On Factors Affecting the Phase Transformation and Mechanical Properties of Cold-Rolled Transformation-Induced-Plasticity-Aided Steel

October 2, 2008

By Soliman, Mohamed Palkowski, Heinz

Two Mo-Nb microalloyed transformation-induced-plasticity (TRIP) steels, with Al contents of 0.23 and 0.65, were subjected to several hot-rolling conditions designed to generate different ferrite morphologies and grain sizes. These structures were then cold rolled and TRIP annealed under different heat-treatment conditions. To further develop TRIP steel in terms of strength and ductility, stabilizing retained austenite by isothermal bainitic transformation was studied in detail. Microstructure observation and tensile tests were conducted, and volume fractions of retained austenite were measured. It was observed that increasing the aluminum content enhances the transformation rate and increases the total amount of bainite fraction at the expense of retained austenite. The latter effect enhances formability by increasing ductility. Furthermore, it was observed that the hot-rolling schedule, prior to cold rolling and heat treatment, has a decisive effect on structure refinement, which enhances the strength-ductility balance of the final product. To study the transformation behavior, dilatometer testing was conducted under conditions similar to that of the heat treatment. Thermodynamic calculations were used to verify the results. DOI: 10.1007/s11661-008-9594-2

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


THE remarkable strength-ductility balance in transformation- induced-plasticity (TRIP) steel results from the occurrence of the TRIP phenomenon during deformation[1] The coexistence of austenite with a certain microstructural stability is of vital importance in order for this phenomenon to occur and, hence, to achieve the desired properties. The austenite retention is usually obtained by combining the effects of chemical composition and typical heat treatment. In this respect, adding large amounts of silicon to TRIP steel ensures that cementite precipitation is unlikely to occur in the microstructure during bainite formation.[2) The absence of cementite ensures that the carbon will enrich the austenite rather than form cementite plates. Therefore, after the bainite transformation finishes by further cooling to room temperature (RT), the austenite is stabilized. Jeong et al. concluded that austenite retention in these low-alloyed steels is almost impossible with silicon concentrations much below 1 wt pct.[3] However, these high required silicon levels are outside standard industrial practice for producing flat products because of the following.

(a) Steel with more than 1 pct Si has a poor Zn coating quality after continuous galvanizing, due to the presence of Si-Mn oxides on the strip surface.[4]

(b) The high Si content of these steels causes red scales to form in bands. After pickling, the oxides are completely removed, but the band remains visible on the surface of the pickled steel.[5]

Consequently, studies have been performed with other elements that can substitute for the role of silicon. Presently, aluminum seems to be the most promising candidate. However, in addition to the fact that high aluminum content in steel causes serious casting problems, a full substitution of silicon by an equivalent amount of aluminum leads to a marked deterioration of the strength-ductility balance.[6,7] De-Meyer et al.[8] proposed that silicon can best be partially replaced by aluminum with an increase in the carbon content. Bleck also suggested that combining silicon, aluminum, and phosphorus is a reasonable compromise and could be the most important alloying concept for low-alloyed TRIP steels.[9]

Little or no information can be found in the literature about the sole effect of aluminum content on the phase transformation and mechanical properties of the TRIP-aided steels. De-Meyer et al.[8] and E. Emadoddin et al.[10] presented some results about Si-Al- alloyed TRIP steels. However, their comparison of the number of phases and the mechanical properties is misleading, because the carbon content was different in their alloys. Carbon is the main alloying element by which both all transformations are noticeably affected and the final microstructure and the mechanical properties are controlled.[11] On the other hand, without changing the carbon content of the alloys, Jacques et al.[7] studied the effect of a partial substitution of Si by Al while varying the silicon content.

The present study is aimed at ascertaining how Al content variations, the prior hot-rolling conditions, and heat-treatment parameters affect the microstructure and mechanical properties of the cold-rolled Mn-Si-Al TRIP-aided steel alloyed with Mo-Nb.


A. investigated Materials

The Si-Al-Mo-Nb steel studied in this work was produced in the laboratory. The two alloys studied, steels 1 and 2, differ in their Al content. The chemical composition is given in Table I. The alloys contain Mo and are microalloyed with Nb. The Nb in solid solution has been found to improve the TRIP properties; the Mo retards the precipitation of Nb(C,N), thus potentially improving the effectiveness of Nb as a TRIP enhancer [12.13] furthermore, Mo retards austenite transformation to both ferrite and pearlite, effecting more manageable process control.[14]

B. Heat Treatment and Dilatometry

The heat treatment had been conducted on the mechanical testing samples and on samples for microconstituents investigations using salt baths. This had been done by austenitizing in Durferrit GS 540/ R2 and austempering in Durferrit AS 140 salt baths*. The inertor R2 was added, to prevent any oxidation or decarburization during austenitizing.

* Durferril GS 540/R2 and Durferrit AS 140 are trademarks of Durferrit GmbH, Mannheim, Germany.

Dilatometric measurements were conducted on a Baehr dilatometer DIL 805A/D** (Figure 1), which has a resolution of 0.05 [mu]m/0.05 K. All the dilatometric measurements were performed using specimens 2.5 x 5 x 10 mm in size. The test specimens were degreased with an acetone solvent. Sheathed type S Pt/Pt-10 pct Rh thermocouples with a nominal diameter of 0.1 mm were individually spot welded to the surface of the specimen in the central position of the 5 x 10- mm surface-to-monitor temperature. Each sample was held between two quartz rods, with its 10-mm side along the rods. One of the rods is fixed; the other one is connected to a linear variable differential transducer (LVDT). A reference rod is also connected to the LVDT. The dimension variations of the specimens during the thermal cycle are transmitted via the moving quartz pushrod to the LVDT sensor. After placing the sample between the pushrods, the insulating sheaths on the thermocouple wires had been moved along the thermocouple wires until they contacted the specimen surface.[15] The thermal cycles were performed under a vacuum of 0.005 Pa, and helium was used for cooling. A computer and data acquisition system recorded the dilatometric change and temperature as a function of time, and cross correlated the relative change in length as a function of temperature.

