2024 Volume 64 Issue 2 Pages 326-337
Simulated thermochemical controlled processing (TMCP) was performed on four microalloyed plate steels with Nb and Mo contents varying from 0.03 to 0.045 and 0.03 to 0.15 wt pct, respectively, to investigate influences of both processing and alloying on transformation behavior and microstructural evolution. Dilatometry was performed in situ in a Gleeble® 3500 during thermomechanical simulation to construct continuous cooling transformation (CCT) diagrams for all alloys. A range of cooling rates between 2°C/s and a target 100°C/s along with two deformation levels, −0.4 total true strain and −0.6 total true strain in the austenite regime, were employed to create a range of microstructures. Increased deformation in the austenite non-recrystallization region promoted both polygonal ferrite and acicular ferrite transformation through an increase in nucleation sites. The increase in nucleation sites also resulted in a finer resultant microstructure with increased deformation. Increased cooling rates reduced transformation start temperatures and favored non-polygonal transformation products. Intermediate cooling rates led to the more desirable microstructures consisting of acicular ferrite and bainite. Both Nb and Mo increased the hardenability of the steel through interactions with the polygonal ferrite transformation. Nb and Mo retarded the polygonal ferrite transformation and favored an acicular morphology. Molybdenum alloying also favored bainite transformation. Desirable microstructures of acicular ferrite and bainite were able to be produced with the combination of higher deformation, intermediate cooling rates, and increased Nb/Mo alloying.
Demand continues for steels with an increased combination of high strength, high toughness, and good weldability for applications such as oil and gas pipelines.1,2,3,4) Microalloyed steels are prevalent in structural applications due to the combination of superior mechanical properties and cost-effectiveness.1,2,3,4) Additions of Nb, Mo, Ti, and V provide strengthening primarily through includes grain refinement and transformation strengthening, so that adequate toughness can be maintained. Thermomechanical controlled processing (TMCP) and accelerated cooling of microalloyed steels can produce as-rolled plates with the desired mechanical properties and microstructures without subsequent heat treatments.2,3) Updated cooling systems allow for a larger range of cooling rates, and therefore microstructures, that can be achieved as well as improved flatness control.5,6) With TMCP, significant deformation in the austenite non-recrystallization regime conditions the austenite leads to a ‘pancaked’ microstructure, resulting in increased boundary area as well as deformation substructure within the austenite grains that act as nucleation sites for ferrite transformation,7) which allows for refinement of the final microstructure.8) The accelerated cooling step serves to avoid the polygonal ferrite transformation region and retard phase transformation to lower temperatures to produce fine microstructures.1,2,9)
High strength line-pipe steels, such as the American Petroleum Institute (API) X70 and X80 grades, employ microstructures of acicular ferrite (AF) and bainite (B) to meet strength and toughness requirements.2,10) Acicular ferrite is desirable due to its interlocking microstructure that resists crack propagation and provides increased strength and toughness.11,12,13) The transformation to AF is displacive, leading to supersaturation of carbon in the AF plates. This excess carbon is then rejected into the remaining austenite.14) The carbon-enriched austenite then transforms into pearlite (P), bainite (B), or martensite (M) depending on the cooling rate. The transformation of austenite during cooling is a competition between the formation of polygonal ferrite (PF), a microstructure with low dislocation density and equiaxed grains that nucleate intergranularly on grain boundaries, and AF, which nucleates intragranularly. The most common nucleation site for AF is inclusions within austenite grains.13,15,16) Bainite also forms at intermediate transformation temperatures. Bainite forms in sheaves made up of either plates or laths with a high dislocation density of low angle boundaries between laths. In some morphologies, the low angle boundaries can be difficult to etch and may result in a featureless appearance in light optical micrographs. The type of bainite is generally defined by the presence and location of carbides in the microstructure. The two classical forms of bainite are upper and lower bainite. Upper bainite (UB) contains interlath carbides, while lower bainite (LB) contains intralath carbides. Another common type is granular bainite (GB) which has a matrix of bainitic ferrite laths with islands of M/A dispersed throughout. The ferrite morphology may appear acicular or more featureless.9,17,18,19,20)
Alloying elements for microalloyed steels pertinent to the present work include Nb and Mo. Each microalloying element plays a significant role in achieving desired mechanical properties and microstructures after TMCP through effects on recrystallization kinetics, hardenability, and secondary precipitation strengthening. Niobium and Mo delay static recrystallization through the mechanism of solute drag. Both Nb and Mo have larger atomic sizes than Fe, causing a tendency to segregate to grain boundaries.21) This causes a drag effect on the movement of grain boundaries. Niobium also forms fine strain-induced precipitates which can pin grain boundaries, slow defect recovery, and cause incomplete recrystallization, which assists in strain accumulation.1,7,21,22,23,24,25,26) Solute effects also influence the hardenability of the steel. Along with the solute drag effect, solute Nb and Mo lowers the activity of carbon.24,25,27,28,29) In the case of Mo, this promotes bainite formation. It has also been suggested that Nb and Mo lower the interfacial energy of austenite/austenite grain boundaries, which lowers the driving force for nucleation.18) Thus, the effect of Mo on Nb precipitation causes a synergistic effect on the delay of the austenite to ferrite phase transformation.
