy (r-value) are known to be critical material parameters in the stretch-flangeability [[21], [22], [23]]. The TWIP steels having low SFEs generally show a relatively low stretch-flangeability because of the absence of post-necking elongation, thereby impeding the localized deformation [22]. However, those researches related to determining factors for the formability of high-Mn TRIP or TRIP-TWIP steels are hardly investigated so far. In the present study, therefore, effects of Cu addition on formability in the austenitic high-Mn TRIP-TWIP steels were investigated. The formability was evaluated by hole-expansion tests and cup-drawing tests, and the surface characteristics were examined. The test results were basically correlated with various mechanical parameters. In a microstructural aspect, the high-Mn steels contain segregation bands of Mn and Cu which lead to the inhomogeneous precipitates and deformation behaviors. Therefore, roles of the segregated microstructures, deformation behavior, and mechanical parameters on the formability and surface characteristics are discussed in detail.

2. Experimental

2.1. Austenitic high-Mn steels

Three austenitic high-Mn steels were fabricated by a vacuum-induction melting method, and their nominal chemical composition was Fe-0.4C-15Mn-(0,1,2)Cu (wt.%). 38-mm-thick plates were reheated for 1 h at 1200 °C, hot-rolled from 1100 to 900 °C, and cold-rolled to produce 1.5-mm-thickness sheets. These sheets were annealed in a continuous annealing line simulator at 750 °C for 1 min. For convenience, the annealed sheets are referred to as ‘0Cu’, ‘1Cu’, and ‘2Cu’ according to the Cu content.

2.2. Microstructural characterization

The detailed microstructures in the 1/4 sheet-thickness area of steel sheets were examined by using X-ray diffraction (XRD), scanning electron microscope (SEM), electron back-scatter diffraction (EBSD), and transmission electron microscopy (TEM). The phase characterization was conducted by using XRD (target; CuKα, step size; 0.02 deg, scan rate; 2 deg min^-1), and the volume fraction of phases was measured by a direct-comparison XRD method [24]. EBSD analysis was conducted by using an FE-SEM (Quanta 3D FEG, FEI Company, USA). EBSD specimens were prepared by the conventional mechanical grinding and subsequent electro-polishing in a solution of 90 %CH₃COOH+10 %HClO₄ at the operation voltage of 30 V. The EBSD operation voltage, measuring step size, and working distance were 25 kV, 0.04 μm, and 12 mm, respectively. The resulting data were interpreted by orientation imaging microscopy (OIM) analysis software. The segregation behavior of Mn and Cu was investigated by using EPMA microprobe (JXA 8530F, JEOL, Japan) at the operation voltage of 15 kV. For the TEM analysis, specimens were prepared by a focused ion beam (FIB), and observed by using a TEM (JEM-2100F, JEOL, Japan) installed with energy dispersive X-ray spectroscopy (EDS) at an accelerating voltage of 200 kV.

2.3. Tensile and formability tests

Dog-bone-shaped plate-type tensile specimens (gage length; 25 mm, width; 5 mm, thickness; 1 mm, longitudinal direction) were tested by using a universal testing machine (8801, Instron, USA, capacity; 100 kN) at a strain rate of 10^-3 s^-1. The tensile tests were conducted three times for reliability of each datum.

The stretch-flangeability was measured by a hole-expansion test (HET) using square-shaped specimens (90×90×1.4 mm). A 10-mm-diameter central hole was made by a drilling process, and then the HET was conducted by using an Erichsen hydraulic universal sheet metal testing machine (145-60, Erichsen, Germany). Each hole-expansion ratio (HER) was measured by pushing a 60°-conical punch up until a main crack was observed in the thickness direction. The hole diameter was measured after the HET, and the HER was calculated by the following equation [25]:

HER(%)=$\frac{ d_{ f } -d_{0}}{ d_{0}}$×100 (1)

where d_f and d_o are final and initial hole diameters, respectively. The HER was presented as the mean value obtained after three tests for each specimen.

Cup-drawing tests of disc-shaped specimens (diameter; 80 mm, thickness; 1.4 mm) were performed by using a universal sheet and strip metal testing machine (USM-50-2, Tokyo Testing Machine Manufacturing, Tokyo) under a load of 19.6 kN. The diameter of extrusion die, extrusion distance, extrusion rate, and drawing ratio were 53.88 mm, 50 mm, 60 mm/min, and 1.6, respectively.