** Baehr Dilatometer DIL 805A/D is a trademark of Bahr- Thermoanalyse GmbH, Huellhorst, Germany.

C. Materials Characterization

The materials were machined to the required specimen size and geometry prior to heat treatment. This was done to avoid the transformation of retained austenite to martensite due to machining forces. Standard subsize tensile specimens were machined in accordance with ASTM standard E8-03; the specimens had a width of 6.4 mm and a gage length of 25.4 mm and were machined transverse to the rolling direction from the cold-rolled bands.

For investigating micro structural constituents, the samples were ground and polished using the normal metallographic preparation procedure. The microstructures were examined with an optical microscope after etching, using LePera[16,17] or nital etchant. The samples for the scanning electron microscope (SEM) were tempered at 473 K for 2 hours before mechanical preparation and deep etching in nital. The tempering treatment at 2 hours and 473 K was performed to enable a good resolution of the martensite substructure.[16]

A magnetic measurement technique that made use of a hysteresis recorder was employed to estimate the amount of the retained austenite (K^sub gamma^). As compared to X-ray diffraction (XRD), the advantages of the magnetic technique are that magnetic measurement is performed on the whole volume, the specimens require no special preparations, and the measurements are fast, reliable, and more sensitive to retained austenite.[18,19]

The thermodynamic calculations were performed using THERMO- CALC[dagger] and the database TCFE3; the ferrite, austenite, and cementite phases were considered.

[dagger] THERMO-CALC is a trademark of Thermo-Cale software AB, Stockholm, Sweden.


A. Hot-Rolling Conditions

1. Estimation of nonrecrystallization temperature

When dynamic recrystallization (DRX) takes place during material deformation, e.g., a rolling process, grain size is determined by the steady-state flow stress. Whenever the critical strain epsilon^sub c^ for the onset of DRX is reached and exceeded during hot deformation, a metadynamic recrystallization (MDRX) takes place after the straining is interrupted, consequently coarsening the austenite grains. If deformation is interrupted before reaching the strain epsilon^sub c^ and if the temperature is high enough, then static recrystallization (SRX) takes place. The austenite grains are undergoing refinement by deformation in the recrystallization region. This is important because the grain size of the austenite strongly affects both the kinetics of the subsequent gamma [arrow right] alpha transformation as well as the ferrite grain size, i.e., smaller austenite grains lead to smaller ferrite grains. When deformations are applied at temperatures below that of nonrecrystallization (T^sub nRX^), the austenite grains elongate and deformation bands form within the grains. The process is called “pancaking.” As the deformation amount increases in this region, the number of nucleation sites at the austenite grain boundaries and within the austenite grains increases. Thus, the gamma [arrow right] alpha transformation from deformed austenite yields much finer ferrite grains than does the transformation from recrystallized and strain-free austenite. Therefore, T^sub nRX^ is a very important parameter and its determination represents a crucial step in designing rolling schedules. Accordingly, the present hot-rolling schedules were selected according to this temperature. The T^sub nRX^ was determined using the method described elsewhere.[20-22] The estimated values of T^sub nRX^ for steels 1 and 2 are 1138 and 1159 K, respectively. 2. Hot-rolling schedules

With these estimated temperatures, the deformation part for the schedules was determined in such a way that all three possibilities were covered, namely, all deformations conducted above T^sub nRX^, deformations conducted below T^sub nRX^, and deformations conducted at mixtures of temperatures below and above T^sub nRX^. Figure 2 shows a schematic representation of the employed schedules. The schedule that resulted in microstructure formed from the recrystallized austenite is denoted R, whereas that which resulted in microstructure formed from pancaked austenite is denoted P. The RP schedule is for the microstructure resulting from the recrystallized and then pancaked austenite.

A reheating temperature of 1523 K was selected to dissolve the Nb in solution entirely, as shown in Figure 3. In this figure, the Nb in austenite increases by increasing the reheating temperature, so that it finally reaches its bulk content in the alloy at approximately 1523 K. If the homogeneous annealing temperature prior to hot rolling is less than the dissolving temperature of Nb(C,N), the undissolved carbonitrides will exist and abnormally grow so that they weaken the refinement of austenite grains and the precipitation strengthening of the Nb(C,N).[23]

Thus, during reheating before hot rolling, the Nb completely dissolves in austenite; then, during the subsequent cooling and hot- rolling process, the Nb(C1N) precipitates. The addition of Nb to TRIP steels effectively refines the austenite grain during the hot- rolling process because the precipitates retard austenite recrystallization and, in turn, refine the final microstructure after intercritical annealing. The recrystallization process is prevented by copious precipitation below T^sub nRX^.[24,25] An extensive study of the precipitation processes in Nb-microalloyed TRIP-steel is reported elsewhere.[23]

The cast ingots were hot rolled in four passes (Figure 2), from a thickness of 19 to 4 mm, and with a true strain value phi = 0.38 at each pass. The temperature was continually monitored with a pyrometer during the final cooling of the hot-rolled strips in air. It took between 20 and 31 minutes to cool the strips from the finish rolling temperature to 623 K. The surface oxide scale was then removed from the slabs using shot blasting; finally the 4-mm-thick hot-rolled plates were cold rolled to a thickness of 2.5 mm.

