Effect of ferrite-martensite interface morphology on bake hardening response of DP590 steel

Materials Science & Engineering A 676 (2016) 463–473

Contents lists available at ScienceDirect

Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea

Effect of ferrite-martensite interface morphology on bake hardening response of DP590 steel Arnab Chakraborty b, Manashi Adhikary c, T. Venugopalan c, Virender Singh b, Tarun Nanda b, B. Ravi Kumar a,n a

CSIR-National Metallurgical Laboratory, Jamshedpur 831007, India Mechanical Engineering Department, Thapar University, Patiala 147004, India c Tata Steel, India b

art ic l e i nf o

a b s t r a c t

Article history: Received 20 May 2016 Received in revised form 2 September 2016 Accepted 2 September 2016 Available online 3 September 2016

The effect of martensite spatial distribution and its interface morphology on the bake hardening characteristics of a dual phase steel was investigated. In one case, typical industrial continuous annealing line parameters were employed to anneal a 67% cold rolled steel to obtain a dual phase microstructure. In the other case, a modified annealing process with changed initial heating rates and peak annealing temperature was employed. The processed specimens were further tensile pre-strained within 1–5% strain range followed by a bake hardening treatment at 170 °C for 20 min. It was observed that industrial continuous annealing line processed specimen showed a peak of about 70 MPa in bake-hardening index at 2% pre-strain level. At higher pre-strain values a gradual drop in bake-hardening index was observed. On the contrary, modified annealing process showed near uniform bake-hardening response at all prestrain levels and a decrease could be noted only above 4% pre-strain. The evolving microstructure at each stage of annealing process and after bake-hardening treatment was studied using field emission scanning electron microscope. The microstructure analysis distinctly revealed differences in martensite spatial distribution and interface morphologies between each annealing processes employed. The modified process showed predominant formation of martensite within the ferrite grains with serrated lath martensite interfaces. This nature of the martensite was considered responsible for the observed improvement in the bake-hardening response. Furthermore, along with improved bake-hardening response negligible loss in tensile ductility was also noted. This behaviour was correlated with delayed micro-crack initiation at martensite interface due to serrated nature. & 2016 Elsevier B.V. All rights reserved.

Keywords: Dual phase steel Bake-hardening Martensite Pre-strain Ferrite/martensite interface

1. Introduction The presence of martensite, a hard second phase within a soft and ductile ferrite matrix in a dual phase steel promotes better distribution of strain. In other words, this results in efficient load transfer between deformable ferrite and relatively non-deformable martensite [1,2]. The combined microstructure creates an excellent blend of strength and ductility. The dual phase (DP) steel offers a characteristic continuous yielding phenomena, high work hardening rate, low yield to tensile strength ratio and high uniform and total elongation [3]. As a consequence, DP steels find extensive use in automotive industries typically for gauge thinning maintaining the same level of strength and stiffness while lowering the overall vehicle weight thus improving the fuel efficiency n

Corresponding author. E-mail address: [email protected] (B. R. Kumar)

http://dx.doi.org/10.1016/j.msea.2016.09.018 0921-5093/& 2016 Elsevier B.V. All rights reserved.

[4]. A major strategy of the automotive industry is to achieve simultaneously optimized combination of properties which include cost, passenger safety and fuel economy through reduction of vehicle weight by using high strength steels [3–5]. DP steels also show excellent bake hardening (BH) properties for which it assumes the role of prime material in automotive industry for the exterior members of a vehicle like door panels and B-Pillars etc [1]. These press formed finished components are passed through a paint curing operation that helps bind the paint coat as well as enhance the final yield strength post forming and painting, thereby, improving the dent resistance at no extra production cost. The BH effect comes into play during paint curing operation which is performed within a chamber held at a temperature of about 170 °C for about 20 min. It is well studied and reported now that this low temperature baking aids in the diffusion of solute carbon atoms present in the microstructure of DP steels to the dislocations generated at (a) ferrite grain boundaries during the phase

464

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

transformation from austenite to martensite during the cooling stage and (b) from the external deformation (pre-straining) while forming of auto-body components [6–8]. The accumulation of these solute interstitial atoms near the generated dislocation sites initiate “Cottrell Atmosphere” formation and simultaneous pinning of the mobile dislocations by these solute atoms begin. The driving force for the solute redistribution emanates from the reduction in lattice energy. Often this stage is referred as first stage of BH and can occur even at room temperature. The second stage of BH is observed under a higher baking temperature (100–250 °C) when carbon precipitates in the form of ϵ carbides around the pinned dislocation sites [9,10]. Prolonged segregation of interstitial solute C atoms on the dislocation sites can lead to an eventual saturation stage when the precipitate concentration induces cluster formation with increasing aging time leading to coarsening of ϵ carbides. This stage is reported to be the third stage of BH [11–13]. Although several authors have discussed the role of second phase (martensite) morphology, its size, and distribution on the mechanical properties of dual phase steels, not much concrete investigation has been reported on its influence on the BH properties [14–16]. The consequence of baking treatment on the final ductility and the mode of deformation mechanism when tensile deformed till fracture with respect to the marteniste spatial distribution in DP steels is another area which has not been examined very distinctly. Kadkhodapoura et al. [4] analysed the failure mechanism of a DP steel material as function of strain, Lai et al. [5] established the fracture mechanism with the influence of martensite volume fraction, Ghadbeigi et al. [17] studied the failure mechanism, its initiation, evolution and growth using in-situ tensile testing experiments. Meanwhile, the second phase development primarily relies on the processing route adopted. It is quite well established that the final microstructure of DP steels depends on lot of factors like: heating rate, soaking temperature (austenitizing temperature), soaking time, cooling medium, and cooling rate [2,18–22]. The relationship between processing factors and final microstructure and hence mechanical properties in DP steel is majorly determined by austenite stability during the annealing process which in turn depends on chemistry of the steel and time given for pearlite dissolution and austenite homogenization with respect to carbon [18]. In the present study an industrially processed cold rolled full hard steel sheet with 67% cold deformation was used. The authors have attempted to work upon few of the several combinations of processing routes which can affect the morphology, fraction and distribution of the martensite islands. Considering different stages of industrial production of DP Steel and the post processing treatments, many of them are found to be associated with a reduction in ductility. BH phenomena is one such finishing process which results in enhanced yield strength of finished component, but often at the cost of its final ductility or total elongation [1]. BH operations in industrial (automotive) processing, is a finishing treatment, where the paint coating of the automotive body is cured/ baked at temperatures of 150–200 °C for 20–30 min [1,20,23,24]. This leads to the diffusion of solute atoms towards the dislocations and the formation of Cottrell atmosphere and hence the strengthening effect [1,3,20,24]. This process of baking is beneficial in ways that it improves the final overall dent resistance without any additional production cost. Industrially, DP steels are produced through continuous annealing route. In DP steel, the BH would also depend on available solute C and its interaction with the dislocations post straining and paint baking operations [1,19,20,23,25,26]. The solute C available would depend on factors like annealing temperature, austenite stability, and amount of C in solution in ferrite during rapid cooling [22]. Effect of BH can be modulated by altering the interaction of solute C atoms with the dislocations. While it has been reported that the