3. Results

3.1. Microstructures of annealed austenitic high-Mn steels

The XRD analysis data of the annealed 0Cu, 1Cu, and 2Cu steels are shown in Fig. 1. Peaks of austenite are present in the three steels, while a very small peak of ε-martensite is found in the 0Cu steel. The volume fractions of austenite (V_γ) calculated by the direct-comparison XRD method [24] are 99.8, 100, and 100 % in the 0Cu, 1Cu, and 2Cu steels, respectively, and are indicated inside Fig. 1.

Fig. 2a-c shows SEM backscattered-electron (BE) images of the 0Cu, 1Cu, and 2Cu steels. All the steels consist of austenite grains. The austenite grain size (D_γ) was measured by an orientation imaging microscopy (OIM) analysis, and the data are also listed at the bottom of each image. In this study, the grain size of the three steels is similar as about 3 μm. In the 1Cu steel, a dark contrast exists along the longitudinal direction (Fig. 2b). When the dark area is observed in detail, a few fine secondary-phase particles are found as indicated by a yellow-dotted box. This dark area with fine particles is more pronounced in the 2Cu steel (Fig. 2c). Fig. 2d presents the EDS line-profile data along the red line of Fig. 2c. The Mn content fluctuates within 11.3-18.0 wt.%, and the Cu distribution follows the trend of Mn distribution. This is because high-Mn steels generally contain the band-shaped solute segregation along the rolling direction [15,[26], [27], [28]]. The segregation is not readily dissolved by the homogenization process due to the sluggish diffusion of substitutional elements. Fine particles are mainly formed along the high-(Mn,Cu) bands (Fig. 2c, d).

Fig. 3a shows a high-magnification BE image in the high-(Mn,Cu) band of the 2Cu steel. Cuboidal particles are densely formed along the high-(Mn,Cu) band. Fig. 3b shows a TEM bright-field (BF) image of cuboidal particles. These particles are identified to be Cu-rich FCC-phase particles according to the selected area diffraction pattern and quantitative EDS analysis data (Fig. 3b).

3.2. Tensile properties and deformation mechanisms

Fig. 4 exhibits engineering stress-strain curves of the 0Cu, 1Cu, and 2Cu steels at room temperature, and measured tensile properties are displayed in the inset of Fig. 4. As the Cu content increases, the yield and tensile strengths decrease, while the total elongation (ε_total) is highest in the 1Cu steel. The 1Cu steel shows a good combination of tensile strength and total elongation (1093 MPa and 65.1 %, respectively). Furthermore, the 2Cu steel shows the quasi-continuous yielding, whereas the 0Cu and 1Cu steels exhibit the continuous yielding behavior. It is well known in Al alloys and Mg alloys having ultra-fine grains that they show a discontinuous yielding behavior because of a lack of mobile dislocations or mechanical twinning [[29], [30], [31]]. The mentioned phenomenon thus might be related to the increased SFE by the Cu addition, implying the decreased twinning activity in the 2Cu steel. However, the origin of quasi-continuous or discontinuous yielding in austenitic TWIP steels having ultra-fine grains should be clarified further [2,31,32]. In addition, it should be noted that serrated flows appear in the three steels during the tensile deformation. The strain for the first serration peak increases from 12.5-22.0 and 28.2 % with increasing Cu content.

Fig. 5a-c shows EBSD image quality (IQ) + phase maps of the 20 %-strained 0Cu, 1Cu, and 2Cu steels. In the 0Cu and 1Cu steels, transformed α’- and ε-martensite are observed along the longitudinal direction (Fig. 5a, b), whereas they are not in the 2Cu steel (Fig. 5c). According to the previous study [18], this band-shaped martensitic transformation originates from the existence of high- and low-(Mn,Cu) bands. In low-(Mn,Cu) bands, the austenite having a low mechanical stability is actively transformed to the α’- or ε-martensite [8,33]. Measured volume fractions of austenite, α’-martensite, and ε-martensite (V_γ, V_α’, and V_ε, respectively) are shown inside each map. The V_α’ and V_ε reach 10.7 and 19.5 %, respectively, in the 0Cu steel, decrease to 3.4 and 10.4 %, respectively, in the 1Cu steel, and disappear in the 2Cu steel. In the 1Cu and 2Cu steels, some sharp parallel lines are found inside austenite grains (arrow marks in Fig. 5b, c), and are identified as deformation twins having misorientation angles of 60° to the matrix.