Figure 4 shows the microstructure obtained after the different hot-rolling schedules. Acicular ferrite (AF) dominates and polygonal ferrite (PF) is also present. Vickers microhardness tests HV0.05 were performed within regions occupied exclusively by PF and AF (Figures 5(a) and (b)). The average HV0.05 values for PF and AF phases are 1370 and 2900 MPa, respectively.

The AF is a nonequiaxed ferrite (Figure 5(c)) formed upon continuous cooling by a mixed diffusion and shear mode of transformation that begins at a temperature slightly higher than the transformation temperature of upper bainite. The Mo and Mn elements (Table I) causes the beginning of transformation to be retarded, because the ferrite start curve on the continuous cooling transformation (CCT) diagram is displaced to longer times. Therefore, the austenite decomposes at lower temperatures during cooling, yielding fine-grained, AF.[26,27]

Figure 5(c) shows a remarkable characteristic of this type of microstructure: it possesses an irregular configuration, which has various grain sizes distributed in a heterogeneous manner with random orientations. The PF grains are equiaxed with different average sizes, depending on the applied rolling schedule.

Table II compares the PF contents (ferrite percent) and their grain sizes (d^sub f^) that resulted from the different hot-rolling schedules for both alloys. The higher ferrite content of steel 2 can be explained by its lower hardenability, because its higher aluminum content shifts the ferrite nose in the CCT diagram toward a lower time.[11] It is also observed that the hardenability is sensitive to the deformation temperature. Decreasing the deformation temperature decreases the steel hardenability; hence, the ferrite content increases by moving the deformation schedule toward a lower temperature.

On the other hand, the variation in the ferrite grain size that occurs when the hot-rolling schedule is varied is caused by the different austenite grain morphology that forms the ferrite. For the P schedule, the grains are produced from work-hardened (pancaked) austenite through an austenite-to-ferrite transformation. The sizes of these ferrite grains are smaller than those produced from the recovered and recrystallized austenite grains of the R schedule. It is well known that ferrite tends to nucleate at the austenite grain boundaries.[28'29] Thus, the ferrite grain size developed after the transformation strongly depends upon the austenite grain structure that appears just before the start of the transformation. During the straining of the austenite in the nonrecrystallization region, deformation bands and twinning boundaries form and the dislocation density inside austenite grains is greatly increased; the increased dislocation density provided favorable nucleation sites and an enhanced nucleation rate. The grain refinement effect of deformation in the nonrecrystallization region is greater than that in the recrystallization region.[30,31] The higher austenite decomposition kinetics causes the faster nucleation rate of ferrite. The nonrecrystallized austenite transforms to ferrite at a faster rate than does recovered or recrystallized austenite, due to two effects:[31]

(a) the higher internal energy of the deformed and, thus, less stable austenite, and

(b) the larger number of nucleation sites provided by defects.

B. Intercritical Annealing Conditions

During the current investigation, the transformation kinetics were studied using the dilatometer. All the dilatometric measurements were performed on samples produced from the cold- rolled strips with the long specimen side oriented parallel to the rolling direction. Unless mentioned, the dilatometric tests were performed on samples from the RP materials.

1. Continuous heating transformation

In order to define the intercritical region, dilatometric measurements were applied by heating specimens up to 1373 K. The Ac1 and Ac3 temperatures are plotted in the continuous heating transformation (CHT) diagram in Figure 6.

Figure 6 shows that the increased heating rate resulted in increased Ac1 and Ac3. The current results indicate that the healing rate exerts a stronger influence on the finish temperature when all the ferritic phase is exhausted than on the start temperature at the beginning of carbide dissolution. On the other hand, the dependence of the transformation temperature on the heating rate decreases as the heating rate decreases. At a very slow heating rate, the Ac1 and Ac3 are independent of the heating rate. This feature is associated with the transformation temperatures at equilibrium (slow heating and cooling). The equilibrium Ae1 and Ae3 transformation temperatures depend only on the chemical composition and are unaffected by the heating or cooling rates.[32] This infers that the Ac1 and Ac3 are close to the Ae1 and Ae3, respectively, when the transformation temperatures become independent of the heating rate. In this context, fine grain size resulted from the hot- and cold- rolling processes; this favors the transformation process in such a way that a very slow heating rate (less than 0.05 K/s), is not required to trace the equilibrium points. This is because the greater density of nucleation sites, resulting from the fine- grained structure, enhances the transformation kinetics.

2. Phases in equilibrium

From the variation of the relative change in length as a function of temperature, the lever rule was employed to calculate the formed austenite fractions (f^sub gamma^). Figure 7 shows the results for both steels. Depending on the previous discussion, for better tracing of the equilibrium points, the 0.05 K/s dilatation curve was used for these calculations.

Figure 8 compares the measured formed austenite fraction with the one predicted using the THERMOCALC TCW3 software. This figure also shows the predicted dependence of the ferrite and cementite on the temperature. During the continuous heating of steels, the transformation reportedly takes place by the initial growth of austenite into the carbide-rich areas, and subsequently by the growth of the austenite into ferrite and the redistribution of carbon between the former and the latter phases.[33-34] Figure 8 is marked by T^sub Cp^ for the predicted and by T^sub Cm^ for the measured kinetics, and shows a clear demarcation of these two steps. The change in slope of the line represents how f^sub gamma^ depends on the annealing temperature, which corresponds to the change in the austenite formation mechanisms. On the other hand, at the beginning of transformation (near the Ae1), the measured kinetics are slower than the predicted rates for both alloys. However, this difference decreases gradually with an increasing transformation temperature. The predicted and the measured transformation kinetics are almost identical at the end of the transformation, close to the Ae3. The reason for this difference could be the sluggish kinetics of the austenite formation at the low temperature, close to the Ae1.[34- 351

C. Heat Treatment and Microstructure Formation

The intercritical annealing temperature was chosen on the basis of the dilatometric measurement shown in Figure 8. A phase content of 70, 50, and 30 pct PF was required at the end of intercritical annealing. In equilibrium conditions, this phase distribution is obtained in the intercritical annealing temperatures (T^sub A^) shown in Table III. It should be mentioned that all the intercritical annealing temperatures used are above the measured cementite dissolution temperature T^sub Cm^. Note that cementite is a cleavage and void-initiating phase that is best eliminated from strong steels.