The effects of Nb, Mo, and deformation individually on austenite recrystallization and decomposition behavior have been widely investigated, especially with respect to alloying effects. Deformation is often studied by comparing recrystallized and deformed austenite and is sometimes less focused on the extent of deformation. Effects on phase or constituent morphology, rather than recrystallization and phase transformation behavior, are often less emphasized. Studies of alloying effects that study the effects of both differing alloy levels as well as differing deformation levels on microstructural development, rather than recrystallization, are less common. The current study utilizes simulated thermomechanical processing and accelerated cooling to examine both processing and alloying effects on phase morphology and transformation in microalloyed steels with the goal of achieving a greater understanding of deformation, alloying, and hardenability relationships to inform future rolling and cooling processes.
Table 1 shows the alloy compositions of the steels used in this study. The ‘base’ alloy is an industrially processed X70 line pipe steel containing 0.045 wt pct Nb and 0.072 wt pct Mo. Niobium and Mo levels were then varied independently in a set of laboratory heats to study specific alloying effects. A ‘low Mo’ composition was produced with 0.03 wt pct Mo, a ‘low Nb’ was produced with 0.03 wt pct Nb, and a ‘high Mo’ was produced with 0.14 wt pct Mo.
| Alloy | C | Mn | Si | Ni | Cr | Mo | Ti | Nb | V | Al | N | S | P | Cu |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Base | 0.051 | 1.48 | 0.24 | 0.14 | 0.14 | 0.072 | 0.015 | 0.045 | 0.022 | 0.032 | 0.0063 | 0.0008 | 0.008 | 0.17 |
| Low Mo | 0.049 | 1.47 | 0.24 | 0.14 | 0.14 | 0.030 | 0.015 | 0.046 | 0.019 | 0.060 | 0.0057 | 0.0015 | 0.008 | 0.16 |
| Low Nb | 0.052 | 1.44 | 0.25 | 0.14 | 0.14 | 0.071 | 0.013 | 0.031 | 0.024 | 0.023 | 0.0052 | 0.0013 | 0.007 | 0.18 |
| High Mo | 0.044 | 1.46 | 0.24 | 0.14 | 0.14 | 0.142 | 0.013 | 0.043 | 0.022 | 0.026 | 0.0067 | 0.0014 | 0.006 | 0.16 |
Hot compression testing was completed in a Gleeble® 3500 with the ISO-T® anvil assembly.30,31) Material was machined into cylindrical samples with a 10 mm diameter and a length of 15 mm. Temperature was measured and controlled using K type thermocouples spot welded onto the surface of the sample. Testing was performed under a vacuum of less than 1333 Pa. A two-step deformation path was chosen to simulate thermomechanical controlled rolling. Samples were reheated to 1250°C at a rate of 5°C/s and held for 10 min. The non-recrystallization temperature (Tnr) of the base alloy was previously determined to be 1044°C, so deformation temperatures were chosen to be above and below this temperature. An initial deformation pass was performed at 1075°C with an applied true strain of −0.2 at a rate of 1 s−1. An additional deformation pass was then performed at 850°C with an applied true strain of either −0.2 (low deformation condition) or −0.4 (high deformation condition) at a rate of 1 s−1. The low deformation condition thus had a total of −0.4 true strain while the high deformation condition had a total of −0.6 true strain before cooling. Accelerated or controlled cooling was used to cool the samples to room temperature at a target rate of 2, 10, 20, 30, 50, or 100°C/s with gas quenching; however, some cooling rates were unable to be met with the quench conditions and will be reported as the average cooling rate measured from 800 to 500°C. Dilatometry testing was performed in situ during Gleeble® 3500 hot compression testing. A contact dilatometer with a linear variable differential transducer (LVDT) measuring head was placed around the sample’s diameter prior to straining to measure diametral changes during testing. Phase transformations were indicated by non-linear diametral changes. Dilation was measured relative to the undeformed sample dimensions. The transformation temperatures were then determined using temperature versus relative dilation curves. Raw data was smoothed using a moving average filter for both temperature and relative dilation fluctuation before transformation temperatures were determined using the first derivative method.32,33) The transformation temperatures were then used to construct CCT diagrams. A replicate was performed for each condition.