3.3. Delamination and formability

Fig. 6a-c shows optical photographs of the tensile-fractured specimens. The 0Cu steel specimen shows a generally-deformed flat surface (Fig. 6a). However, a delamination is found in the specimen surface of 1Cu steel (Fig. 6b), and becomes severe in the 2Cu steel (Fig. 6c). This surface delamination occurs mainly along the longitudinal direction, and is deepened into the interior up to about 0.1 mm in the 2Cu steel.

Hole-expansion ratio (HER) values calculated after the hole-expansion test (HET) are listed in Table 1. The 0Cu steel shows the lowest HER (36.67 %) which is somewhat lower than those of conventional high-Mn steels having similar chemical compositions [22]. As the Cu content increases, the HER enhances dramatically up to 58.05 % in the 1Cu steel, and then slightly decreases in the 2Cu steel. The post elongation (ε_post), normal anisotropy ($\bar{r}$), planar anisotropy (ǀǀǀΔrǀ), and strain hardening exponent (n) of the 0Cu, 1Cu, and 2Cu steels are shown in Table 2. The normal ($\bar{r}$) and planar anisotropy (Δr) can be calculated by using the following equations [34]:

$\bar{ r }=\frac{ r_{0}+2r_{45}+r_{90}}{4} $ (2)

|Δr|=|$\frac{ r_{0}-2r_{45}+r_{90}}{2} $| (3)

where r₀, r₄₅, and r₉₀ are plastic strain ratios of loading directions. The n-values are calculated from true stress-strain curves according to ASTM standard E 646-98, as expressed by the following equation [35].

$n=\frac{N\sum\limits_{i=1}^{N}(log\varepsilon_{i}\times log\sigma_{i})-(\sum\limits_{i=1}^{N}log\varepsilon_{i} \times \sum\limits_{i=1}^{N}log\sigma_{i})}{N(log\varepsilon_{i})^{2}-(\sum\limits_{i=1}^{N}log\varepsilon_{i})^{2}} $ (4)

where N, ε, and σ are the number of data pairs, true strain, and true stress, respectively. Total elongation, strain hardening rate (n), and planar anisotropy (|Δr|) tend to follow the trend of HET results in Cu-added high-Mn steels. The relationship between HET results and various mechanical parameters (ε_total, ε_post, |Δr|, $\bar{ r }$, n, and circumferential strain limit (ε_θ)) are visualized in Fig. 7a-d. Detailed calculation of ε_θ will be discussed in the later section.

Fig. 8a-c shows photographs of the cup-drawn specimens of the 0Cu, 1Cu, and 2Cu steels after the drawing test with a drawing ratio of 1.6. The 0Cu steel does not show any earing or wrinkle (Fig. 8a), while the drawing ratio of 1.6 is not satisfied because of the edge cracking. The 1Cu and 2Cu steels do not show any edge cracking, earing, or wrinkle (Fig. 8b, c), while the drawing ratio is satisfied. However, a severe surface delamination occurs on inner and outer cup-specimen surfaces of the 2Cu steel, as indicated by arrows in Fig. 8c.

4. Discussion

4.1. Effects of Cu addition on stretch-flangeability

Since Cu raises the SFE or austenite stability related to deformation mechanisms [[16], [17], [18]], the transition of deformation mechanism occurs as the amount of Cu addition increases. In the previous study [18], the SFE fluctuates in the range of 17.7-24.2, 19.8-27.1, and 21.37-29.21 mJ/m² for the 0Cu, 1Cu, and 2Cu steels, respectively, when the segregation of Mn and Cu is considered. This implies that the change in deformation behavior can be affected significantly by the solute segregation. However, since the volume fraction of 30-70 nm-sized Cu-rich particles is only 0.5 % or less (according to TEM and SEM micrographs) even in the 2Cu steel, their effects on the TRIP or TWIP phenomena is negligible in this study. In particular, the 1Cu steel is a kind of hetero-structural material which simultaneously shows the TRIP in the low-(Mn,Cu)-segregated region and the TWIP in the high-(Mn,Cu)-segregated region. In the 2Cu steel, the TWIP is predominant because the SFEs of both high- and low-(Mn,Cu)-segregated regions (21.37-29.21 mJ/m²) are higher than 20 mJ/m². These results conclude that deformation mechanisms transit from TRIP to TWIP with increasing Cu content.