For heat-treatment processes, the salt baths mentioned in Section II-B were used. The intercritical annealing temperatures listed in Table III were implemented in combination with austempering temperatures, T^sub B^, of 365 [degrees]C, 400 [degrees]C, and 435 [degrees]C. The isothermal holding time for each of the two steps was 8 minutes. After austempering, the samples were quenched in water.

1. Effect of heat-treatment conditions

Figure 9 shows representative microstructural results using the same magnification for the two steels, after applying the two-step annealing conditions. In all specimens, the minor microstructure is the martensiteaustenite (MA) structure (white areas). Figures 9(a) through (c) show the influence of the intercritical annealing temperature on the microstructure of steel 2, for specimens austempered at 673 K. An increase in the intercritical annealing temperature resulted in a decrease in the PF amount to approximately reach the expected amount (70, 50, and 30 pct). A concurrent increase in the amount of the MA phase was observed.

Geib et al.[36] reported that intercritical annealing produces well-developed precipitate distributions in ferrite retained during intercritical annealing, whereas the austenite that forms on intercritical annealing dissolves any carbonitride precipitates initially present. Austenite formation at ferrite-ferrite boundaries during partial austenitization would lead to the dissolution of these stable carbides/carbonitrides residing at ferrite boundaries, making the austenite with niobium in solution sluggish to transform on quenching.[37] It is reported that microalloying steel with Nb increases the amount of retained austenite.[38]

On the other hand, the intercritical annealing temperature has no significant effect on the grain size of ferrite and retained austenite. A similar observation has been reported by Shi et al.[39] The grain sizes of the ferrite and retained austenite varied between 3 and 7 [mu]m and 0.7 and 3 [mu]m, respectively. This study detected no significant influence of the chemical composition and austempering temperature on the location and morphology of the retained austenite.

Additionally. Figure 9 shows that the structure is fully recrystallized. Consistent with that result, Petrov et al. observed no interaction between recrystallization and transformation phenomena, because the SRX was already completely finished before the start of the alpha [arrow right] gamma phase transformation.[40]

Figure 10 shows the dependence of the retained austenite content (V^sub gamma^) on the heat-treatment parameters for both alloys. In general, the retained austenite content measured at RT is dependent on two phenomena:[41]

(a) the isothermal transformation of austenite to bainite at a given temperature, and

(b) the a thermal transformation of austenite to martensite on cooling to RT after isothermal holding.

Figure 10 shows that decreasing the PF amount (increasing the intercritical annealing temperature) leads to a higher amount of retained austenite (V^sub gamma^) in the final microstructure for the two alloys. This could be attributed to the higher intercritical austenite content.

For the alloys annealed to the expected PF levels of 50 and 70 pct, increasing the T^sub B^ temperature from 638 to 708 K resulted in increasing the V^sub gamma^. For these intercritical annealing conditions, the authors also noted that the richer aluminum alloy (steel 2) had a lower V^sub gamma^ compared with steel 1. The latter observation is consistent with the light optical microscopic one.

These two observations were not recorded for the materials intercritically annealed for an expected PF level of 70 pct. Such results for V^sub gamma^ can be justified by tracking the austenite and its composition throughout the heal-treat ment process as follows: during the intercritical annealing stage, the austenite is enriched with carbon, due to the partitioning of carbon between the ferrite grains and the intercritical austenite (gamma^sub i^) during the formation of gamma^sub i^. Figure 11 uses THERMO-CALC and gives the predicted dependence of the elemental concentrations in gamma^sub i^ on the PF content (i.e., T^sub A^). During the course of the second isothermal holding (austempering), the austenite is further enriched with the carbon rejected from the bainitic ferrite. This reaction can occur until the point at which the free energy of ferrite is equal to the free energy of austenite (G^sub alpha^ = G^sub gamma^); thus, at this point, no further transformation of austenite to ferrite can occur. This residual austenite, because of its high carbon content, has a martensite start temperature (M^sub S^) below RT; thus, a certain amount of austenite from this process can be retained at RT. The amount of carbon that can enrich the austenite during this process has been found to depend on the isothermal austempering temperature, according to the T^sub 0^ concept.[27] Figure 12 shows the T^sub 0^ curve drawn for both steels I and 2 using THERMO-CALC software (TCCQ). However, the amount of carbon that enriched the V^sub gamma^ does not necessarily match the carbon content predicted by the T^sub 0^ curve. For example, if the holding time is insufficient to reach the T^sub 0^ curve, especially for low austempering temperatures that require a long holding time, the expected carbon concentration value cannot be reached. On the other hand, for a high holding temperature, the carbides can form easily and the carbon in the austenite decreases.[42] Reportedly, carbide starts to precipitate between 723 and 748 K, and this temperature range is hardly affected by chemical composition.[43]

In the current case, considering the same amount of gamma^sub i^ at the beginning of the bainitic transformation, the bainitic reaction will last to a lower concentration of carbon at a higher temperature. Thus, the final austenite will be of a lower carbon concentration (Figure 12), but of a higher content (Figure 10). The use of the relatively narrow austempering temperature range around 673 K is the reason why the results could be justified in the light of the T^sub 0^ curve. Hashimoto reported that the best combination of bainite transformation rate and carbon content at T^sub 0^ curve occurs at this temperature.[42] However, the trend may not continue for lower or higher austempering temperatures, at which the carbon in the retained austenite may not fit with the T^sub 0^ curve, due to the insufficient holding time or the carbide precipitation, respectively, as explained previously.

