July 24, 2008

Strain Rate and Temperature Effects on the Formability and Damage of Advanced High-Strength Steels

By Winkler, S Thompson, A; Salisbury, C; Worswick, M; Van Riemsdijk, I; Mayer, R

In order to understand the crashworthiness and formability of advance high-strength steels, the effects of strain rate and temperature on the constitutive response of DP 600 and DP 780 steel tubes were investigated and compared with commercial drawing quality (DQ) and high strength low alloy (HSLA) 350 steel tubes. Uniaxial tensile tests were conducted at quasi-static (QS) (0.003 and 0.1 s^sup -1^), intermediate (30 and 100 s^sup -1^), and high (500, 1000, and 1500 s^sup -1^) strain rates using an Instron, instrumented falling weight impact tester and tensile split Hopkinson bar (TSHB) apparatus, respectively. Elevated temperature tests at 150 [degrees]C and 300 [degrees]C were also conducted at high strain rates. Following testing, metallography and microscopy techniques were used for material and damage characterization. The results obtained show that the steels studied exhibit a positive strain rate sensitivity. Compared to DQ and HSLA 350, the DP steels were found to have less formability at QS rates but enhanced formability at higher strain rates. A decrease in strength and ductility was measured with increasing temperature for the DP steels, indicating a reduction in energy adsorption due to adiabatic heating during a crash event. DOI: 10.1007/s11661-008-9495-4

(c) The Minerals, Metals & Materials Society and ASM International 2008


IN the interest of improving fuel efficiency, automotive manufacturers are driven to produce lighter weight vehicles without compromising passenger safety and environmental standards. Vehicle weight reduction can be achieved by substituting mild steel in automotive bodies with advanced high-strength steels, such as DP 600 and DP 780. These low-carbon microalloyed steels can attain very high strengths while retaining moderate ductility, making them ideal for energy absorption during automotive crash. Their inherent high strengths allow for thinner structural components, making a lighter and more fuel efficient vehicle.

Commercial cold-rolled dual-phase grades are produced from steels of suitable chemical composition through cold reduction and intercritical annealing between the A1 and A3 temperatures of the Fe- C diagram. Soaking at the intercritical annealing temperature creates small pools of austenite in the ferrite matrix. Upon rapid cooling, austenite is transformed into martensite, producing a ferrite-martensite dual-phase microstructure. The volumetric expansion of the martensite creates a high mobile dislocation density in the ferrite, which is responsible for the continuous yielding behavior and high initial work-hardening rate attributed to dual-phase grades.[1-3] The mechanical properties of dual-phase steels at low and intermediate strain rates have been studied by a number of researchers. Beynon et al.[3,4] conducted tensile tests on DP 500 and DP 600 sheets at strain rates of 0.001 and 100 s^sup - 1^. The alloys studied displayed an increase in strength and a slight decrease in elongation to fracture with increasing strain rate. DP 600 and DP 800 alloys tested by Schael and Bleck[5] at quasi-static (QS) rates to 250 s^sup -1^ were found to exhibit a positive strain rate sensitivity. At rates above 100 s^sup -1^, a sharp increase in strain rate sensitivity and a sharp decrease in ductility were noted. Tarigopula et. al.{6] carried out tensile split Hopkinson bar (TSHB) tensile tests on DP 800 at strain rates up to 444 s^sup -1^. The results obtained are in agreement with Reference 5, but no change in ductility was noted. Recent work published by Huh et al.[7] on DP 600 and DP 800 type steel sheets snowed that both flow stress and ductility increased with increasing strain rates when tested from 0.003 to 200 s^sup -1^.

In general, the formability, energy absorption, and damage evolution of advanced high strength dual-phase steels in a crash event are not yet fully understood. The objective of this work is to characterize these steels throughout a complete range of strain rates (0.003 to 1500 s^sup -1^) and temperatures and to compare their properties with commercial drawing quality (DQ) and high strength low alloy (HSLA) 350 steels.


The materials studied are 3-in.-diameter DQ, HSLA 350, DP 600, and DP 780 steel tubes. The tubes were manufactured by Dofasco Inc. (University of Waterloo, Waterloo, Canada) using a roll forming and electric resistance welding process from hot-rolled sheets for DQ and cold-rolled hot-dip galvanized sheets for the other steels. The tube thicknesses are 1.8 mm for DQ, HSLA 350, and DP 600; and 1.5 mm for DP 780. The microstructures of the DP steels consist of approximately 5.5 pct martensiteand 10 pct bainite for DP 600 and 7 pct martensite and 30 pct bainite for DP 780. The remaining phase present in both DP steels is ferrite. The chemical compositions of the steels studied are listed in Table I.