different morphology and distribution of martensite (second phase) is expected to behave differently under stress/ tensile loading with respect to deformation, it is also expected to alter the interaction between available C atoms with the mobile dislocations near the interface and therefore the BH [2,4]. BH response also depends on the amount of pre-strain, bakehardening temperature-time and on the extent of formation of mobile dislocations in ferrite as a result of martensitic transformation. Therefore, contribution of dislocation pinning leading to an increase in bake-hardening response is significant [20]. In this work, authors have made an attempt to study the ferrite/martensite interface size and shape effect along with the martensite spatial distribution on BH and total elongation of two different inter-critically annealed DP590 steel grade. Second phase morphology and distribution was engineered by simulating industrial and lab designed annealing processing conditions. The change in the interfacial area of the second phase was found to impact the BH and tensile behaviour in this steel under the simulated annealing conditions.

2. Experimental procedure 2.1. Material and processing An industrially processed 67% cold rolled steel with chemical composition as stated in Table 1 was used for this study. Flat tensile specimens of 25 mm gauge length as per the ASTM standard E-8 M[11] were machined from the as cold rolled sheets for annealing experiments. Annealing experiments were carried out in a custom designed annealing simulator, having a hot and cold chamber. The annealing atmosphere and cooling medium used was a gas mixture of 10% H2 and balance N2 for all experiments, similar to industrial practice, to prevent oxidation. Two different annealing cycles were used to study their effect on microstructure and consequently on the bake-hardening characteristics. First annealing cycle used was similar to industrial continuous annealing process line (CAPL) process, which comprised (a) fast heating to inter-critical annealing temperature 790 °C, (b) subjecting to short isothermal annealing treatment, (c) slow cooling to 675 °C, (d) rapid cooling to 400 °C and (e) short aging treatment at 275 °C before fast cooling to room temperature. The second annealing cycle was employed to produce DP microstructure with almost similar martensite fraction and tensile properties but with different morphology and distribution of martensite. Annealing cycle comprised of three different annealing stages in the inter-critical annealing temperatures with interrupted heating rates at 710 °C, 790 °C and 840 °C. Specimens were continuously heated to peak annealing temperature at different heating rates without any isothermal annealing stage. The higher average heating temperature compared to CAPL was expected to influence the austenite nucleation and growth. Thus continuous heating annealing process (CHAP) was expected to affect formation of martensite morphology upon rapid cooling. 2.2. Material characterization Both optical and electron microscope techniques were used for the microstructure analysis. Specimens were prepared for microstructure studies using standard metallographic technique. An etchant of 2% Nital was used for revealing microstructure for microscopy studies [5]. Quantitative microstructure analysis for determining grain size distribution and phase fraction of second phase was done by using image analysis software such as “Analysis 5” and “Image J” respectively. X-ray diffraction (XRD) technique using Bruker AXS D8 Discover diffractometer with Cu Kα of

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

Table 1 Chemical composition of given steel.

465

(a)

800

o

Mn

Si

S

P

Al

N

Fe

% wt.

0.074

1.83

0.43

0.002

0.012

0.026

0.0032

Balance

0.154 nm wavelength along with fast detector Lynx Eye was used for diffraction experiments. Nano indentation hardness measurements were made using MTS Nano Indentor XP.

o

790 C /60s

700

675 C

o

710 C

600 o

C

Temperature, C

Element

500 o

400 C

400 o

o

10 C/s

300

275 C

200

2.3. Tensile property evaluation 100

Tensile tests were conducted at room temperature under displacement control at a strain rate of 1.3
10  4 s  1 using Instron 8862 system of 100kN load capacity. To study the BH effect, initial pre-strains fixed at various levels (5.0% max) were subjected to the tensile specimens followed by paint curing simulation leading to bake-hardening phenomena. Thereafter, tensile properties were determined.