On the other hand, the V^sub gamma^ of the 70 pct PF material does not follow the trend expected from the T^sub 0^ curve. In this case, the gamma^sub i^ is enriched with the carbide former elements (C and Mn), whereas the carbide suppression element (Al) is at a lower level (Figure 11). This motivates the carbide formation in bainite. Carbide formation withdraws carbon from the austenite, so its stability drops noticeably. This may proceed to the extent that some of the austenite formed during the isothermal holding cannot be stabilized down to RT and, thus, transforms to martensite. Furthermore, Kim et al. have reported that the improvement of the austenite hardenability because of the high Mn content (2.52 wt pct, in their case) results in martensite transformation during cooling after the isothermal bainite holding.[44] For the material annealed to 70 pct PF, a similar excessive Mn content is predicted in gamma^sub i^ (Figure 11).

The occurrence of the martensite phase in the 70 pct PF material was confirmed with investigation using a SEM. The samples for the SEM were prepared following the procedure proposed by Girault et al.; in that manner, distinguishing martensite from retained austenite in the micrograph is possible.[16] Figure 13 shows representative SEM micrographs.

For the material annealed to 70 pct PF, the martensite phase was observed. Figure 13(c) shows a representative example. The martensite grain is characterized by its well-delineated substructure, while the retained austenite looks rather smooth. Using transmission electron microscopy, Sakuma et al[41] observed a high dislocation density in the ferrite surrounding the martensite region, whereas the ferrite surrounding the retained austenite and bainite regions showed a very low dislocation density. On the other hand, increasing the aluminum content leads to a lower V^sub gamma^ (Figure 10). The ability of steel 2 to proceed to the higher carbon enrichment of austenite during bainitic holding (Figure 12) consequently results in lowering V^sub gamma^, as discussed earlier. Therefore, increasing the aluminum content not only reduces the cementite stability but also motivates the enrichment of austenite with carbon and leads to a lower retained austenite volume fraction (with a higher carbon content and greater stability).

2. Bainite transformation kinetics

The dilatometric measurements were used to compare the bainite transformation kinetics of the two steels. During the dilatometric tests, a cooling rate of 50 K/s was used to quench the samples from the T^sub A^ temperature (Table III) to 673 K.

Figure 14 compares the dilatation curves for steels 1 and 2 after bainitic transformation at 673 K. The bainite reaction proceeds to a higher final amount of bainite for the alloy with a higher Al content. Figure 14 also demonstrates that bainite formation is accelerated by increasing the aluminum content (assuming the same PF content). THERMO-CALC calculations of the driving force for the transformation of austenite into ferrite (DeltaG^sup gammaalpha^) for steels 1 and 2 (Figure 15) confirm the thermal dilatometric results. Thus, the micro structure evolution throughout the isothermal bainite transformation demonstrated that both the bainite formation rate and the total bainite amount increase with an increase in the aluminum content.

The dilatometric experiments have shown that a cooling rate of 50 K/s is sufficient for avoiding the formation of the allotriomorphic ferrite during cooling to the T^sub B^. This was detected from the linearity of the temperature-dilatation curve observed during cooling.

3. Effect of hot-roiling conditions

As explained in Section A-2, the different hot-rolling schedules (Figure 2) resulted in pronounced differences in the hot-rolled structure size (Figure 4). Due to the latter variation, different final TRIP-aided steel structure sizes are observed. Figure 16 shows representative microstructures that result due to applying different hotrolling schedules using the same magnification. Table IV compares the final structure grain sizes of ferrite and retained austenite. Making use of the beneficial effect of the rolling below T^sub nRX^, a pronounced finer cold-rolled TRIP-aided steel structure was produced. Based on the current result (as well as on a previous report[39]), the employed heat-treatment parameters have no significant effect on the ferrite and retained austenite grain size. The current study infers that the structure fineness depends only on the hot-rolling conditions. Thus, controlling the hot-rolling schedule, prior to the cold rolling and heat treatment of TRIP- aided steel, has a decisive effect on the structure refinement and results in a more convenient structure size. On the other hand, the observed effect of the prior hot-rolling schedule on the V^sub gamma^ was very limited.

Additionally, dilatometric investigations showed that the difference in the hot-rolling schedules influences the bainite transformation kinetics in a pronounced way. Figure 17 presents the evolution of DeltaL/L^sub 0^ as a function of the transformation time for the different hot-rolling schedules applied on steel 2. Figure 17 shows that the transformation in smaller grains starts faster but proceeds at a slower rate. Indeed, the grain-size reduction causes an increase in the grain-boundary area, at which the first bainitic ferrite subunits nucleate. Thus, the transformation starts more quickly, due to an enhanced nucleation rate. The transformation then proceeds by means of the nucleation and the growth of new subunits from the tip of the previous ones toward the interior of the austenite grain; thus, when the austenite grain size is reduced, the transformation proceeds at a slower rate.[45,46] The influence of the austenite grain size on the transformation rate was already shown and modeled by Rees and Bhadeshia.[47]

In addition to the effect of the chemical composition on the bainite transformation kinetics (Figure 14), another feature highlights the transformation occurring in the cold-rolled TRIP- aided steels: the prior hot-rolling schedule.