All samples used for microscopy were sectioned parallel to the compression axis, mounted in Bakelite, and prepared using standard metallographic procedures and polished to 1 μm. The Vickers microhardness of all samples was measured with a load of 300 g. 9 indents were taken and averaged for each condition. Samples for light optical microscopy (LOM) were etched using a 2 pct nital solution and imaged using an Olympus DSX500. Additional observation of selected microstructures included electron backscatter diffraction (EBSD). Samples were prepared using standard metallographic techniques, then finished by vibratory polishing in a colloidal silica solution. EBSD was completed on an FEI Helios NanoLab 460F1 with an accelerating voltage of 20 keV and a current of 5.5 nA. A single scan at 500x magnification was analyzed for each condition. EBSD data was first cleaned in OIM® analysis software using the ‘Neighbor CI Correlation’ method at a minimum confidence index (CI) of 0.1. EBSD data was then transferred to Matlab® where the MTEX toolbox was used for processing. Grains of less than 5 pixels were removed and boundaries of less than 1.5° misorientation were not included in further analysis. To determine effective grain size, only boundaries greater than 15° misorientation were considered. Grain size was measured as a diameter and determined by measuring the longest distance between any two vertices of the grain boundary. All imaging, microhardness testing, and EBSD was done in the center of the samples.
The CCT diagram constructed for the base alloy for the low deformation (−0.4 total true strain) condition is shown in Fig. 1 and corresponding microstructures are shown in Fig. 2. The presence of polygonal ferrite (PF) in the microstructure was only detected at the slowest cooling rate, 2°C/s, along with quasi-polygonal ferrite (QF) and bainite (B). Each of the constituents have distinct etching responses, shown in Fig. 2(a). The PF regions have a light etching response and are equiaxed with no internal substructure.16) The QF regions are identified by their irregular boundaries.16) The B regions have a lath-like morphology and some visible substructure.16) As stated previously, B is also characterized by the presence and location of carbides. Carbides are not easily observed in LOM/SEM analysis of these microstructures, so distinction between bainitic constituents will be made here based on the morphology of the laths and M/A constituent. At 8°C/s, Fig. 2(b), the microstructure is a mixture of QF and B. At 13°C/s, Fig. 2(c), acicular ferrite (AF) is present in the microstructure along with QF and B. The AF regions are distinguishable by their darker etching response, non-parallel laths, and irregular shape.16) Acicular ferrite also contains M/A islands. Figure 3(a) shows an SEM micrograph of the 13°C/s to better understand the morphology of the constituents. The QF regions appear as dark, mostly featureless grains with an irregular shape. M/A islands can be identified dispersed among QF grains and within the B packets. At 26°C/s, Fig. 2(d), the microstructure is primarily AF and B. An SEM image is shown in Fig. 3(b), and reveals the more interlocking shape of the AF as well as a slight change in morphology of the B. The M/A constituent with the B regions appears as interlath films, rather than globular islands in the 13°C/s sample. At the highest cooling rates, 42 and 58°C/s, the microstructure became increasingly martensitic, with some B. Martensite, specifically lath martensite, forms with a packet morphology. Carbides are also absent in LOM from this phase.17) The SEM image at this cooling rate is shown in Fig. 3(c). The M/A islands within the B morphology are again interlath films. The M appears as dark, nearly featureless grains.