0 0

25

50

75

100

125

150

175

200

225

Time, s

(b)

800

o

840 C o

1.5 C/s

700

o

1 C/s

o

790 C

o

675 C

o

BH is a strain aging phenomena which helps in improving the yield strength of a pre-strained or formed component [10,20]. Migration of solute atoms to the mobile dislocations present in microstructure causes pinning effect or creation of “Cottrell Atmosphere” resulting in increase of yield strength [22,23]. Bakehardening effect was studied to a maximum pre-strain of 5% in this study. After pre-straining, specimens were bake-hardened by annealing at 170 °C for 20 min.

600 Temperature, oC

2.4. Bake-hardening

710 C

500 o

400 C

400 300

o

275 C

o

6 C/s 200 100 0 0

25

50

75

100

125

150

175

200

225

Time, S

3. Results and discussion

Fig. 1. Temperature-time profile of two different annealing processes (a) CAPL process and (b) CHAP is shown.

3.1. Effect of annealing process on microstructure and properties Bake-hardening characteristics were studied for specimens processed under two different annealing cycles. Fig. 1(a and b) shows CAPL and CHAP annealing cycle used in the experiments to produce DP microstructure in a DP590 grade steel. Microstructures of both specimens obtained under CAPL and CHAP cycles are shown in Fig. 2(a and b). A subtle difference in martensite formation sites in the microstructure was observed, as indicated by arrows in figure. CAPL process dominantly produced martensite at ferrite grain boundaries, whereas, in case of CHAP, martensite formation tendency was more within the ferrite grains or “in-grain” ferrite sites. Microstructure phase quantification for martensite phase fraction revealed about 16.5% and 18% fraction for CAPL and CHAP processes respectively, which is similar to DP590 grade steel produced industrially. Further, grain size distribution of both the phases, ferrite and martensite, is shown in Fig. 3(a,b). Martensite size distribution, irrespective of shape, in CHAP showed relatively large fraction of fine martensite in class range of 0–0.5 μm2 as compared to CAPL. For the purpose of better resolution of martensite, several SEM microstructures at high magnification were used for determining size fraction of martensite. Ferrite grain size distribution, for two processing routes, showed a random distribution nature, which is generally observed for annealed microstructures [6]. Further nano-indentation hardness was conducted to evaluate the second phase (martensite) hardness for both CAPL and CHAP processed specimens. Hardness variation with depth is given in Fig. 4. The average hardness of both lath and other martensite morphology observed in CAPL and CHAP were 350 HV and 387 HV

respectively. Slightly higher hardness of martensite phase produced by CHAP was noticed. Probably, this was due to presence of more carbon in the martensite phase. Tensile properties were evaluated for both the microstructures. Desired UTS of about 590 MPa was obtained under both processing conditions. Additionally, a continuous yielding behaviour along with high work hardening and low yield to tensile strength ratio was also observed (Fig. 5). A minor increase in total elongation was observed for CHAP. 3.2. Effect of aging/bake hardening treatment Room temperature stability or aging resistance of both specimens subjected to two different cycles was examined by aging treatment. In other words, steel was subjected to bake-hardening treatment in the as processed state. This is considered a very important parameter for steels, because in case of lack of room temperature aging resistance, steel will exhibit undesirable stretcher strains during forming operation. Even after bake-hardening treatment, not much change in the tensile properties was noted as shown in Fig. 6. Tensile strength of the steel was observed to increase by a small amount of about 10 MPa. There was slight corresponding loss in ductility. The slight change in yield strength and little or no appearance of yield point elongation behaviour reflects the negligible BH response of as processed microstructure state. The slow cooling rate below 400 °C or aging during annealing cycle seems to have provided effectively stress relieving action of matrix and strain field around martensite and helped in

466

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

(a)

(b)

Fig. 2. SEM microstructure of (a) CAPL processed and (b) CHAP processed DP590 steel. Arrows indicate grain boundary as well as in-grain martensite.

30

60 0

CHAP CAL

CHAP CAL

50 0

20 Frequency, %

40 0 Frequency, %

Size distribution off Ferrite

25

Size distribution of martensite

30 0 20 0

15 10 5

10 0 0

0 (0-0.5)

(0.5-1)

(1-1.5 5)

(1.5-2)

(2-2.5)

(2.5-3)

(3-8)

(<8)

(1-2)

2

(2-3)

(3-4)

(4-8)

(8-16) (16-32) (32 2-64) (64-100) (<100) 2

Class C Interval, μm

C Class Interval, μm

Fig. 3. Grain size distribution of martensite and ferrite in specimens processed by CAPL and CHAP route.

5500

5500

CAPL p Martensite phase

5000

5000

4500

4500

4000

4000

3500

Hardness, HV

Hardness (HV)

CHAP e Martensite phase

(

3000 2500 2000

3500 3000 2500 2000

1500

1500

1000

1000

500

500

0

0 0

50

100

Displacement, nm n

150

0

50

100 Displacement, nm

1 150

Fig. 4. Nano-indentation hardness vs depth plots of martensite phase.

reducing available dislocation density for pinning effect [24]. In summary, it was possible to produce DP 590 grade steel comparable to industrially processed same steel grade by simulating CAPL conditions in the custom designed annealing simulator. Further, CHAP simulation also produced steel with

comparable tensile properties. In fact, it showed a shade better properties than that of CAPL. It is important to mention here that the second hard phase, ‘martensite’ showed different characteristics, with reference to its nucleation sites and size distribution in case of CHAP cycle.

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

700

700 CHAP

600 Engineering Stress, MPa

600 Engineering Stress, MPa

467

500 CAPL

400 300 200 100

5%

500

3%

1%

2%

400 300 200 100

0

CAPL Processed DP 590

0

4

8

12

16

20

24

28

32

0 0

Engineering Strain, %

15

20

25

Engineering Stress, MPa

500

CAPL 400

After Aging Treatment

200

35

5%

600

CHAP 600

30

700

700

Engineering Stress, MPa

10

Elongation,%

Fig. 5. Engineering tensile stress-strain curves of CAPL and CHAP processes showing nearly same properties.