Effects of Cu addition on formability in high-Mn steels have not been carefully investigated yet, although they have been concretely investigated in stainless steels or multi-phase TRIP steels [[36], [37], [38], [39], [40], [41], [42]]. The basic relationship between the sheet-formability and various mechanical parameters have been well established [22,[36], [37], [38], [39]]. The formability generally increases as the total elongation, post elongation, strain hardening exponent (n), yield strength, and fracture toughness increase [22,[36], [37], [38], [39]]. However, conflicting opinions on the relationship between the formability and mechanical parameters have been raised in high-Mn steels [8,39,40]. Chung et al. [40] reported that TWIP steels having high strain hardening rate and total elongation were deformed much earlier than the predicted onset of diffuse necking for all deformation modes, thereby leading to the failure without the strain localization, which might be a major drawback of the TWIP steels. Chen et al. [22] also reported the relationship between r-value and formability, and explained the low stretch-flangeability of TWIP steels by the absence of localized necking. Several literatures [22,[36], [37], [38], [39], [40]] on formability of high-Mn steels reached the similar conclusion, but most of them described the comparison of the TWIP steels with interstitial-free (IF) steels, instead of TRIP steels or high-Mn steels.

To reveal critical factors of formability, the relationship between the formability and mechanical parameters is discussed in this study. Firstly, the total elongation shows an obvious relationship with the HER results (Fig. 7a). The 0Cu steel shows the lowest total elongation and formability because of the rapid martensitic transformation from the initial strain stage (Fig. 5a). The total elongation is not a sole factor determining formability, but various factors are also complexly related to the formability.

The material showing the high post-elongation generally suppresses the final fracture after the strain localization. However, it is a common characteristic of high-Mn steels that exhibit the relatively low post elongation and sudden fracture after the ultimate tensile strength point [8]. Several researchers mentioned this as one of reasons for the lower formability in high-Mn steels than in IF steels [8,22]. The present 0Cu, 1Cu, and 2Cu steels also show the poor post elongations of 0.23, 1.08, and 3.14 %, respectively, along with a sudden fracture after the necking (Table 2). Furthermore, they do not follow the tendency of HER results (Fig. 7a).

In addition, the strain hardening rate (n) works for delaying the strain localization as well as for enhancing the formability. The n-values are calculated to be 0.341, 0.415, and 0.395 for the 0Cu, 1Cu, and 2Cu steels, respectively, and show a similar tendency with the HER results (Fig. 7c). In the 1Cu steel, synergetic strengthening effects of TRIP and TWIP occurring during the deformation contribute to the highest n and HER, along with the delay of strain localization.

The anisotropy of materials is well known to be a critical factor for determining the stretch-flangeability, and is widely utilized in Mg and Al alloys as well as steels [[43], [44], [45]]. There are several studies on the relationship between normal anisotropy ($\bar{r}$) and formability. Wang et al. [46] reported the relationship, which is expressed by the following equation [46]:

ε_θ=（1+$\bar{r}$）n-ε₀ (5)

where ε_θ, n, and ε₀ are the limit of circumferential strain, strain hardening rate, and pre-strain, respectively. The pre-strain can be negligible because a central hole of an HET specimen is made by the drilling process, instead of the usual punching process. Thus, the limit of circumferential strain increases with increasing $\bar{r}$ and n, and is calculated to 0.71, 0.85, and 0.81 for the 0Cu, 1Cu, and 2Cu steels, respectively. This shows a similar tendency with the HER values (Fig. 7d). The present steels do not exhibit a significant difference in anisotropy along the normal direction, but the difference of n-value determines the tendency of the circumferential strain limit and HET results. In addition, the planar anisotropy (Δr) also follows the tendency of the HET results. However, the Δr of the three fully-recrystallized steels is close to 0, which cannot drive a noticeable difference in formability because of nearly-isotropic austenite grains (Table 2 and Fig. 7b).

On the other hand, a large difference in hardness for each phase can deteriorate the formability in dual- or multi-phase steels. In the 0Cu steel, a relatively large volume fraction of α’- or ε-martensite formed during the deformation can be one of probable reasons for the poor formability (Fig. 5a). Even though the 1Cu steel also exhibits the martensitic transformation, the volume fraction of transformed martensite is somewhat smaller in the 1Cu steel than in the 0Cu steel. In addition, not only the martensitic transformation but also appropriate formation of deformation twins leads to strong strain hardening effects to achieve the highest HER value (Fig. 5b). Therefore, the 1Cu steel exhibits the optimized formability because of the highest total elongation and strain hardening rate and the lowest planar anisotropy.