Beyond LOM and SEM inspection, misorientation distributions from EBSD can help to identify different microstructural constituents. The austenite to acicular ferrite/bainite transformation can be described using the Kurdjumov-Sachs (K-S) and the Nishiyama-Wasserman (N-S) orientation relationships (ORs).34,35,36,37) These ORs produce misorientations within the Bain region, that are below 20° and above 47°.34,35,36,37) Two characteristic peaks in intensity appear between 47° and 60° misorientation with little to no misorientations between 20° and 47° and a large peak in intensity appearing below 20° misorientation as well. It has also been suggested that bainitic microstructures have a higher proportion of low angle (<15°) boundaries than acicular ferrite microstructures due to the sub-grain structure of bainitic packets.36) Additionally, acicular ferrite microstructures have been suggested to have higher proportions of high angle boundaries due to the lack of sub-grains within acicular ferrite grains. Any misorientations that lie outside the range between 20° and 47°, are indicative of primary ferrite, such as PF and QF, or prior austenite grain boundaries as these constituents produce misorientations across the entire distribution.10,35,36,37) In this study only the samples cooled at intermediate target rates of 20, 30, and 50°C/s were analyzed with EBSD. These intermediate cooling rates were chosen as they closely resemble cooling rates which are attained in industrial processing and therefore produce industrially relevant microstructures. Each sample was analyzed using a single large scan of the representative microstructure. Figure 4 show inverse pole figure (IPF) maps and misorientation distribution histograms for the 3 conditions. Each misorientation distribution shows a distinct peak in frequency below 20°, near-50°, and near-60° misorientation, as expected for acicular ferrite and bainitic microstructures. There is also a small distribution of misorientations across 20–47°, likely due to a small amount of QF present in the structure. Gourgues et al. has hypothesized that lower transformation temperature non-polygonal microstructures, such as lower bainite (LB) and M, produce stronger near-60° misorientation peaks typical of a near-K-S OR, while higher transformation temperature non-polygonal microstructures, such as granular bainite (GB) and AF, produce stronger near-50° misorientation peaks typical of a near-N-W OR.35) Zajac et al. reported a similar conclusion from their work. They found that GB has a broad peak around 45° misorientation, while LB has a high proportion of boundaries above 50° misorientation.20) In the current work, along the studied range of cooling rates, the microstructure shifted from predominantly AF and B with globular M/A constituents, Figure 3(a), to a microstructure of predominately AF and B with interlath M/A films, Fig. 3(b), to a microstructure predominately of B with interlath M/A films and some M present, Fig. 3(c). The sample cooled at 13°C/s (target rate of 20°C/s), Fig. 4(a), had all 3 characteristic peaks in misorientation; however, there was less distinction between the peaks in frequency of the near-50° and near-60° misorientation when compared to the other samples. The lack of distinction could be due to the increased presence of AF in the microstructure. Along with the microstructural progression between the samples, the peak in frequency of the near-60° misorientation also grew in relative frequency as cooling rate increased. The sample cooled at 42°C/s (target rate of 50°C/s), Fig. 4(c), had the highest peak in frequency of the near-60° misorientation of the samples at a frequency of 0.1, compared to a frequency of ~0.05 for the other samples. The appearance of martensite in the microstructure is likely responsible for this shift as the microstructure of the 26°C/s sample (target rate of 30°C/s), Fig. 4(b), does not contain M and the peak in frequency of the near-60° misorientation has a lower relative frequency. However, the peak in frequency of the near-50° misorientation remains relatively the same over all samples, regardless of the microstructure changes, suggesting that both B and AF have near N-W ORs. All samples also have a large peak in frequency at less than 20° misorientation of similar relative frequency, indictive of either N-W or K-S ORs. The less than 20° peak in frequency is expected as all microstructures contained some amount of a bainitic or martensitic structure, which have numerous sub-grain boundaries of low misorientation.10,17,18,19)
The CCT diagram constructed for the high deformation (−0.6 total true strain) condition of the base alloy is shown in Fig. 5 and the resultant microstructures in Fig. 6. At 2°C/s, Fig. 6(a), the microstructure consists of PF and QF along with AF. Polygonal ferrite is barely present at 10°C/s, Fig. 6(b), and QF and AF are still present along with B. At 14°C/s, Fig. 6(c), the microstructure consists mainly of non-polygonal transformation products, namely AF and B, with some QF still present. The SEM micrograph, Fig. 7(a), shows a morphology of B with interlath M/A films and some globular M/A islands. The QF regions can be identified by the dark appearance and presence of M/A islands while AF is distinguished by the interlocking structure. The microstructure is similar at 26°C/s, Fig. 6(d), and with a shrinking QF fraction. The SEM images in Fig. 7(b) and show films of interlath M/A, with some globular M/A islands still present. At 40°C/s, Fig. 6(e), M is found in the microstructure. The SEM image, Fig. 7(c), show dark, relatively featureless M and a mixture of B with interlath M/A films and globular M/A islands. At 51°C/s, Fig. 6(f), the AF, B, and M remain, however, without noticeable QF present.
The CCT diagram for the low deformation condition of the low Mo alloy is shown in Fig. 8(a). At 2°C/s the microstructure consists of a mixture of majority PF with some P, Fig. 9(a). At 10 and 14°C/s, there is some PF accompanied by QF and B. At 26°C/s, 40°C/s, and 67°C/s the microstructure consists of a mixture of AF and B.