300

5

3% 1%

500 400 300 200 100

100

CHAP Processed DP590

0

0 0

5

10

15

20

25

30

35

Engineering Strain, %

0

5

10

15

20

25

30

35

Elongation,%

Fig. 6. Engineering stress-strain curves for CAPL and CHAP process after aging treatment.

Fig. 7. Engineering stress-strain tensile curve CAPL and CHAP processed steel after pre-straining and bake-hardening treatment.

3.3. Effect of bake-hardening treatment to maximum Pre-straining of 5.0% max

Flow stress is the value of stress at unloading point of the specimen.

Pre-straining at different strain levels from 1% to 5% followed by bake-hardening (BH) treatment led to increase in yield strength along with appearance of yield behaviour (discontinuous yielding). This also affected ductility which showed drastic decrease at higher pre-strain levels. Fig. 7 shows typical tensile stress-strain curves after bake-hardening treatment. Appearance of discontinuous yield behaviour can be noted after bake-hardening treatment. At small pre-strain values to about 2%, yield strength was noted to increase without any significant loss in ductility for CAPL which was followed by a continuous decrease. With prestraining beyond 2%, it resulted in loss of ductility. On the contrary, CHAP did not show much loss in ductility even at high pre-strain values (3–5%), while other tensile properties were found similar to that of CAPL. Bake-hardening treatment is given to automotive grade steels to enhance yield strength after pre-straining. The increase in yield strength after bake-hardening of pre-strained steel depends on mobile dislocation density (newly created due to pre-straining) and available solute carbon in steel. Baking process helps in diffusion of the solute carbon to dislocations and pin newly created dislocations. Pinned dislocations help in the increase of yield strength of bake-hardened steel. The magnitude of increase in yield strength due to BH effect depends on the pre-strain value and the presence of available solute carbon. BH index (BHI) is determined according to Eq. (1) for any given pre-strain condition.

BHI=Lower yield point ( after baking) – Flow stress ( at pre−strain level) (1)

The results of BHI are plotted against pre-strain and shown in Fig. 8. CAPL showed BHI peak at 70 MPa for 2% pre-strain indicating a very promising BH effect. However, it showed a dip in BH effect to below 40 MPa at pre-strain of 3% and continued to drop to about 35 MPa at maximum pre-strain of 5%. In case of CHAP, even at 1% pre-stain the BHI reached to about 60 MPa and achieved peak value of 72 MPa at 4% pre-strain. Further, prestraining to 5% it dropped to 52 MPa. Unlike CAPL, nearly stable BHI were observed in CHAP till 4% pre-strain. BHI for CHAP process always remained above 50 MPa whereas, CAPL showed drop in BHI values to below 40 MPa at higher pre-strains. Further, a comparison of ductility and BHI between the two processes showed a contrast. Beyond peak BHI-pre strain combination, CAPL showed decreasing trend of % elongation to values as low as 10%. This was a very sharp decrease with respect to the as processed state ductility. After bake-hardening treatment there was no significant change in tensile properties as noticed in CHAP steel specimens, unlike CAPL process. Tensile properties were observed to be comparable after bake-hardening treatment even with as processed state values. The interesting feature observed by CHAP was that there was considerable increase seen in yield strength, while the loss in ductility was far lesser than that was observed for CAPL. CHAP processed specimens continued to retain % elongation even

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

100

36

90

32

80 28 70 CHAP

60 50 40

CAPL

30

Total Elongation, %

Bake Hardening Index, MPa

468

CHAP

24 20 16 12

20

CAPL

8

10 0

4 0

1

2

3

4

5

0

1

2

Pre-Strain,%

3

4

5

Pre-Strain,%

Fig. 8. Comparison of BHI and total elongation at different pre-strain values.

250 Open Symbol: CHAP Solid Symbol : CAPL

225 200 175

Work Hardening

150 Δ, MPa

at maximum 5% pre-strain. This was a quite contrasting behaviour for steel processed by CHAP with respect to CAPL. Results of this study indicate that bake-hardening characteristics could be improved by process variation, which in turn induces subtle microstructure changes. CAPL processed steel very distinctly showed drop in bake-hardening properties with increase in pre-strain %. This is reported by several researchers [27]. However, even at high pre-strain % bake-hardening properties could be improved by varying the annealing process parameters (CHAP). Interestingly the ductility was also retained even at 5% pre-strain. It is clearly evident that CHAP processed specimens showed better bake-hardening response along with higher total elongation as compared to CAPL specimens at higher pre-strain values. The minor variation in second phase distribution and morphology thus resulted in a profound effect on the overall BH and tensile properties of CHAP processed DP steel. Maximum achievable BH Index depends on available free carbon that would pin mobile dislocations. In case of ultra low carbon steels, the amount of free C needed for desired BH effect has been investigated by many authors [28,29]. To obtain BHI in the range of 30–60 MPa a minimum of 15–25 ppm of free carbon is reported to be necessary in ferrite [28]. Chiang [29] predicted about 15– 20 ppm of free C needed for a minimum BH effect of 30 MPa. For multiphase steels apart from solute carbon, defect concentration or pre-strain (PS) percentage is also reported to be an important parameter in the evaluation of BH effect. Several authors have highlighted the role of pre-straining on BH [30–36]. While Krieger [35] reported a BHI peak of 60 MPa at 1% PS which further increased with higher pre-straining values, Bruhl [31] observed a maximum BHI of 100 MPa at a PS of about 2% along with a drop in BHI value at higher baking temperature and time due to overaging phenomena. To understand the response of steel due to work hardening (WH) and BH, WH was determined. This measurement was done by determining the difference between the flow stress at unloading (at different pre-strain levels) and the 0.2% offset yield strength. In the present study, continuous increase in work hardening was observed for steel processed by CAPL and CHAP route as can be seen in Fig. 9. However, BH effect did not show trend similar to work hardening. After an initial constant BH effect, a continuous drop in BHI was observed at pre-strains above 2% values for CAPL, whereas constant trend was noted for CHAP even till 4% pre-strain. This indicated more efficient dislocation pinning action by carbon in CHAP. This also indicates insufficient solute carbon in ferrite/martensite phase in CAPL that makes ineffective