Uniaxial tensile QS experiments were performed using an Instron servo-electric machine (University of Waterloo, Waterloo, Canada) at a strain rate of 0.003 and 0.1 s^sup -1^. Intermediate strain rates of 30 and 100 s^sup -1^ were carried out using an Imatek instrumented falling weight impact tester (University of Waterloo, Waterloo, Canada).[8,9] High strain rate testing was performed at strain rates of 500, 1000, and 1500 s^sup -1^ using a TSHB apparatus.[8] Elevated temperature tests were also conducted at 150 [degrees]C and 300 [degrees]C at high strain rates with the aid of a radiative furnace. The time taken to reach the desired test temperature is 13 and 11 seconds for 150 [degrees]C and 300 [degrees]C, respectively.[9] Mini-tensile samples with a gage length of 12.5 mm and width of 1.75 mm were employed. These samples were designed to obtain uniaxial constitutive data with a stress-strain response prior to necking that is in agreement with standard 50-mm ASTM E8 samples.[10-13] All samples were fabricated with the rolling direction oriented along the tensile loading axis. A gripping mechanism consisting of a wire-EDM cut curved slot for a single specimen thickness with a steel screw for clamping pressure was used for all tests. For the TSHB apparatus, the grips are directly incorporated into the pressure bars to avoid wave distortion caused by threads.[10,11]

After testing, the specimens were analyzed using metallography and microscopy techniques. The specimens were mounted in epoxy resin, ground, and polished to a mirror finish using 500, 1200, and 4000 grit SiC paper, followed by 3-, 1-, and 0.25-[mu]m diamond paste. Quantitative wall thickness and damage measurements were performed at regular intervals, from every 50 [mu]m (in the highly necked region) to 200 [mu]m (in uniform deformation region) from the fracture surface. These measurements and microstructural characterization were carried out using an Olympus BH60 optical microscope equipped with ImagePro Plus 5.1 image analysis software. Fractography and area reduction measurements were conducted using a JSM 840 scanning electron microscope (SEM, University of Waterloo, Waterloo, Canada).


Figure 1 shows the effect of strain rate on the true stress vs effective plastic strain response of the different tubes at room temperature (23 [degrees]C). For all steels, an increase in strength was observed with increasing strain rates. This positive strain rate dependence is attributed to the increase in the activation energy of dislocations to bypass obstacles.[13,14] The increase in strength with strain rate is most notable between 0.003 and 500 s^sup -1^. Little difference is seen between the high-strain-rate TSHB results. This may be due to the limitation and higher noise-to-signal ratio of the TSHB apparatus. Note: because the effective plastic strain range of the TSHB is governed by the size of the apparatus, the specimens tend not to fail on the first loading pulse but require multiple loading cycles to induce fracture.[16-18]

A similar trend is observed for sheet specimens. Figure 2 compares the tube and sheet response of the steels at different strain rates. For all steels and strain rates, the tube materials tend to exhibit a higher yield stress and strength than the sheet material. This is expected because some work hardening would have been imparted to the material during the tube forming process, and is in agreement with previous work.[15] Since the sheet and tube material responses are similar, only the results of the tube material will be included in further discussions.

The elongation to failure of the tube specimens was measured and plotted against strain rate in Figure 3. The error bars in the figures indicate the standard deviation of at least three repeated measurements. It can be seen in Figure 3 that at strain rates of 100 s^sup -1^ or less, the DQ steel has superior ductility, followed by HSLA 350, DP 600, and DP 780. For the DQ tube, increasing strain rates have the overall effect of reducing the total elongation of the material. In contrast to this, the total elongation of the DP and HSLA steels was found to initially decrease from 0.003 to 0.1 s^sup -1^, before increasing with increasing strain rate. A drop in elongation was measured from 1000 to 1500 s^sup -1^ for HSLA 350 and DP 780. From a strain rate of 1000 s^sup -1^ and above, both the DP and HSLA steels exhibit similar ductility or better than the DQ steel. This suggests that the HSLA 350, DP 600, and DP 780 steels may be more formable than the DQ steel at strain rates of 1000 s^sup -1^ or higher. Figure 4 plots the percentage area reduction vs strain rate for the four different tubes. As the strain rate increases, the area reduction for DQ gradually decreases. For the DP steels, the area reduction values initially decrease before increasing from 30 s^sup -1^ onward. This result is consistent and is in agreement with the elongation data presented in Figure 3. No significant change in area reduction was measured for the HSLA tube.

