Investigation of Strain-Induced Precipitation of Niobium Carbide in Niobium Micro-Alloyed Steels at Elevated Temperatures

Abstract

Two steels with a base composition of Fe-0.2C-0.8Mn-1.2Cr (wt%) but with different niobium (Nb) contents (0.02 and 0.03 wt%) were employed to study the effect of precipitate evolution on the softening resistance in the austenite region under elevated temperature deformation. The thermomechanical procedure was executed by a deformation-dilatometer and involved double deformation processes with 25% strain at a 0.25 s⁻¹ strain rate at 900, 925, 950, and 1000 °C. The softening ratios, reflecting the competition between recrystallization and precipitation, were evaluated. The results indicated that both steels showed better softening resistance at 900 °C than at other temperatures. However, the softening ratio of 0.03 wt% Nb-containing steel (Steel 3N) rose after 100 s at 900 °C, while 0.02 wt% Nb-containing steel (Steel 2N) maintained a low softening ratio within 300 s at 900 °C, indicating that Steel 3N was relatively non-durable. A microstructural characterization showed that in the Steel 3N sample deformed at 900 °C, recrystallization occurred more strongly than for Steel 2N after a 1000 s holding time. A follow-up analysis then showed that Steel 3N treated at 900 °C revealed a faster coarsening of the carbides than Steel 2N even in the early stage of precipitation, evidencing that Steel 2N exhibited a lower softening resistance at 900 °C.

1. Introduction

High-strength low-alloyed (HSLA) steels have been used in a wide range of applications [1,2,3,4]. The mechanical properties of HSLA steels can be enhanced by precipitation strengthening and grain refinement through the addition of micro-alloying elements and the use of appropriate processes. Therefore, high strength can be achieved while maintaining toughness, weldability, and formability [2,5,6,7,8,9,10,11]. Common micro-alloying elements include niobium, titanium, and vanadium, which can form small carbide precipitates with carbon and nitrogen. In addition to precipitation strengthening, precipitates occurring at defects such as dislocations or grain boundaries are presumed to have a significant effect on delaying recrystallization and impeding grain growth [12,13,14,15,16,17].

In the machinery and automotive industries, carburizing technology is widely used on gears, bearings, and other workpieces after machining. The surfaces are hardened by carburizing but the softer structure inside the steel is retained, thereby maintaining toughness, and also improving the wear resistance and fatigue strength [18,19,20]. The carburizing process requires high temperature treatment and increasing the temperature can effectively reduce the carburizing time and improve the efficiency of the process. However, the degree of grain growth is significant at high temperatures [21,22,23], and coarse grains may degrade the fatigue and impact properties of steels. It is worth noting that excessively coarse grains will cause anisotropic properties, affecting the accuracy of workpieces. To improve performance, it is necessary to develop a method that can suppress austenite grain coarsening at high temperatures. It is clear that precipitates can pin the grain boundaries of austenite [24,25,26,27,28], and the favorable factors that can prevent abnormal grain coarsening are illustrated as follows. When the volume fraction of precipitates is large and the precipitate’s size is small, the precipitates have a better pinning effect on grain boundaries [29]. Second-phase ultra-fine particles, such as aluminum nitride or niobium, titanium, and vanadium carbon–nitrogen compounds, can be used to prevent grain boundary movement due to their pinning force on the grain boundary [18,23,30,31,32,33,34]. Modeling is an important auxiliary tool to support the analysis of carbonitride precipitation processes in HSLA steels. Some models of thermodynamics and kinetics of precipitation in micro-alloyed steels have been reported [35,36]. The key to the pinning force of the second phase particles is whether they are sufficiently small and well dispersed in the matrix. It is critical to establish a proper precipitation process to achieve these characteristics. As a result, when the steel is re-austenitized for high-temperature carburization, the existing precipitates in the matrix will hinder the movement of austenite grain boundaries and prevent abnormal grain coarsening. In view of the above, careful consideration should be given to the composition of the steel and the thermomechanical treatment, both of which are relevant to the control of carbide precipitation.

The thermo-mechanical controlled process (TMCP) is an advanced process for producing micro-alloyed steel plates. The plastic deformation of steel at high temperatures introduces a large number of defects and induces the early precipitation of particles at high-energy defect positions, a phenomenon known as strain-induced precipitation. Since the precipitates delay recrystallization, the deformed grains with a high strain energy can be stored at the subsequent phase transformation temperature of the ferrite phase, leading to the grain refinement of the produced ferrite [14,15,37,38,39]. Besides the effect of delaying recrystallization at high temperatures, the fine and dispersed precipitates can also be maintained at room temperature to provide precipitation strengthening, which together with the refinement of grains enhances the mechanical properties of steels.