125 100 75 50 Bake Hardening

25 0 1

2

3

4

5

Pre-Strain, % Fig. 9. Work hardening and bake hardening response at various pre-strain levels.

pinning above 2% pre-strain wherein number of mobile dislocations increase with higher pre-straining values. X-ray diffraction studies very clearly indicated the availability of solute carbon even after BH of 5% pre-strained state of CHAP specimens. Fig. 10 shows high resolution x-ray diffractogram depicting the split of {112} and {211} peak doublet of martensite in the as processed state, this is indicated by arrows in the figure. The split in the peak doublet measures tetragonality of the martensite phase, which increases with increasing octahedral lattice site occupancy of carbon atoms in martensite lattice. This doublet was not found for CAPL after BH of 5% pre-strained specimen indicating absence of solute carbon. On the other hand, a subtle peak doublet continues to be present in CHAP, indicating surplus solute carbon. Presence of high solute carbon content in CHAP may be attributed to high annealing temperature of 840 °C, which allows sufficient time for carbon dissolution in the austenite phase. Similar to our observation in CAPL, Ramazani et al. [20] made observations on the critical parameters affecting BH kinetics for DP600 grade steel, wherein an increase in either baking temperature or time along with pre-straining (2–5%) caused overaging in these steels. Furthermore, on increasing BHI with higher pre-strain, Waterschoot et al. [37] reported increase in BHI values at higher PS values in case of CMnCrMo dual phase steel (0.08% C) due to the release of C from martensite at temperature around 170 °C due to (over) aging.

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

7000

469

6000

CHAP

CAPL 6000

5000

5000 4000

cps

cps

4000 3000

3000 2000

2000

As processed

As processed 1000

1000 0 81.50 0

5% Pre-strain +BH H 81.75

82.00

82.25

82.50

82.75

83.00

Two Theta

0 81.50

5% Pre-strain + +BH 81.75

82.00

82.25

82 2.50

82.75

83.0 00

Two Theta

Fig. 10. High resolution x-ray diffractogram of {112} and {211} peak doublet of martensite phase in the as processed state and 5% pre-strained and bake hardened state. Arrows indicate doublets.

The effect of second phase morphology on retention of total elongation after bake-hardening treatment has been addressed in more detail and discussed in the ensuing section. 3.4. Role of second phase microstructure morphology and distribution The CAPL cycle offered a mix of fine to medium sized second phase (martensite), mostly at/near grain boundaries, while CHAP cycle showed presence of finer martensite within ferrite grains/ “in-grain” martensite. Thus CHAP seems to offer a better strain partitioning effect between the soft and hard microstructure constituents, which render good strength and ductility combination. To investigate this further an attempt was made to understand the strain partitioning response of ferrite/martensite interface. High resolution microscopy studies were conducted to elucidate the differences at interface. Interestingly, a marked variation in the martensite lath interface could be observed. CAPL processed specimen showed lath interfaces, in general, as sharp, and on the contrary, CHAP processed specimens lath interfaces were serrated as can be seen in Fig. 11. The plausible cause of serrated edges observed in CHAP has been explained in detail in another manuscript which is under publication process. From Fig. 11 it is quite evident that the serrated interfaces of lath martensite in CHAP processed specimen offered better strain

partitioning response and more resistance to deformation thereby delaying the micro-void/crack initiation at the interface which also explains better elongation observed at higher pre-strain values. To understand effect of interface morphology and its behaviour during tensile deformation, tensile fractured specimens were studied. Specimens at 5% pre-strain and bake-hardened condition were considered because of poor ductility of CAPL. CAPL offered uniform distribution of second phase along the ferrite grain boundaries, while, CHAP offered a more random distribution of second phase both along ferrite grain boundary as-well as within the ferrite grains. Near tensile necking region, severe local strain concentration at ferrite/martensite interface was observed. Further, microstructure studies near fracture tip with martensite present at grain boundary showed extensive micro-crack tendency at interface for CAPL cycle, whereas, CHAP showed micro void nucleation at the “in-grain” ferrite/martensite interfaces. Microstructure revealing interface deformation characteristic near fracture tip at very high magnification is shown in Fig. 12. Moreover, irrespective of processing cycle, grain boundary martensite showed early fracture, as can be seen in Fig. 12. However, severity of micro-crack initiation was observed at sharp ferrite/martensite interfaces in CAPL cycle, independent of its location. Further, shape of hard martensite phase was found to influence severity of micro-cracking. Fine lath martensite at grain boundaries as shown in Fig. 11(a) was more sensitive to interface strain localisation compared to oblong or polygonal shaped martensite as in

Fig. 11. Microstructure showing marteniste/interface morphology (a) CAPL produces finer and sharper/smooth ferrite-lath martensite interface; (b) CHAP morphology with finer and serrated laths.