The CCT diagram for the high deformation condition of the low Mo alloy is shown in Fig. 8(b). At 2°C/s the microstructure consists almost entirely of PF with a small amount of P, Fig. 9(b). At 10 and 14°C/s, PF is still predominant in the microstructure but with increasing QF. At 26°C/s the microstructure consists of entirely non-polygonal constituents, in this case AF and B. This microstructure is maintained at the 41 and 52°C/s cooling rates.
3.3. Transformation Behavior of the High Mo AlloyThe CCT diagram for the high Mo alloy in the low deformation condition is shown in Fig. 10(a). At 2°C/s the microstructure is a mixture of PF, QF and B, Fig. 11(a). At 9 and 19°C/s, the microstructure consists only of B. At 22°C/s and 42°C/s, the microstructure is a mixture of AF and B. At 64°C/s the microstructure is a mixture of B and M.
The CCT diagram for the high Mo alloy in the high deformation condition is shown in Fig. 10(a). At 2°C/s consists of a mixture of PF, QF, and B, Fig. 11(a). At 10°C/s the microstructure is mostly that of AF and B, with some QF. At 18°C/s and 22°C/s, the microstructure consists exclusively of AF and B. At 41°C/s and 56°C/s, the microstructure consists of AF, B and M.
3.4. Transformation Behavior of the Low Nb AlloyThe CCT diagram for the low Nb alloy in the low deformation condition is shown in Fig. 12(a). At 2°C/s, the microstructure is a mixture of PF, QF and B, Fig. 13(a). At 9°C/s, some PF remains along with QF, AF and B. At 17°C/s, there is no longer PF in the microstructure but QF, AF and B remains. At 25°C/s, 38°C/s, and 59°C/s, the microstructure is a mixture of AF and B.
The CCT diagram for the low Nb alloy in the low deformation condition is shown in Fig. 12(a). At 2°C/s, the microstructure consists predominantly of PF, with some QF and a small amount of P, Fig. 13(b). At 10°C/s, some PF and QF remains in the microstructure, but the majority is made up of AF and B. At 18°C/s, only some QF remains along with AF and B. From 19°C/s to 52°C/s, some QF remains, and the microstructure consists mainly of AF and B.
There are a few notable differences between the microstructural development of the high and low deformation conditions of the base alloy. In both conditions PF is only visible in the microstructure at low cooling rates. For the low deformation conditions, PF was visible only at 2°C/s and for the high deformation condition there was some PF visible up to 10°C/s. With the increased deformation, the transformation start temperature of PF at 2°C/s rose from 630°C for the low deformation condition to 645°C for the high deformation condition. This increase resulted in a higher percentage of PF in the final microstructure of the high deformation condition, Fig. 6(a). The QF percentage in the microstructure also increased with the additional deformation. Polygonal ferrite nucleates intergranularly on austenite grain boundaries and deformation increases the amount of nucleation sites in two different ways. As the austenite grains are deformed, they become elongated or ‘pancaked’ which increases the grain boundary area, and therefore nucleation sites. Deformation also creates substructure (e.g. deformation bands) within austenite grains, which can act as nucleation sites for PF.1,9,21,22,23) Increased deformation also increased the amount of AF formation. In the high deformation condition, AF was present across the entire range of cooling rates. The bainitic microstructures are also markedly finer with shorter laths in the high deformation condition. The martensitic transformation also differed slightly between the low and high deformation conditions. While M is visible in the two highest cooling rates for the low and high deformation conditions, it is less prevalent in the high deformation condition, likely due to accelerated ferrite nucleation.
To further study the effects of deformation, EBSD was performed on the base alloy samples cooled at 20, 30, and 50°C/s target cooling rates in the high deformation condition. As with the low deformation condition, a single large scan was used for analysis. Figure 14 show IPF maps and misorientation distributions for these samples. In all cases the characteristic peaks in frequency of less than 20°, near-50°, and near-60° misorientation remained consistent. However, there are a few key differences in the misorientation distributions noted with increased deformation. Across all cooling rates of the high deformation condition, the frequency of the 20–47° misorientations increased. This indicates that QF persists across these cooling rates, though in relatively small amount compared to the other constituents at the higher cooling rates. The presence of QF is especially prevalent at 14°C/s which can be observed both in the LOM image, Fig. 7(a), and the misorientation distribution, Fig. 14(a). In the low deformation condition, the peak in intensity of the near-60° misorientation increased significantly at the higher cooling rates. In the high deformation condition this effect is less pronounced, only a slight increase in the intensity of the near-60° peak in disorientation with each successive cooling rate is observed. The sample cooled at 26°C/s, Fig. 7(b), has a microstructure predominantly of AF and B with interlath M/A constituent. Compared to the 14°C/s sample, the 26°C/s sample had a larger relative near-50° and near-60° peak in misorientation, Fig. 14(b), due to the decrease of QF, and increase of AF and B in the microstructure. At 40°C/s, a small amount of M is visible in the structure, Fig. 7(c), which is consistent with the increase in intensity of the near-60° peak in misorientation, Fig. 14(c).