On the other hand, instead of differing transformation rates, the bainite reaction proceeds to the same final amount of bainite (Figure 17). Accordingly, by the mass-balance relationship among the material phases, the V^sub gamma^ in the final microstructure should remain unaffected by the prior structure fineness. This dilatometric observation is consistent with the fact that the measured V^sub gamma^ is only insignificantly affected by the hot-rolling schedule.

D. Mechanical Properties

Figures 18 and 19 show the effect of the processing route on the ultimate tensile strength (UTS), yield strength (YS), and the total elongation percent (pct El) for steels 1 and 2, respectively. The YS values plotted in the histograms are the lower yield strengths or the 0.2 pct offset YS (0.2 pct YS) values, in the case of the absence of a yield point. In Figure 20, the pct El values are plotted vs the UTS values for all combinations of the rolling conditions and heat-treatment parameters (3 structure sizes x 3 annealing temperatures x 3 austempering temperatures).

It can be concluded from Figures 18 to 20 that it is the PF percent (the intercritical annealing condition) that has the most pronounced effect on the mechanical properties of both steels. The UTS values increase with a decrease in the PF percent. This can be explained by the higher austenite content that formed during the intercritical annealing. A higher amount of gamma^sub i^, results in a greater amount of high-strength bainite, after the isothermal bainitic transformation step.

On the other hand, increases in the PF percent from 30 to 50 pct results in El values with increasing percents. In addition to the fact that an increase in the soft PF content affects the pct El, the expected higher amount of retained austenite surrounded by the bainite phase in the 30 pct PF microstructure can also contribute to the observed lower pct El. That is because the retained austenite grains with the film shape located at the bainite are not transformed to martensite, even when a considerable amount of deformation is applied; thus, no contribution is made to the ductility improvement.[44]

An additional increase in the PF percent to 70 pct has no further enhancement effect on the percent of El. This observation can be justified by the retained austenite characteristics associated with the 70 pct PF material. Principally, the total elongation of TRIP- aided steel is controlled by the volume fraction and stability of the retained austenite and the difference in strength between the matrix and second phase.[48] In the 70 pct PF material, the V^sub gamma^ is obviously lower than that of the 50 pct PF material (Figure 10) and has lower stability, as well, as discussed earlier.

Furthermore, Timokhina et al. have reported that the PF in the C- Mn-Si-Nb-steel showed a ferrite deformation lower than that in the non-Nb steels; the PF also did not flow around the bainitic region during straining. This has been justified by the fact that the Nb addition has the effect of promoting precipitation hardening of the ferrite, either directly or through grain refinement, which leads to the strengthening of the ferrite.[49]

The materials annealed to 70 pct PF have recorded relatively high YS/UTS values. For this case, the retained austenite stability is very low, as explained earlier. Consequently, transformation in such low austenite stability would take place even in the elastic region.[38] Thus, the TRIP effect is most likely to take place in the elastic region and would eventually increase the YS values, but subsequent transformation in the plastic region would be limited. Consequently, the contribution of the TRIP effect in enhancing the UTS and the percent of El is also limited.

Figures 18 and 19 show that the mechanical properties do not vary similarly with respect to the isothermal bainitic holding temperature, T^sub B^. The complex nature of the microstructures and the interaction among their phases result in an unpredictable effect of T^sub B^, especially in the relatively narrow range investigated (638 to 708 K).

Since TRIP-aided steels combine high strength and high ductility, their mechanical properties are often characterized by the product of the tensile strength and the total elongation (UTS x pct El), which is known as the formability index.[42,50] At each combination of PF content and hot-rolling condition, the average form-ability index was calculated from the values recorded at the three corresponding T^sub B^ temperatures. Figure 21 shows a histogram that compares these averages. This figure shows that the highest formability index was obtained for the steels annealed to 50 pct PF. This result is associated with the strength-ductility balance recorded at this PF content (Figure 20).

Based on Figure 21, the hot-rolling schedule has a pronounced effect on the formability index for both steels- It is interesting that the hot-rolling schedule has a very limited effect on the V^sub gamma^ values. Thus, the improvement in properties seems entirely due to microstructural refinement (Figure 16). In addition to the well-known effect of grain refinement on improving mechanical properties, the greater stability of the smaller retained austenite grains has an additional improving effect. Brandt et al.[51] have stated that smaller retained austenite particles contain fewer potential nucleation sites for the transformation to martensite; consequently, these require a larger total driving force for the nucleation of martensite. Thus, by controlling the deformation temperature and the degree of deformation below T^sub nRX^ during the hot-rolling process, it was possible to improve the strength- ductility balance of the cold-rolled TRIP-aided steel. The deformation below the T^sub nRX^ results in the distinct refinement of the final TRIP-aided steel microstructure. A poorer formability index for the lower-Al-content alloy (steel 1) in spite of its higher austenite content is related to lower carbon enrichment during isothermal bainite transformation (Figure 12). This results in less stable austenite, which is then transformed to martensite in an earlier stage of plastic deformation. In cases in which the austenite stability is very low, the transformation occurs even in the elastic region, so that the dual-phase behavior can be observed. Nevertheless, a minimum amount of retained austenite should be present, in order to guarantee TRIP behavior and enhanced strength and formability.[9]

For additional analysis, the average mechanical properties of steels 1 and 2 were calculated from all the values recorded under the different employed processing routes. Table V lists the calculated values. The values are highly reliable, because they were obtained from many samples processed under different conditions, but these conditions are quite similar for the two alloys. It can be concluded that increasing the aluminum content has improved the formability of steel 2 by improving its ductility. The higher ductility of steel 2 is correlated with its higher bainite content (Figure 14) and higher retained austenite stability (Figure 12). Sakuma et al. stated that more stable austenite transforms at high strains and increases strain hardening at necking, thereby improving ductility.[41,52]


This work provided an investigation of the microstructure formation and mechanical properties obtained after different hot- rolling and two-step annealing treatment conditions of two Mn-Si-Al cold-rolled TRIP-aided steels alloyed with Mo and Nb. From the present investigation, the following conclusions can be drawn.