470

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

Fig. 12. Micro-crack formation /micro-void formation in 5% pre-strained and bake hardened specimens. (a) CAPL process microstructure showing extensive micro-cracking of ferrite-martensite interface and (b) CHAP processed microstructure shows interface micro-void nucleation. White arrows indicate various cracks and black arrow indicate the tensile loading direction.

Fig. 11(b) at ferrite grain boundaries. Four major factors are considered important for determining bake-hardening behaviour of dual phase steels. These are: (a) amount of free carbon in solution, (b) presence of dislocation substructure in ferrite phase, (c) kinetics of strain ageing behaviour of soft and hard phases and (d) strain partitioning between soft and hard phases [27]. In the present study, bake-hardening response of two steels was similar to the previous available knowledge in case of CAPL. Therefore, strain partitioning characteristics between soft and hard phases appeared to be responsible for the loss of BH response with increase in pre-strain along with ductility [1,10,24]. Micro-cracking is a well observed phenomenon in DP steels [38–42]. It may occur via martensite particle fracture or martensite-ferrite interface rupture, depending on volume fraction of martensite [4]. The transition from interface controlled damage mechanism to martensite particle fracture increases with increasing volume fraction of martensite [42–44]. CAPL processed steel showed high tendency towards microcracking at or near martensite laths at the ferrite grain boundaries in this study. It is reported that in ferrite, void nucleation is identified with the presence of hard phases viz inclusions and carbides and interface strength varies with type of second phase particle [44,45]. From Fig. 11(a) it may be noted that micro-crack/ ferrite-martensite interface delaminating propensity was high at fine and long martensite present at the grain boundary. The presence of hard phase at the grain boundaries is reported to weaken the grain boundaries by void nucleation causing early interface delamination [46,47]. This may limit the maximum limit to which steel can be pre-strained for BH response as observed in this study. 3.5. Interrupted tensile test Interrupted tensile test was carried out to study critical truestrain required for initiation of interface micro-crack or void nucleation. Tensile deformation test of bake-hardened specimen with 5% pre-strain was carried out till the onset of necking. Before tensile deformation, specimens were marked along gauge length at 2 mm interval and width of the marked location was measured. After interruption of the tensile test, final width was measured at the same location to determine true-strains. Onset of necking was observed at similar true strain for the two processed materials. CAPL processed specimen showed necking at about 9% elongation

whereas it was about 20% in case of CHAP processed material, indicating delayed necking in case of latter. Tensile deformed gauge section was cut from the specimen and examined for deformation behaviour at various true-strain levels. The results of microstructure analysis are shown in Fig. 13. It shows evolution of deforming microstructure at increasing true-strain levels. An early micro-crack initiation was noted at grain boundary martensite/ ferrite interfaces in specimen processed by CAPL, resulting in loss of ductility. CHAP processed specimen did not show any microcracking evidence even in the neck region. However, very limited micro-void nucleation at the martensite/ferrite interfaces was noted. The absence of interface cracking implied stronger interfaces thus resulting in high ductility in CHAP processed specimens. Important to note here was that onset of necking to the same extent as in CAPL was seen, though at much higher strain levels. Loss in strain hardenabilty in the absence of micro-crack called for further investigation. To understand this behaviour, plastic deformation characteristics of ferrite grains were studied. Microstructure very distinctly showed extensive plastic deformation in the two processed steels. To distinguish the extent of plastic deformation, ferrite grain elongation in the tensile deformation direction was determined by estimating aspect ratio distribution. Aspect ratio distribution for the two specimens is shown in Fig. 14. The CHAP shows exclusive peak near small values indicating that nearly all the ferrite grains underwent plastic deformation. On the contrary, CAPL showed bimodal distribution in aspect ratio values. First one at nearly same aspect ratio values as in case of CHAP and second one at higher values. This indicates that necking in CAPL occurred even before the plastic strain was imparted to the more ductile ferrite phase than harder martensite phase. It shows inefficient plastic strain partitioning between ferrite and martensite phases or presence of hard phase at the grain boundaries resulting in strain localisation at the grain boundaries causing loss of ductility. On the other hand, CHAP cycle exhibited optimum strain partitioning between ferrite and martensite phases. The onset of necking in this case was a synergistic effect of loss in ferrite phase strain harden-ability and micro-void nucleation.

4. Conclusion Present study shows the possibility of improving bake-

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

471

Fig. 13. SEM microstructure of CAPL and CHAP samples showing micro-void/crack initiation and grain boundary shearing with increasing true strain (a–c) and (d–f) respectively. (a) Initiation of boundary shear; (d) very stable boundary; (b) Initiation of micro-void at several places and grain boundary shearing in CAPL; (e) scattered micro-void nucleation witnessed in CHAP; (c) extensive local deformation and shear along grain boundaries; (f) relatively stable boundaries and extensive ferrite grains deformation with few micro-cracks. Black arrows indicate various defect zones and white arrow indicate the tensile loading direction.

472

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

50 A.R= Width/Length

National Metallurgical Laboratory, Jamshedpur, for supporting this work.