In addition to microstructures, cooling rate also had an effect on mechanical properties. Figure 15 shows a plot of Vickers microhardness versus cooling rate for the high and low deformation conditions for the base alloy. For each condition, there are large increases in microhardness with increasing cooling rate. Both conditions show a significant increase in microhardness following cooling at ~20°C/s, corresponding to the appearance of the interlath M/A film morphology in the B. From 2°C/s to ~20°C/s, the hardness values between the high and low deformation conditions are similar, with the high deformation condition having marginally higher microhardness. As discussed previously, increased deformation led to an increase in nucleation sites, promoting both PF and AF transformation. As a result, the high deformation samples had finer microstructures.1,9,21,22,23) Table 2 shows the average grain sizes for the low and high deformation conditions at the 20, 30, and 50°C/s target cooling rates. On average, there was a 2 μm decrease in grain size for the higher deformation condition. The behavior between the high and low deformation conditions differs beyond ~30°C/s. For the low deformation condition, the microhardness increases with further increases in cooling rate due to a significant amount of martensite appearing in the microstructure at 42°C/s and beyond. For the high deformation condition, there is a plateau in microhardness as the microstructural constituents do not change significantly over these cooling rates. The amount of martensite that formed in the high deformation condition is relatively low, as reflected in the grain boundary misorientation distributions shown in Fig. 14(c), so the hardness did not increase appreciably. Regardless of deformation condition, increased cooling rates are shown to favor non-polygonal transformation products. For this alloy, the cooling rate must exceed 2°C/s to avoid PF formation with increased deformation. The combination of both intermediate cooling rate and high deformation resulted in a fine, non-polygonal microstructure with greater microhardness.
| Condition | 20°C/s target | 30°C/s target | 50°C/s target |
|---|---|---|---|
| Low Def. | 13.7 μm | 14.7 μm | 13.2 μm |
| High Def. | 11.5 μm | 12.1 μm | 11.8 μm |
The most significant impact of additions of Mo is the presence of PF in the microstructure. For the base alloy, PF was only visible in small quantities and only at the lowest cooling rate of 2°C/s for both the low and high deformation conditions. In the low Mo alloy, both PF and QF are predominant in the microstructure up to 20°C/s in the low deformation condition and 26°C/s for the high deformation condition. The transformation start temperatures at 2°C/s are significantly different as well. Figure 16 shows the transformation start temperatures for each of the alloys in the low and high deformation condition. For the low deformation condition, transformation begins at 630°C in the base alloy, 658°C in the high Mo alloy, and 732°C in the low Mo alloy. For the high deformation condition, transformation begins at 654°C in the base alloy, 668°C in the high Mo alloy, and 739°C in the low Mo alloy. The difference in transformation start temperatures, specifically between the base alloy and low Mo alloy, indicates that additions of Mo lower transformation start temperatures and favor non-polygonal transformation products. Solute Mo is known to segregate to austenite grain boundaries, creating a solute drag effect and suppressing the formation kinetics of ferrite,1,9,21,22,23,28,29) making PF formation less favorable with higher Mo content. It should be noted that the high Mo transformation start temperature was higher than that of the base alloy. This behavior was unexpected and might be associated with prior austenite grain size (PAGS) and a slightly greater C content of the base alloy.22,23,38)
Literature indicated that Mo affects the bainite transformation both directly and indirectly.22,25,27,39,40) Additions of Mo directly impact bainitic transformations by lowering the Bs temperature.25) Mo also suppresses ferrite nucleation, therefore promoting B formation.27,39,40) Figures 17 and 18 show LOM and SEM micrographs of the high Mo alloy cooled at 18, 22, and 41°C/s. Apart from AF, all the microstructures at these cooling rates are predominantly bainitic. At 18°C/s and 22°C/s, Figs. 18(a) and 18(b), the B microstructure contains mostly globular M/A constituent, while at 41°C/s, the M/A constituent is present as interlath films in the B microstructure. The appearance of the M/A constituent is similar to the base alloy in the high deformation condition. However, the M/A constituent, both in the globular and film morphology, is markedly finer in the high Mo alloy across all cooling rates. This is likely a result of the influence of Mo on the bainitic temperature range during cooling. Suppression of the Bs temperature also resulted in finer bainitic ferrite laths in the high Mo alloy when compared to the base