Thickness strain measurements were conducted along the gage length of the failed specimens for each test condition. The typical thickness strain distribution measured from the minimum cross- sectional area adjacent to the fracture surface is shown in Figure 5. Note that for all tubes and strain rates, the first 2000 [mu]m closest to the fracture surface is a region of localized necking. Beyond that, the material is deforming uniformly. As expected, the severity of the localized neck is dependent on the strength and ductility of the material. Hence, the most severe necks were observed in the DQ specimens, followed by HSLA 350, DP 600, and DP 780.

The amount of thinning achieved by the different tubes with respect to strain rate were measured and plotted in Figure 6. Measurements were taken away from the region of localized necking at a distance of 3500 [mu]m from the minimum cross-sectional area of the specimen. At QS strain rates, the DQ tube is observed to experience more thinning than the other tubes. However, as the strain rate increases, the amount of thinning that is achieved by the DQ steel decreases; while those of the DP steels were found to increase from intermediate strain rates or higher. For HSLA 350, the amount of thinning was found to increase from a strain rate of 0.003 to 30 s^sup -1^, and then it decreases. These results are in agreement with the data presented in Figures 3 and 4. The improved ductility observed in the DP steels at higher strain rates indicates that thinner automotive parts could be manufactured with these steels when deformed under high-strain-rate conditions.


Elevated temperature tests were carried out at strain rates of 500 and 1500 s^sup -1^, at 150 [degrees]C and 300 [degrees]C. Figure 7 shows the temperature effects on the true stress vs effective plastic strain response of the different tubes. A significant decrease in strength occurs with increasing temperature for all four steels at both strain rates. For a rise in temperature from 23 [degrees]C to 300 [degrees]C at a strain rate of 1500 s^sup -1^, the decrease in strength at 0.1 effective plastic strain is found to be 33, 32, 23, and 14 pct for DQ, HSLA 350, DP 600, and DP 780, respectively. This suggests that the dual-phase steels are less susceptible to thermal softening effects than DQ or HSLA 350. This could be due to higher carbon and silicon contents of the dual- phase steels (Table I). Carbon is a solid solution strengthener for bainite, martensite and austenite, while silicon is a ferrite strengthener.[14,16]

It is observed that the drop in strength from 150 [degrees]C to 300 [degrees]C is less than from 23 [degrees]C to 150 [degrees]C, particularly for HSLA 350, DP 600, and DP 780. This may indicate the presence of dynamic strain aging due to the increased mobility of free carbon at high temperatures, usually around the 200 [degrees]C to 350 [degrees]C temperature range.[10-13] A slower decrease or a slight increase in the strength of steels was observed at 200 [degrees]C to 350 [degrees]C at low strain rates by many researchers.[11-13] Tensile tests conducted by Astakhov et al.[10] at different strain rates and temperatures showed that as the strain rate increases, the rate at which the strength of the material reduces due to increasing temperatures increases. This is due to the carbon atoms having less time to cluster around dislocations and inhibit their motion at higher strain rates. However, when the temperature exceeds 250 [degrees]C, the rate at which strength drops is also reduced at high strain rates (1600 s^sup -1^). Hence, it is hypothesized that as the strain rate increases, a higher temperature is required to create a more intense influx of carbon atoms to inhibit dislocation motion.

Figures 8 through 10 show the temperature effects on the total elongation, area reduction, and thickness strain of the tube specimens, respectively. Note that because no failed specimens were obtainable at the 500 s^sup -1^ elevated temperature tests, only results conducted at 1500 s^sup -1^ are presented. The measurements in Figure 8 show that the total elongation of the DQ specimens tends to increase with increasing temperature, but decreases for the HSLA and DP steels. This could be due to the more intense dynamic strain aging in HSLA 350 and DP steel (both of which have significantly more carbon than DQ), which is known to reduce ductility.[10-13]

The percentage area reduction for all tubes in Figure 9 is found to increase with temperature, resulting in more pronounced localized necks. Thickness strain measurements in Figure 10 show a similar trend to the elongation results in Figure 8 (measurements were obtained from region of uniform deformation, 3500 [mu]m from the minimum cross-sectional area of the specimen). Higher thickness strains were measured for the DQ material at elevated temperatures, while those of DP 600 and DP 780 were found to decrease with temperature. Little change in thickness strains was measured for HSLA 350.