40 Frequency, %

References 30 CAPL

20

10 CHAP Maximum deformed to minimum

0 0-0.2

0.2-0.3

0.3-0.4

0.4-0.5

0.5-0.6

>0.6

Aspect Ratio (A.R) Fig. 14. Ferrite grain shape aspect ratio distribution after post tensile deformation necking of CAPL and CHAP cycled specimens showing high deformation to low deformation of ferrite grains.

hardening response via modification of current industrial continuous annealing process line parameters. With minor modification in the annealing parameters it is possible to improve bakehardening response of the steel. Following things emerged from this study:

 Bake hardening index showed maximum increase of 70 MPa at









2% pre-strain for CAPL processed DP steel beyond which strength was observed to fall 35 MPa at maximum pre-strain of 5%. On the other hand, CHAP processed specimen was observed with a peak in BH at 4% PS to about 72 MPa and did not show any sharp drop even at 5% pre-strain. Most importantly it did not show loss in its ductility even at 5% pre-strain. Both processes showed increasing work hardenability with prestrain, however, bake hardening behaviour was different for the two cases. It showed nearly constant values for CHAP, whereas, a decreasing trend was observed for CAPL at higher pre-strains. This characteristic was correlated with different contents of available solute carbon in the two processes. Nano-Indentation test too confirmed higher hardness of CHAP processed martensite phase indicating presence of higher carbon. The grain boundary distribution of second phase martensite in CAPL microstructure showed high amount of local deformation along the ferrite-martensite interface boundary by means of interfacial crack growth and propagation while “in-grain” ferrite martensite lath showed relatively less strain concentration at boundaries. Therefore, martensite presence at the ferrite grain boundaries seems to increase the propensity of early microcrack formation leading to loss in ductility. The lath structures of CAPL and CHAP were strikingly different, while CAPL showed sharp/smooth interface with ferrite CHAP produced serrated interface. This difference in interface also appears to influence the fracture or delamination of ferritemartensite boundaries. Between CAPL and CHAP the strain distribution behaviour varied. Onset of necking in CHAP case was a synergistic effect of loss in ferrite phase strain hardenability and micro-void nucleation.

Acknowledgments Authors wish to record their gratitude to Director, CSIR-

[1] J. Zhang, R. Fu, R. Liu, X. Wei, L. Li, Bake hardening behaviour of DP and TRIP steels, J. Univ. Sci. Technol. B 15 (2008) 132–137. [2] H. Seyedrezain, A.K. Pilkey, J.D. Boyd, Effect of pre-IC annealing treatments on the final microstructure and work hardening behavior of a dual-phase steel, Mater. Sci. Eng. A 594 (2014) 178–188. [3] Y. Cao, J. Ahlström, B. Karlsson, The influence of temperatures and strain rates on the mechanical behavior of dual phase steel in different conditions, J. Materress Technol. 4 (2015) 68–74. [4] J. Kadkhodapoura, A. Butzb, S. Ziaei-Rad, Mechanisms of void formation during tensile testing in a commercial dual-phase steel, Acta Mater. 59 (2011) 2575–2588. [5] Q. Lai, O. Bouaziz, M. Gouné, L. Brassart, M. Verdier, G. Parry, A. Perlade, Y. Bréchet, T. Pardoen, Damage and fracture of dual-phase steels: Influence of martensite volume fraction, Mater. Sci. Eng. A 646 (2015) 322–331. [6] P. Chang, Temper ageing of continuously annealed low carbon dual phase steel, Met. Trans. A 15 (1984) 73–86. [7] R.G. Davies, Early stages of yielding and strain ageing of a vanadium containing dual phase steel, Met. Trans. A 10 (1979) 1549–1555. [8] T. Waterschoot, A.K. De, S. Vandeputte, B.C. De Cooman, Static strain aging phenomena in cold-rolled dual-phase steels, Met. Trans. A. 34 (2003) 781–791. [9] A.K. De, S. Vandeputte, B.C. De Cooman, Static strain aging behavior of ultra low carbon bake hardening steel, Scr. Mater. 41 (1999) 831–837. [10] S. Das, Bake Hardening in Low and Medium Carbon Steels (PhD Thesis), IIT Kharagpur, 2012. [11] I.P. Kemp, G. Pollard, A.N. Bramley, Static strain aging in high-carbon steel wire, Mater. Sci. Technol. 6 (1990) 331–337. [12] P. Watte, J. Van Humbeeck, E. Aernoudt, I. Lefever, Strain ageing in heavily drawn eutectoid steel wires, Scr. Mater. 34 (1996) 89–95. [13] Y. Yamada, Static Strain Aging of Eutectoid Carbon-Steel Wires, T Iron Steel I Jpn 16 (8) (1976) 417–426. [14] A. Ghaheri, A. Shafyei, M. Honarmand, Effects of inter-critical temperatures on martensite morphology, volume fraction and mechanical properties of dual phase steels obtained from direct and continuous annealing cycles, Mater. Des. 62 (2014) 305–319. [15] A. Bag, K.K. Ray, E.S. Dwarkadasa, Influence of martensite content and morphology on tensile and impact properties of high-martensite dual phase steels, Met. Mater. Trans. 30 (1999) 1193–1202. [16] S. Kim, S. Lee, Effects of martensite morphology and volume fraction on quasistatic and dynamic deformation behavior of dual phase steels, Met. Mater. Trans. 31 (2000) 1753–1760. [17] H. Ghadbeigi, C. Pinna, S. Celotto, Failure mechanism in DP600 steel: initiation, evolution and fractuRe, Mater. Sci. Eng. A 588 (2013) 420–431. [18] P. Li, J. Li, Q. Mengb, W. Hua, D. Xu, Effect of heating rate on ferrite recrystallization and austenite formation of cold-roll dual phase steel, J. Alloy Compd. 578 (2013) 320–327. [19] S. Gunduz, Static strain ageing behaviour of dual phase steels, Mater. Sci. Eng. A 486 (2008) 63–71. [20] A. Ramazani, S. Bruehla, T. Gerber, W. Bleck, U. Prahl, Quantification of bake hardening effect in DP600 and TRIP700 steels, Mater. Des. 57 (2014) 479–486. [21] A.P. Pierman, O. Bouazizb, T. Pardoena, P.J. Jacques, L. Brassart, The influence of microstructure and composition on the plastic behaviour of dual-phase steels, Acta Mater. 73 (2014) 298–311. [22] A. Momeni, K. Dehghani, S. Abbasi, M. Torkan, Bake Hardening of a Low Carbon Steel for Automotive Applications, Metalurgija. UDC:669.14.018.456157.846.217.8 ¼ 20. [23] F. Maresca, V.G. Kuznetsova, M.G.D. Geers, Deformation behaviour of lath martensite in multi-phase steels, Scr. Mater. 110 (2016) 74–77. [24] C.F. Kuang, J. Li, S.G. Zhang, J. Wang, H.F. Liu, A.A. Volinsky, Effects of quenching and tempering on the microstructure and bake hardening behavior of ferrite and dual phase steels, Mater. Sci. Eng. A 613 (2014) 178–183. [25] E.V. Nesterovaa, S. Bouvierb, B. Bacroix, Microstructure evolution and mechanical behavior of a high strength dual-phase steel under monotonic loading, Mater. Charact. 100 (2015) 152–162. [26] R.R. Queiroz, F.G. Cunha, B.M. Gunzalez, Study of dynamic strain aging in dual phase steels, Mater. Sci. Eng. A 543 (2012) 84–87. [27] I.B. Timokhina, P.D. Hodgson, E.V. Pereloma, Transmission electron microscopy characterization of the bake hardening behavior of transformation induced plasticity steels and dual phase steels, Met. Trans. A 38 (2007) 2442–2454. [28] N. Mizui, A. Okamoto, Recent Development in Bake hardenable Sheet Steels for Automotive Body panels. Steel in Motor Vehicle Manufacture, Int. Conf., Würzburg (1990) 85–94. [29] L.J. Chiang, K.C. Yang, I.C. Hsiao, Effects of Annealing Conditions on Bake Hardenability for ULC Steels, China Steel Tech. Rep. 24 (2011), 1( ̶ 6). [30] W. Bleck, A. Frehn, J. Ohlert, H.M. Sonne, G. Steinbeck, Influence oftemperature and pre-straining on the plastic material behaviour of modern sheet steels for autobody applications, Mater. Werkst. 35 (2004) 495–504. [31] S. Brühl, Einfluss der Martensitephase auf das Bake-Hardening-Verhalten vonDualphasen-Stählen (pH.D. Thesis), RWTH Aachen, Germany, 2010.

