Materials and Methods
Low-carbon steel with a chemical composition of 0.12% C, 0.78% Mn, 0.18% Si, 0.020% Al, 0.0011% P, 0.018% S, and 0.0079% N microalloyed with 0.048% Nb was selected for this study. Two variants of industrial thermomechanical processing were implemented: high temperature (HT), in which the start of the final deformation occurred at 1020 °C, and low temperature (LT), in which the start of the final deformation occurred at 987 °C. The heating and austenitization were performed at 1250 °C for 3.1 h. The pre-deformation was carried out in the temperature range 1180–1040 °C with a total reduction of 92.25%.
Austenitizing and pre-deformation were completely identical in both variants. The degrees of deformation in all passes of the final deformation were also almost identical. In the LT variant, the final deformation was performed in the temperature range 987–887 °C, with five passes resulting in a total reduction of 83.10%. In the HT variant, the final deformation was also performed with five passes resulting in a total reduction of 82.98%, but in the temperature range 1020–905 °C. The strip was cooled to 560 °C and reeled. The temperature was measured using an optical pyrometer. A schematic of the final deformation is shown in Figure 1.
Samples for testing were taken at characteristic spots from each phase of thermomechanical treatment. Sampling was carried out in such a way that the strip was stopped at a chosen moment. Then, the samples of interest from the deformation zone and its vicinity were rapidly cooled by applying the maximum amount of cooling water onto the rollers. In this way, the structural development and the processes involved in each phase of thermomechanical finishing were blocked. Reductions in the samples were determined using various methods of analysis. For electrochemical isolation, samples in the form of a plate with dimensions 80 mm × 20 mm were electrolytically dissolved in a cell according to the method described by W. Koch 
. The isolate was identified using an atomic absorption spectrophotometer (AAS, Perkin Elmer, Waltham, MA, USA). The size and distribution of niobium precipitates, as well as their interaction with dislocations, were analyzed using a transmission electron microscope (TEM, JEOL-2000, Tokyo, Japan). For analysis by TEM, samples were prepared using the foil method. For that purpose, samples in the form of a cylinder with a diameter of 3 mm were used. They were mechanically ground on 600 and 1000 grit papers and then electrolytically thinned to a fully transparent foil. Electrolytical thinning was performed in a solution of 5% HClO4
and 95% CH3
COOH using a voltage of 45 V at room temperature. The samples prepared in this way were analyzed using a transmission electron microscope at magnifications of 5000× and 200,000× and voltage of 200 kV. To determine the internal stresses in the material and the dislocation density, the X-ray diffraction method was used. Analyses were performed using a universal camera, a vertical diffractometer with Bragg–Brentano’s focusing, and a graphite monochromator using Co Kα radiation (λ = 1.790 Å). Samples with dimensions 10 mm × 10 mm were analyzed after metallographic preparation by grinding, polishing, and deep etching.
The microstructure of metallographically prepared samples after etching in nital was observed using a light microscope (Leitz Ortholux, Wetzlar, Germany), and images were taken using an Olympus digital camera. The appearance and propagation of the Lüders bands were analyzed by tensile testing, up to breaking point, on a Zwick 50 kN universal tensile testing machine (Zwick, Ulm, Germany). Simultaneous measurements of the temperature change in the deformation zone were taken using a VarioCAM M82910 thermographic camera (Jenoptik, Jena, Germany) with a temperature sensitivity of 80 mK. The measured results were analyzed using Irbis professional software (InfraTec GmbH, Dresden, Germany).
Results and Discussion
Figure 2 shows that in the pre-deformation in both variants, the degree of precipitation was low (11.8–12.1%) and did not change significantly from the onset to the end of pre-deformation. In the final deformation, there was an increase in the degree of precipitation in both variants. Nevertheless, it was slightly lower in the HT variant (25%) than in the LT variant (26%). The precipitation continued during the cooling phase of the strip and was higher for the high temperature variant (36.5%) as compared to low temperature variant (34%). Since in both variants the steels were heated under the same conditions to the same temperature (1250 °C), it was assumed that the achieved solubility of niobium was the same in both variants.