1. By making use of the beneficial effect of the hot-rolling below T^sub nRX^, a pronounced finer cold-rolled TRIP-aided steel structure was produced. This resulted in an improvement in the strength-ductility balance of the steel.

2. A comparison of the transformation kinetics using the thermodynamic calculations and the dilatometric method is only relevant when using a heating rate at which the transformation temperatures are independent of the heating rate; using a higher heating rate is then misleading.

3. As long as the bainite transformation in TRIP steel can comply with the T^sub 0^ curve, the retained austenite content increases by increasing the bainite transformation temperature.

4. It is the composition of intercritical austenite that controls its behavior during the bainite transformation, and not the bulk composition of the alloy. During the current study, it is proposed that the retained austenite has low stability, due to the partitioning of elements during intercritical annealing in such a way that, at a low intercritical annealing temperature, gamma^sub i^ is enriched with the carbide former elements (C,Mn); the carbide suppression element (Al) concentration, however, is at a lower level.

5. Increasing the aluminum content not only reduces the cementite stability but also achieves accomplishes the following.

(a) Motivates the formation of bainitic-ferrite and leads to a higher fraction of bainite.

(b) Lowers the fraction of retained austenite.

(c) Increases the carbon content in retained austenite and improves its stability.

(d) Enhances the formability though increasing the ductility. (This is a result of the latter effect.)

6. For the given alloying elements, the most promising microstructures with respect to the strength-ductility balance are those containing 50 pct PF.


1. V.F. Zackay, E.R. Parker, D. Fahr, and R. Busch: Trans. ASM, 1967, vol. 60, pp. 252-59.

2. O. Matsumura, Y. Sakuma, and H. Takechi: Iron Steel Inst. Jpn., 1987, vol. 27, pp. 570-79.

3. W.C. Jeong and J.H. Chung: HSLA Steels: Processing, Properties and Applications, TMS, Warrendale, PA, 1992, pp. 305-11.

4. J. Maki, J. Mahieu, B.C. De Cooman, and S. Claessens: Mater. Sci. Technol., 2003. vol. 19. pp. 125-31.

5. L. Tosal-Martiez, D. Vanderschueren, S. Jacobs, and S. Vandeputte: Steel Res., 2001. vol. 72 (10), pp. 412-15.

6. E. Girault, A. Metens, P. Jacques, Y. Houbaert, B. Verlinden, and J. Van Humbeeck: Scripta Mater., 2001, vol, 44, pp. 885-92.

7. P.J. Jacques, E. Girault, A. Mertens, B. Verlinden, J. van Humbeek, and F. Delannay: ISIJ Int., 2001. vol. 41, pp. 1068-74.

8. M. De Meyer, D. Vanderschueren, and B.C. De Cooman: ISIJ Int., 1999, vol. 39, pp. 813-22.

9. W. Bleck, E. Kechagias, J. Ohlert, J. Christen, A. Moulin, H. Hofmann, and T.W. Schaumann: Optimisation of Microstructure in Multiphase Steels Containing Retained Austenite, Final Report, Technical Steel Research Series, RFCS Publications, Luxembourg, 2004, pp. 29-73.

10. E. Emadoddin, A. Akbarzadeh, and G. Daneshi: Mater. Charact., 2006, vol. 57, pp. 408-13.

11. W. Bleck: Proc. Int. Conf. TRIP-Aided High Strength Ferrous Alloys, Gent, Steel GRIPS, Bad Harzburg, Germany, 2002, pp. 13-23.

12. M. Bouet, J. Root, E. Es-Sadiqi, and S. Yue: Mater. Sci. Forum, 1998, vols. 284-286, pp. 319-26.

13. S. Jiao, F. Hassani, R.L. Donaberger, E. Essadiqi, and S. Yue: ISIJ Int., 2002, vol. 42, pp. 299-303.

14. C. Capdevila, F.G. Caballero, and C. Garcia De Andres: ISIJ Int., 2005, vol. 45, pp. 238-47.

15. Quantitative Measurement and Reporting of Hypoeutectoid Carbon and Low-Alloy Steel Phase Transformations, ASTM Standard A- 1033-04, ASTM, Philadelphia, PA, 2004.

16. E. Girault, P. Jacques, P. Harlet, K. Mols, J. Van Humbeeck, E. Aernoudt, and F. Delannay: Mater. Charact., 1998, vol. 40, pp. 111-18.

17. F.S. LePera: J. Met., 1980, vol. 32, pp. 38-39.

18. E. Wirthl, A. Pichler, R. Angerer, P. Stiaszny, K. Hauzenberger, Y.F. Titovets, and M. Hackl: Proc. Int. Conf. TRIP- Aided High Strength Ferrous Allovs, Gent, Steel GRIPS, Bad Harzburg. Germany, 2002, pp. 61-64.