A. Chakraborty et al. / Materials Science & Engineering A 676 (2016) 463–473

[32] W.C. Jeong, Effect of prestrain on aging and bake hardening of cold-rolled, continuously annealed steel sheets, Met. Mater. Trans. A 29 (1998) 463–467. [33] M. Krieger, Mechanical Properties and Bake Hardening Behaviour of Cold Rolled Dual Phase Steels Subjected to Advanced Galvanising Process Routes (pH.D. Thesis), Clausthal University of Technology, 2007. [34] T. Waterschoot, Fundamentals of Aging In Multiphase Steels: Static Strain Aging and Tempering in Dual Phase Steels (pH.D Thesis), University of Ghent, Ghent, Belgium, 2003. [35] M. Krieger, Bake-Hardening bei kaltgewalzten Mehrphasenstählen, Diploma Thesis, Clausthal University of Technology, 2003. [36] C.F. Kuang, S. Zhang, J. Li, J. Wang, H. Liu, Effects of pre-strain and baking parameters on the microstructure and bake hardening properties of dual phase steel, Int J. Min. Met Mater. 21 (2014) 766–771. [37] T. Waterschoot, K. Verbeken, B.C. De Cooman, Tempering kinetics of the martensitic phase in DP steel, ISIJ Int. 46 (2006) 138–146. [38] K. Sugimoto, M. Kobayashi, H. Matsushima, S. Hashimot, X-ray residual stress and strain induced transformation of retained austenite in TRIP aided dual phase steels, Trans. Jpn. Soc. Mech. Eng. A 16 (1995) 80–86. [39] W.M. Garrison Jr, N.R. Moody, Ductile fracture, J. Phys. Chem Solids 48 (1987)

473

1035–1074. [40] W.A. Spitzig, R.E. Smelser, O. Richmond, The evolution of damage and fracture in iron compacts with various initial porosities, Acta Met. 36 (1988) 1201–1211. [41] H. Qiu, H. Mori, M. Enoki, K. Teruo, Evaluation of Ductile Fracture of Structural Steels, ISIJ Int. 39 (1999) 358–364. [42] G. LeRoy, J.D. Embury, G. Edwards, M.F. Asbhy, A model of ductile fracture based on the nucleation and growth of voids, Acta Met. 29 (1981) 1509–1522. [43] D. Kwon, Interfacial de-cohesion around spheroidal carbide particles, Scr. Met. 22 (1988) 1161–1164. [44] J.B. Kosco, D.A. Koss, Ductile fracture of mechanically alloyed iron-yttria alloys, Met. Trans. A 24 (1993) 681–687. [45] P. Poruks, I. Yakubtsov, J.D. Boyd, Martensite-Ferrite interface strength in low carbon bainitic steel, Scr. Mater. 54 (2006) 41–45. [46] G. Avramovic-Cingara, C.A.R. Saleh, M.K. Jain, Void nucleation and growth in dual phase steel 600 during uniaxial tensile testing, Met. Mater. Trans. A 40 (2009) 3117–3127. [47] F. Zok, J.D. Embury, Analysis of Delaminated Fractures in High Strength Steels, Met. Trans. A 21 (1990) 2565–2575.