The results presented in Figures 8 through 10 show that the formability of the DQ material is enhanced with increasing temperature, allowing for thinner parts to be produced. In contrast to the higher-solute containing steels, DP 600 and DP 780 in particular, their ductility deteriorated with increasing temperature. This suggests that a combination of thermal softening and dynamic strain aging results in a loss of formability, causing an earlier onset of material instability and failure in the DP and HSLA steels.


The optical micrographs in Figure 11 show the typical cross section and microstructure of the failed tube specimens. For all test conditions, final failure occurred via a ductile-shear fracture mode for all steels. As the strength of the material increases from DQ to DP 780, the specimens undergo less localized necking and the shear mechanism becomes increasingly dominant. Higher resolution images of the failed specimens etched with 2 pct Nital are shown on the right-hand side of Figure 11. For the DQ and HSLA material, damage was observed to nucleate at the pearlitic islands present at grain boundaries or where hard oxides (Figure 12(b)) are present. For both the DP steels, damage was found to occur primarily where thicker bands of martensite are present. This is most notable in the DP 600 specimens, which tend to have one large central martensite band or several bands of martensite in close proximity to one another running along the middle of the specimen (Figures 11(c) and 12(c)). These observations are consistent with the findings of previous work[15]

The SEM images in Figure 12 show the typical fracture surfaces of the failed steel tubes. Similar to the images in Figure 11, the specimens are observed to have failed via a ductile-shear mechanism. Higher resolution images of the center of the specimens presented on the right of Figure 12 show a dimpled morphology that is characteristic of a ductile failure mode. Failure appears to have commenced via a void nucleation, growth, and coalescence mechanism, followed by final failure via shear. The size and depth of the dimples can be seen to reduce in the higher strength materials, indicating a reduction in ductility. As the strength of the material increases from DQ to DP 780, the shear mechanism becomes more dominant, as indicated by the larger shear lips.

Damage measurements were carried out on the failed specimens. The typical damage distribution obtained for the steel specimens is shown in Figure 13. In agreement with observations in Figure 11, higher amounts of damage were measured for the DQ material, followed by HSLA 350, DP 600, and DP 780. For all specimens, damage is most pronounced within 500 [mu]m from the minimum cross-sectional area or the region adjacent to the fracture surface, where strains and hydrostatic stresses are highest. Strain rate and temperature appear to have little or no influence on the damage values obtained. The amount of damage present is observed to be dependent on the local microstructure heterogeneity of the individual specimens studied. In the DQ and HSLA 350 specimens, voids tend to nucleate at the pearlitic islands at the ferrite grain boundaries or at oxide inclusions. For the dual-phase steels, voids are generated at the martensite phase by means of martensite fracture or martensite- ferrite decohesion. Less damage was consistently measured for the higher martensitecontaining DP 780 than DP 600. This is in agreement with the observations of Mazinani et al.,[16] who showed that the number and volume fraction of voids are inversely proportional to the volume fraction of martensite present.


Flow stress obtained from the uniaxial tensile tests conducted show that the DQ, HSLA 350, DP 600, and DP 780 tubes have a positive strain-rate sensitivity for the range of strain rates investigated. Both the elongation and area reduction values were found to decrease with increasing stain rate for the DQ material, suggesting that increasing strain rates have a detrimental effect on its formability. For the DP and HSLA steels, formability was found to increase at strain rates above 30 s^sup -1^, indicating enhanced formability at high strain rates. The flow stress obtained from elevated temperature tests shows a reduction in strength of all steels at elevated temperatures. As the strength of steels increases from DQ to DP 780, the drop in strength due to increasing temperatures is found to decrease. This can be attributed to the higher carbon content and dynamic strain aging influences of these steels. Increasing elongation and area reduction values with increasing temperature for the DQ tube indicate that the ductility of this material is improved at high temperatures. In contrast to this, the total elongation and area reduction of the other steels were found to decrease and increase with temperature, respectively. A combination of thermal softening and dynamic strain aging at elevated temperatures appears to contribute to the loss in ductility and earlier onset of failure.

Metallographie and microstructural analysis showed that the all the tubes failed via a ductile-shear fracture mode. As the strength of the material increases from DQ to DP 780, the shear mechanism is found to be increasingly dominant. The amount of damage present was found to be dependent on the heterogeneity of the local microstructure. Strain rate and temperature have no significant effect on the amount of damage measured. Voids were observed to nucleate preferentially at the pearlitic island along ferrite grain boundaries in DQ and HSLA and at the thicker or denser martensitic bands for both DP steels.