4.2. Effects of Cu addition on surface delamination

The addition of Cu influences the formation of secondary phases such as Cu-rich particles as well as deformation mechanisms. The presence of the Cu-rich particles has been reported in various steel researches [[47], [48], [49]]. In Cu-bearing HSLA steels, the Cu-rich particles having an FCC crystal structure were defined as ε-Cu phases, and worked for the precipitation strengthening [41]. They were also observed in Cu-alloyed austenitic steels [47,49]. Isheim et al. [49] observed a substantial precipitation hardening effect due to the formation of nanometer-sized Cu-rich precipitates in a 6-wt.%-Cu-alloyed austenitic steel. However, effects of Cu-rich particles on formability have not been investigated, although they are widely reported in steels.

In this study, a small fraction of Cu-rich particles is observed in the Cu-added high-Mn steels. Most of them are found along high-(Mn,Cu) bands, and their volume fraction increases with increasing Cu content (Figs. 2b-d, 3 a). Cuboidal particles of 30-70 nm in size are revealed as a Cu-rich FCC phase by the detailed TEM analysis (Fig. 3b, c). Fig. 8a, b shows SEM secondary-electron (SE) and BE images of a delaminated region on the tensile-fractured specimen of the 2Cu steel. The surface delamination is found to propagate along the longitudinal direction (Fig. 6a, b). Fig. 8c, d shows EPMA Mn- and Cu-distribution maps of the region corresponding to Fig. 9a, b. The mapping results indicate that the surface delamination occurs in the high-(Mn,Cu) band region, and that it is closely related to the formation of Cu-rich FCC-phase particles according to Fig. 2c, d.

The delamination-related fracture has been widely reported in fields of composite materials such as fiber-reinforced composites [50,51]. The segregation of solute atoms and severe texture are regarded as main reasons of the delamination. In this study, however, the solute segregation is not easy to be a primary factor. It is because the band-shaped Mn-segregation is usually formed in conventional high-Mn steels [[26], [27], [28]]. In addition, the surface delamination due to the Mn-segregation has not been reported for those steels yet, and the severe texture has not been found in fully-recrystallized Cu-added high-Mn steels. Thus, it is reasonable to conclude that the Cu-rich FCC-phase particle critically affects the surface delamination (Fig. 9a-d).

Achieving an acceptable guarantee of surface quality is primarily needed after the machining or forming of cold-rolled sheets. The addition of Cu effectively controls the SFE and stability of the austenite to change deformation mechanisms in high-Mn steels, but sometimes causes a serious drawback of surface delamination. Therefore, it should be carefully considered and optimized even though it dramatically enhances mechanical properties.

5. Conclusions

In this study, effects of Cu addition on formability and surface delamination phenomenon in the austenitic high-Mn TRIP-TWIP steels were investigated. Roles of the segregated microstructures, deformation behavior, and mechanical parameters on the formability and surface delamination were discussed.

(1)All the steels consisted of austenitic microstructure of similar grain sizes. The 0Cu, 1Cu, and 2Cu steels contained the Mn-segregation band, and the Cu distribution followed accordingly the trend of Mn distribution. Cuboidal Cu-rich FCC-phase particles were densely formed along the high-(Mn,Cu) band, and their volume fraction increased with increasing Cu content.

(2)As the Cu content increased, the yield and tensile strengths decreased, while the elongation was highest in the 1Cu steel. The major deformation behavior varied in the order of TRIP, TRIP-TWIP, and TWIP, which was affected by the existence of high- and low-(Mn,Cu) bands. The serrated flows appeared in the three steels during the tensile deformation, and the strain for the first serration peak increased from 12.5-22.0 and 28.2 % with increasing Cu content.

(3)The hole-expansion ratio (HER) enhanced dramatically up to 58.05 % in the 1Cu steel, compared to 36.67 % in the 0Cu steel, and then slightly decreased in the 2Cu steel. Among the various mechanical parameters, the total elongation, strain hardening rate, and planar anisotropy tended to follow the trend of the HERs of the Cu-added high-Mn steels.

(4)The Cu addition showed a significant role in enhancing the formability. However, a delamination was found in the specimen surface of the 1Cu steel, and became severe in the 2Cu steel. The Cu-rich particles affected the surface delamination rather than the segregation or texture itself.

Acknowledgments

This work was supported by the Korea Institute for Advancement of Technology (KIAT) grant funded by the Korea Government (MOTIE) (Grant No. P0002019), by the Korea University Grant for the fifth author, and by the Brain Korea 21 PLUS Project for Center for Creative Industrial Materials.

---