Transmission electron analysis of the pre-deformed samples showed niobium precipitates (Figure 3). These precipitates were coarse, of spherical shape, and larger than 20 nanometers and were heterogeneously distributed in structure.
In principle, precipitation of Nb carbides can occur in austenite or ferrite on grain boundaries, dislocations, subgrain boundaries, or interphase boundaries during the austenite-to-ferrite transformation 
In the final phase of thermomechanical treatment in both variants, the samples were examined after each pass (Figure 4a). The degree of precipitation in the HT variant was lower than that in the LT variant, and for the first three passes, they remained low and were similar for both variants (11.8–12.1%).
Although it is known from the literature  
that strain-induced precipitation is a function of temperature, deformation, and time, the results obtained in this study (shown in Figure 4) imply that SIP is more temperature dependent. In particular, in the low temperature variant, intensive, strain-induced precipitation started after the second pass (at 963 °C). In the high temperature variant, strain-induced precipitation began after the third pass (at 960 °C). Figure 4b clearly shows that the onset of intense strain-induced precipitation is more temperature dependent than the degree of deformation (Figure 4a). Strain-induced precipitation begins at a temperature of about 960 °C. It was found that, until the onset of strain-induced precipitation in both variants, dynamic and static recrystallization proceeded undisturbed. This is very important in the sense of the determination of hot rolling parameters. At 930 °C, the degree of precipitation was the same in both variants. In the fourth pass and in the fifth pass, there were significant increases in the degree of precipitation, amounting to 19% and 25%, respectively. Figure 4b clearly shows that with the decrease in temperature, the degree of precipitation increased in the LT as well as in the HT variant.
The beginning of strain-induced precipitation was also confirmed by transmission electron microanalysis (Figure 5). The strain-induced precipitates were very small, below 10 nanometers, and arranged in rows. The fraction of strain-induced precipitates and their average size are directly influenced by the amount of deformation 
In the last passes of the final thermomechanical processing in both variants, the degree of total deformation decreases (Figure 6) while dislocation density increases, as is the case with the degree of precipitation (Figure 4a). In investigations of the same steels 
it was proven that the low temperature rolling schedule leads to considerably higher rolling forces in the last three passes.
In order to determine the quantitative amount of changes in the structure, a study of the diffraction line profile was carried out. By studying the diffraction line profile in each pass of the HT and LT thermomechanical processing, a significant difference was observed both from the beginning to the end of the thermomechanical process and in each pass in both variants. From the profile of diffraction lines, the density of dislocations, ϕ (m−2), was calculated according to Equation (1).
The onset of strain-induced precipitation depended on the temperature, as is shown in Figure 7. In the first two passes, the degree of deformation was high and almost identical for both variants. Static and dynamic recrystallization took place undisturbed, so the dislocation density was lower. When strain-induced precipitation began at 960 °C, the recrystallization became inhibited and the dislocation density increased. In the last passes, the degree of deformation was small, but the density of dislocations increased as the degree of precipitation increased. Precipitates arranged in close-packed arrays became a strong obstacle to the movement of dislocations. Therefore, dislocations became immobile and their density increased.
Figure 7a shows that in the LT variant, the degree of precipitation almost linearly increased with the decrease in temperature after two passes. Dislocation density at the onset of deformation increased, but after the third pass, there was an evident slowdown, as can be seen in the values. Figure 7b shows similar tendencies for the HT variant, with the only difference being that a significant increase in the degree of precipitation and deformation occurred after the third pass. Further, it can be clearly seen that a change in the density of dislocations was associated with a change in the degree of precipitation. At the beginning of strain-induced precipitation, interaction of the precipitate and dislocations occurred and the dislocation density increased, as is shown in Figure 8.