1. G.R. Speich: Fundamentals of Dual-Phase Steels, The Metallurgical Society, Warrendale, PA, 1981, pp. 1-45.

2. W.F. Smith: Structure and Properties of Engineering Alloys, 2nd ed., McGraw-Hill Inc., New York, NY, 1993, pp. 119-23.

3. N.D. Beynon, T.B. Jones, and G. Fourlaris: Mater. Sci. Technol., 2005, vol. 21 (1), pp. 103-12.

4. N.D. Beynon, S. Oliver, T.B. Jones, and G. Fourlaris: Mater. Sci. Technol., 2005, vol. 21 (T), pp. 771-78.

5. W. Bleck and I. Schael: Steel Res., 2000, vol. 71 (5), pp. 173- 78.

6. V. Tarigopula, M. Langseth, O.S. Hopperstad, and A.H. Clausen: Int. J. Impact Eng., 2006, vol. 32, pp. 847-82.

7. H. Huh, S.-B. Kim, J.-H. Song, and J.-H. Lim: Int. Mech. Sci., 2003, vol. 45 (10), pp. 1645-60.

8. A. Thompson, C. Salisbury, W. Worswick, and R. Mayer: J. Phys. IV France, 2006, vol. 134, pp. 281-86.

9. A. Thompson: MASc Thesis, University of Waterloo, Waterloo, ON, Canada, 2006.

10. R. Smerd, S. Winkler, C. Salisbury, MJ. Worswick, D.J. Lloyd, and MJ. Finn: Int. J. Impact Eng., 2005, vol. 32 (1-4), pp. 541-60.

11. R. Smerd: MASc Thesis, University of Waterloo, Waterloo, ON, Canada, 2005.

12. Annual Book of ASTM Standards: Metals-Mechanical Testing; Elevated and Low Temperature Tests; Metallography, SJ. Bailey, N.C. Baldini, E.I. Barkley, K.A. Peters, J.L. Rosiak, ST. Simms, E.A. Whealen, and R.F. Wilhelm, eds., ASTM International, West Conshohocken, PA, 2003, pp. 61-82.

13. K.P. Staubhammer and L.E. Murr: Mater. Sci. Eng., 1980, vol. 44, pp. 97-113.

14. J. Van Slycken, P. Verleysen, J. Degrieck, L. Samek, and B.C. De Cooman: Metall. Mater. Trans. A, 2006, vol. 37A, pp. 1527-39.

15. S.L. Winkler, Y. Ososkov, A. Bardelcik, M. Jain, D.S. Wilkinson, and MJ. Worswick: Proc. SAE 2005, SAE Inc., Houston, TX, SAE paper no. 2005-01-0092.

16. M. Mazanani and WJ. Poole: Metall. Mater. Trans. A, 2007, vol. 38A, pp. 328-39.

17. M.M. Astakhov, A.N. Golenchenko, A.V. Loginov, L.P. Loshmanov, and A.E. Trotsenko: Special Features of Dynamic Strain Aging in High-Speed Tension of Steel 20, Moscow Institute of Engineering Physics, Moscow; translated from Problemy Prochnosti, 1991, No. 9, pp. 28-30.

18. S.H. Mousavi Anijdan and H. Vahdani: Mater. Lett., 2005, vol. 59, pp. 1828-30.

19. S. Gunduz and R.C. Cochrane: Mater. Sd. Technol., 2003, vol. 19, pp. 422-28.

20. S. Gunduz and M. Acarer: Industr. Lubric. Tribal., 2005, vol. 57, pp. 145-49.

S. WINKLER and C. SALISBURY, Research Associates, and M. WORSWICK, Professor and Associate Dean of Engineering, are with the Department of Mechanical Engineering, University of Waterloo, Waterloo, N2L 3G1, ON, Canada. Contact e-mail: [email protected] lagavulin.uwaterloo.ca A. THOMPSON, formerly with the Department of Mechanical Engineering, University of Waterloo, is with Babcock and Wilcox, Cambridge, NlR 5V3, ON, Canada. I. VAN RIEMSDIJK, Senior Research Associate, is with Research and Development, Dofasco Inc., Hamilton, L8N 3J5, ON, Canada. R. MAYER. Staff Research Engineer, is with the GM R&D Center, MC 480-106-256, Warren, MI 48090-9055.

Manuscript submitted September 20, 2007.

Article published online April 15, 2008

Copyright Minerals, Metals & Materials Society Jun 2008

(c) 2008 Metallurgical and Materials Transactions; A; Physical Metallurgy and Materials Science. Provided by ProQuest Information and Learning. All rights Reserved.