alloy. The formation of QF is greatly reduced in the high Mo alloy. In the base alloy, QF is present at cooling rates up to 51°C/s, while in the high Mo alloy, QF is present at cooling rates up to 18°C/s. M is also found at slower cooling rates. Martensite was present in the microstructure at 22°C/s in the high Mo compared to 40°C/s in the base alloy. Both instances illustrate that greater additions of Mo increase hardenability. One additional observation in Fig. 18 is the appearance of more ‘fragmented’ B laths, best seen in Fig. 18(b). Previous work by Regier et al. reported that bainitic growth in deformed austenite may be interrupted by dislocation debris and austenite deformation substructure, promoting a fragmented bainite morphology with shorter lath sections having less needle/plate-like character.41)
Like the low Mo alloy, the low Nb alloy has significantly more PF formation at 2°C/s compared to the base alloy, in both deformation conditions. PF formation persists until the cooling rate exceeds 10°C/s for both the low and high deformation condition. The transformation start temperature of the low Nb and the base alloy are also significantly different, shown in Fig. 16. In the low deformation condition, transformation starts at 630°C when cooled at 2°C/s in the base alloy, while the transformation starts at 686°C in the low Nb alloy when cooled at the same rate. In the high deformation condition, transformation starts at 654°C in the base alloy, while the transformation starts at 713°C in the low Nb alloy. This indicates that Nb alloying shifts the polygonal ferrite transformation region to lower cooling rates, causing non-polygonal microstructures to be more favorable. While the effect of Nb and Mo in this case is the same, the mechanisms differ. Niobium has a more complicated effect on transformation of austenite. Like Mo, solute Nb can exhibit a drag effect on austenite boundaries, inhibiting polygonal ferrite transformation. However, during deformation strain-induced precipitates such as NbC and Nb(C,N) can form which can act as ferrite nucleation sites, enhancing the transformation kinetics.21,22,23) The hardenability effect of Nb can therefore be attributed to both solute Nb and precipitated Nb effects. The presence of NbC and Nb(C,N) can also affect the PAGS. In a study by Zurutuza et al., it was found that additions of Nb reduced PAGS significantly, from 12.9 ± 0.8 μm in a 0.15C-1.05Mn-0.002B steel to 5.5 ± 0.2 μm in a 0.16C-1.07Mn-0.002B-0.026Nb steel.21) A lower PAGS will lead to finer microstructures which leads to higher strength and toughness.42)
Vickers microhardness testing was performed on all alloys in the low and high deformation condition to examine the differences in mechanical properties and evaluate microstructural changes, the results of which are shown in Fig. 19. With increasing cooling rate, the microhardness of all alloys generally increased as the microstructures of the alloys shifted from polygonal to non-polygonal microstructures. The base alloy and the high Mo alloy had the highest microhardness over the range of cooling rates due to the increased hardenability with Nb and Mo alloying. In the low deformation condition, Fig. 19(a), there was an average of 6.3 HV0.3 difference between the high Mo and base alloys. The low Nb alloy had lower microhardness values than the base and high Mo alloys with an average of 15.6 HV0.3 difference between the low Nb and the base alloy. The low Mo had the lowest microhardness with an average microhardness 32.7 HV0.3 lower than the base alloy. In the high deformation condition, Fig. 19(b), there was an average of 7.1 HV0.3 difference between the base and high Mo alloy over the range of cooling rates. The low Nb had marginally lower microhardness with an average of 13.2 HV0.3 lower microhardness throughout the range of cooling rates when compared to the base alloy. The low Mo had the lowest microhardness of all the alloys with an average of 39.9 HV0.3 lower microhardness than the base alloy. These trends in microhardness are consistent with the differences in microstructural evolution between the alloys as discussed above. The low Mo alloy had the highest amount of relatively softer polygonal ferrite in the microstructures and polygonal ferrite appeared over the widest range of cooling rates, which resulted in the low microhardness values overall in both the low and high deformation conditions. It should be noted, however, that the hardness of the low Mo alloy at the target cooling rate of 30°C/s appears much higher than expected. The standard deviation of the hardness measurements for the low Mo alloy at that cooling rate is also very high, so it cannot be stated with certainty that the hardness is truly higher than those that follow it. The low Nb alloy showed similar trends but had less polygonal ferrite in the microstructures over a narrower range of cooling rates than the low Mo, resulting in a lowered microhardness when compared to the base but still higher than the low Mo alloy.