19. J.Z. Zhao, C. Mesplont, and B.C. De Cooman: ISIJ Int., 2001, vol. 41, pp. 492-97.

20. L.P. Karjalainen, T.M. Maccagno, and J.J. Jonas: ISIJ Int., 1995, vol. 35, pp. 1523-31.

21. D.Q. Bai, S. Yue, W.P. Sun, and J.J. Jonas: Metall. Trans. A, 1993. vol. 24A, pp. 2151-59.

22. H. Palkowski, M. Soliman, and G. Kugler: Proc. TMS Ann. Mtg. Exhib., Orlando, FL. TMS, Warrendale, PA, 2007.

23. X.D. Wang, B.X. Hunag, L. Wang, and Y.H. Rong: Metall. Trans. A, 2008, vol. 39A, pp. 1-7.

24. B. Dutta and C.M. Seller: Mater. Sci. Technol., 1987, vol. 3, pp. 197-207.

25. L.P. Karjalainen: Mater. Sci. Technol., 1995, vol. 11, pp. 557-65.

26. R.W. Cahn and P. Haasan: Physical Metallurgy, 4th ed., NorthHolland, Amsterdam, Netherlands, 1996, p. 1601.

27. H.K.D.H. Bhadeshia: Bainite in Steel, 2nd ed., IOM Commercial, Ltd., London, 2001, pp. 122-26.

28. J.R. Bradley, J.M. Rigsbee, and H.I. Aaronson: Metall. Trans. A, 1977, vol. 8A. pp. 323-33.

29. H. Beladi, G.L. Kelly, A. Shokouhi, and P.D. Hodgson: Mater. Sci. Eng. A, 2004, vol. 371A, pp. 343-52.

30. J.B. Qu, Y.Y. Shan, M.C. Zhao, and K. Yang: Mater. Sci. Technol., 2002, vol. 18, pp. 145-50.

31. B. Eghbali and A. Abdollah-zadeh: J. Mater. Process. Technol., 2006, vol. 180, pp. 44-48.

32. F.C. Garcia de Andres, F.G. Caballero, C. Capdevila, and L.F. Alvarez: Mater. Charact., 2002, vol. 48, pp. 101-11.

33. W.C. Jeong and C.H. Kim: J. Mater. Sci., 1985, vol. 20. pp. 4392-98.

34. M. De Meyer, J. Mahieu, and B.C. De Cooman: Mater. Sci. Technol., 2002. vol. 18, pp. 1121-32.

35. M. Soliman and H. Palkowski: ISIJ Int., 2007, vol. 47. pp. 1703-10.

36. M.D. Geib, D.K., Matlok, and G. Krauss: Metall. Trans. A. 1980. vol. 11, pp. 1683-89.

37. M.K. Manna and P.C. Chakraborti; Trans. Ind. Inst. Met., 2004, vol. 57, pp. 149-56.

38. J. Ohlert, W. Bleck, and K. Hulka: Proc. Int. Conf. TRIP- Aided High Strength Ferrous Alloys, Gent, Steel GRIPS, Bad Harzburg, Germany, 2002, pp. 199-206.

39. W. Shi, L. Li, C. Yang, R. Fu, L. Wang, and P. Wollants: Mater. Sci. Eng., A. 2006, vol. 429. pp. 247-51.

40. R. Petrov, L. Kestens, and Y. Houbaert: ISIJ Int., 2001, vol. 41, pp. 883-90.

41. Y. Sakuma, D.K., Matlock, and G. Krauss: Metall. Trans. A, 1992, vol. 23A, pp. 1221-32.

42. S. Hashimoto, S. Ikeda, K. Sugimoto, and S. Miyake: ISIJ Int., 2004. vol. 44, pp. 1590-98.

43. S. Traint, A. Pichler, K. Hauzenberger, P. Stiaszny, and E. Werner: Proc. Int. Conf. TRIP-Aided High Strength Ferrous Alloys, Gent, Steel GRIPS, Bad Harzburg, Germany, 2002, pp. 121-27.

44. S.-J. Kim, C.-G. Lee, I. Choi, and S. Lee: Metall. Mater. Trans. A, 2001, vol. 32A, pp. 505-14.

45. H. Matsuda and H.K.D.H. Bhadeshia: Proc. R. Soc. London, Ser. A, 2004, vol. 460, pp. 1707-22.

46. P. Jacques: Curr. Opin. Solid Slate Mater. Sci.., 2004, vol. 8, pp. 259-65.

47. G.I. Rees and H.K.D.H. Bhadeshia: Mater. Sci. Technol., 1992, vol. 8, pp. 985-93.

48. K. Sugimoto, T. Muramatsu, S. Hashimoto, and Y. Mukaid: J. Mater. Process, Technol., 2006, vol. 177, pp. 390-95.

49. L.B. Timokhina, P.D. Hodgson, and E.V. Pereloma: Metall. Mater. Trans. A, 2004. vol. 35A, pp. 2331-12.

50. L. Barbe. L. Tosal-Martinez, and B.C. De Cooman: Proc. Int. Conf. TRIP-Aided High Strength Ferrous Alloys, Gent, Steel GRIPS, Bad Harzburg, Germany. 2002, pp. 147-51.

51. M.L. Brandt and G.B. Olson: Ironmaking Steelmaking. 1993, vol. 20 (5), pp. 55-60.

52. Y. Sakuma, D.K. Matlock, and G. Krauss: Metall. Trans. A, 1992, vol. 23, pp. 1233-41.

MOHAMED SOLIMAN, Doctor, and HEINZ PALKOWSKI, Professor of Metal Forming, are with the Institute of Metallurgy, Clausthal University of Technology, 38678 Clausthal-Zellerfeld, Germany. Contact e-mail: mohamed.soliman@tu-clausthal.de Manuscript submitted February 24, 2008.

Article published online July 15, 2008

Copyright Minerals, Metals & Materials Society Oct 2008

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

comments powered by Disqus