After the third pass, the major influence on the dislocation density was shown to be the degree of precipitation rather than the degree of deformation. Specifically, in microalloyed steels, recrystallization and precipitation can interact in at least three distinct ways 
. Firstly, a decrease in dislocation density through recrystallization reduces the number of precipitate nucleation sites available and can thus retard the onset and rate of precipitation; this occurs after the onset of intense deformation-induced precipitation (960 °C) and is evident in the first two passes in the LT variant, as well as in the first three passes in the HT variant. Secondly, precipitate dispersion can provide a pinning (Zener) force that can retard or halt the progress of recrystallization; this is evident in the latest passes of final TMCP, because by increasing the degree of precipitation, the density of blocked dislocations also increases, as shown in Figure 7. Precipitates are arranged in rows at the onset of SIP (Figure 5), but after it, they are dispersed, accumulating in certain places and blocking the dislocations (Figure 8). Thirdly, the mobility of grain boundaries is strongly affected by the solute content of the matrix and its diffusivity. Particularly, the diffusivity of Nb is an important factor contributing to the significant retarding effect on recovery and recrystallization; it slows down boundary motion since it interacts with the boundary and does not diffuse rapidly enough to keep up with the migrating boundary. The higher degree of precipitation leads to the decrease in the amount of free Nb that can diffuse. Furthermore, as a result of the larger atomic radius of Nb than those of carbon, nitrogen, and iron, Nb moves more slowly; thus, it cannot track the grain boundary, and its diffusivity decreases. Further, the dislocation density increases even though the energy of the system and the temperature decrease, and so the total deformation slows down. Therefore, close-packed carbonitride rows and interaction between precipitates and dislocations are present in the structure (Figure 9). The further increase in deformation and precipitation also increases the density of dislocations, as was also observed in 
. Dislocations pile up on rows of precipitates and form the boundary of the subgrain (Figure 9).
In the last passes, the recrystallization and recovery mechanisms were halted and completely stopped. Specifically, fine precipitates pinned dislocations and thus retarded recovery. In such conditions, further deformation leads to an increase in the density of dislocations.
Before the phase transformation, there was enough solute niobium present in the steel to delay the transformation. Specifically, solute niobium delays transformation significantly, which lowers the temperature, resulting in a higher ferrite nucleation rate and lower ferrite growth rate. As the phase transformation takes place relatively slowly, there is a possibility for interphase precipitation.
After slow cooling, which occurs in the air, there is a chance for spontaneous precipitation 
. In this study, it was found that in the HT variant, precipitation of 36.5% was achieved, and in the LT variant, 34% was achieved. As a final step of TMCP, after rolling, both strips (the LT and HT variants) were cooled at the same speed. Precipitation continued, and these precipitates had a beneficial effect in preventing ferritic perlite grain growth. A fine-grained, ferrite–pearlite microstructure with almost the same grain size was achieved (Figure 10).
In both variants of TMCP, the precipitation of Nb, the interaction of precipitates, and dislocations resulted in a fine-grained, homogeneous structure. In the cooled strip, the degree of precipitation in the high temperature variant (36.5%) was higher as compared to that in the low temperature variant (34%). This indicates that the binding of niobium to precipitates (NbC, NbN, Nb(CN)) was not accomplished entirely. Free niobium may affect the appearance of Lüders bands. The dislocation density was about the same, 108 m−2. Steels of similar chemical composition but not containing Nb have a dislocation density of 106 m−2.
The subsequent cold deformation of the strip shows the appearance of inhomogeneous deformations at the beginning of the plastic flow of the material. These inhomogeneous deformations (stress–strain curve) are related to the appearance of Lüders bands (Figure 10). Specifically, Lüders bands are localized regions of plastic deformation that are often observed in cold-rolled low-carbon steel following the appearance after the limit of proportionality
When the interaction energy between dislocations and easily diffused solute atoms (C, N, and Nb) is strong, they gather around dislocations to form Cottrell’s atmospheres. Then, dislocations are strongly anchored and the stored energy of the dislocation is decreased. When the upper yield point is reached in cold deformation, the number of moving dislocations increases because of the formation of new dislocations as a result of the multiplication mechanism. Moving dislocations are formed only at the places of the highest stress concentration. This moment corresponds to the development of Lüders bands. The appearance of Lüders bands was observed in both variants 
as a consequence of the interaction between niobium strain-induced precipitates and dislocations during cold deformation 
The article has been published on 10.3390/met10020294