Both processing and alloying have distinct effects on microstructural evolution, however, the interactions between processing and alloying are important to understand as well. In the investigation of deformation effects in the base alloy, it was found that increased deformation favored polygonal ferrite transformation and led to finer microstructures. It was also found that Nb and Mo additions retarded polygonal ferrite transformation. The combination of deformation and increased alloying then led to fine, desirable microstructures of AF and B. The high Mo alloy provides an example of this. In the low deformation condition of the high Mo alloy, QF was present in the microstructure only at the cooling rate of 2°C/s, while QF was visible up to 18°C/s, Fig. 8(b), in the high deformation condition. When compared to the high Mo alloy in the high deformation condition, the base alloy in the high deformation condition showed QF in the microstructure in cooling rates up to 41°C/s, Fig. 5. While deformation increased QF formation, the increased Mo content in the high Mo alloy retarded the transformation to slower cooling rates. Microstructures of AF and B can therefore be produced with slower cooling rates that are easier to achieve with industrial cooling equipment. The ‘fragmentation’ of the microstructure in the high Mo alloy is another example of the combined effects of deformation and alloying. The ‘fragmentation’ is believed to result from interruption of bainite growth due to dislocation substructure and debris caused by increased deformation.41) However, this phenomenon was only seen in the high Mo alloy, indicating that the increased Mo content may have led to greater strain accumulation in the microstructure.
Deformation and Nb alloying also have combined effects. It was concluded that both PF and AF formation increased with deformation, but that Nb retarded the PF transformation. It is also likely, however, that Nb content could increase AF formation as well. The low deformation condition of the base alloy consists of PF, QF and B, Fig. 2(a), when cooled at 2°C/s. In the high deformation condition of the base alloy, the microstructure consists of PF and QF like the low deformation condition, but rather than B this condition contains AF, Fig. 6(a), when cooled at 2°C/s. At 2°C/s, the low Nb alloy has a microstructure of PF, QF, and B, in the low deformation condition, Fig. 13(a), and PF, P, and QF in the high deformation condition Fig. 13(b). It is clear in the base alloy that the increased Nb led to increased AF formation when paired with increased deformation. In the low Nb alloy, the increased deformation only led to increased PF and QF formation, and no AF was visible. This indicates that increased Nb content leads to enhanced AF formation likely due to increased Nb precipitates that act as nucleation sites for AF.
The current work examined the changes in transformation behavior induced by prior deformation and Nb/Mo alloying through simulated thermomechanical heat treatments and accelerated cooling in 4 different alloys. Two different deformation levels were utilized along with a range of cooling rates to study the differences in microstructural development of a base alloy. EBSD analysis was also applied to selected conditions to further examine effects of deformation. Additionally, 3 laboratory heats were designed in which Nb and Mo levels were adjusted independently to examine alloying effects with the high deformation condition and a range of cooling rates applied. The following conclusions were found:
(1) At intermediate cooling rates, the base alloy exhibited a range of non-polygonal constituents such as acicular ferrite, bainite and martensite. It is interpreted from the misorientation distributions that these microstructures followed near-K-S and near-N-W ORs.
(2) Higher deformation of the base alloy led to an expansion of the polygonal ferrite transformation region while also favoring acicular ferrite transformation at slower cooling rates.
(3) Higher cooling rates resulted in lower transformation start temperatures which favored non-polygonal transformation products.
(4) Additions of Mo were found to shift the polygonal ferrite transformation region to slower cooling rates, favoring non-polygonal transformations and increased hardenability.
(5) Additions of Mo led to finer ‘fragmented’ bainitic ferrite morphologies and M/A constituents.
(6) Additions of Nb were found to shift the polygonal ferrite transformation region to slower cooling rates, favoring non-polygonal transformations and increased hardenability.
This research was funded by the sponsors of the Advanced Steel Processing and Products Research Center, an industry/university cooperative research center. The authors gratefully acknowledge Dengqi Bai (SSAB), Matt Merwin (US Steel), Brian Lin (ArcelorMittal), and Matt Enloe (Steel Dynamics) for their guidance and feedback on this work. The authors would also like to extend a special thanks to Dengqi Bai (SSAB) and Matt Merwin (US Steel) for providing the experimental materials.