Strain-induced precipitation involves the application of strain to the steel before precipitation, as mentioned in the previous paragraph. The strain introduces a considerable number of defects, and the type of the defect that affects precipitation behavior the most is dislocation. Many investigations have reported that niobium carbides nucleate preferably at dislocations instead of homogeneously throughout the matrix [40]. B. Dutta et al. [13,41,42] have proposed a model of strain-induced precipitation in which the precipitate characteristics can be properly illustrated as follows. When the steel is subjected to plastic deformation, many dislocations are generated, and the dislocations become entangled with one another, forming many dislocation nodes. Precipitation at dislocation nodes further lowers the energy barrier of precipitation by eliminating dislocation cores; thus, precipitates prefer to form at these high energy locations. Moreover, since precipitates form preferably at dislocation nodes, they are connected to one another by dislocations and dispersed widely. The solute atoms can diffuse faster along dislocations owing to the higher diffusion coefficient than that in bulk, i.e., pipe diffusion; therefore, precipitates will rapidly coarsen through dislocations after strain-induced precipitation. The precipitation reaction can be divided into three stages: nucleation, growth, and coarsening. However, these three evolution stages of strain-induced precipitation cannot be individually distinguished. The growth and coarsening of precipitates can occur simultaneously in strain-induced precipitation.

The present work was an attempt to explore the optimum conditions for Nb-containing steels to avoid abnormal grain coarsening during carbonization. Two different amounts of Nb additions (0.02 and 0.03 wt%) in the Fe-0.19C-0.82Mn-1.16Cr (wt%) base alloy were investigated. Previous works [38,43] have examined the effects of adding niobium on strain-induced precipitation and the recrystallization of austenite. One work compared no niobium content to 0.16 wt% niobium content in 0.02C-1.5Mn (wt%) steels [36], and another varied the carbon/manganese contents, comparing 0.1C-1.4Mn (wt%) steels to 0.04C-1.8Mn (wt%) steels with the same 0.08 wt% niobium content [43]. However, the influence of the niobium content remains to be further studied. Therefore, the objective of the present work was to study the recrystallization and precipitation conditions of steels to elucidate the influences of different amounts of niobium content.

2. Materials and Methods

Two experimental steels used in this work were fabricated by melting pure metals and alloying elements in appropriate proportions in a vacuum induction furnace. The molten alloyed metal was cast in copper molds and hot-forged to rods of about 32 mm in diameter. Their chemical compositions are listed in

Table 1. Chemical composition of the experimental steels (wt %).

No.	Fe	C	Cr	Mn	Si	Nb
Steel 2N	Bal.	0.19	1.20	0.82	0.38	0.02
Steel 3N	Bal.	0.19	1.16	0.85	0.32	0.03

. Nitrogen content of these two steels were analyzed to be about 0.006 wt%. The stable phases and their detailed compositions at 900 °C in the two steels had been preliminarily studied via their calculation through Thermo-calc. The results show that in austenite, the content (mole fraction) of the N is less than 1/10 of that of the C or Nb content. It is clear that at 900 °C, the equilibrium precipitate phase, NbC, contains a very small level of N in these two steels. The main difference in the compositions of the two steels was the amount of niobium. The one labeled Steel 2N contained 0.02 wt% Nb, and the other, Steel 3N, had 0.03 wt% Nb. The dilatometer specimens were prepared axially from the half-radius regions of the rods, and finally machined into cylindrical specimens with a diameter of 5 mm and a length of 10 mm for heat treatments, which were conducted in a deformation-dilatometer (DIL805A/D, TA instrument, New Castle, DE, USA). Nitrogen content of two steels were analyzed to be about 0.006 wt%. The stable phases and their detailed compositions at 900 °C in the two steels studied had been preliminarily studied via the calculation using Thermo-calc. The results showed that N content (mole fraction) is less than 1/10 of C or Nb contents (mole fractions) in austenite. It is clear that at 900 °C, the equilibrium precipitate phase, NbC, contains a very small level of N in these two steels.

The thermomechanical procedure performed by a deformation-dilatometer involved double compressive deformation processes, as presented in Figure 1a. In the first process, the specimens were austenitized at 1300 °C for 6 min; each cooled to 900, 925, 950, and 1000 °C; immediately deformed with a strain of 25% at a strain rate of 0.25 s⁻¹; and then held at that temperature for periods ranging from 1 to 800 s. After the isothermal holding, the second process was subsequently carried out with the same strain (25%) and at the same strain rate (0.25 s⁻¹) and temperature as the first process.

After double compressive deformations, two corresponding stress–strain curves were obtained from the dilatometry data, as illustrated in Figure 1b. The offset yield stresses (S_I,2% and S_II,2%) were determined in the first and second curves, respectively, by using a plastic strain of 2%, and ${\mathrm{S}}_{\mathrm{m}}$ was the maximum stress of the first curve. According to a previous work [44], the 2% offset method is one of the best methods to describe the static recrystallization kinetics of the Nb-containing steels under investigation, and the softening ratio (X) is determined by using the 2% offset method from these three stresses: X = (${\mathrm{S}}_{\mathrm{m}}-{\mathrm{S}}_{\mathrm{II},2\%}$) / (${\mathrm{S}}_{\mathrm{m}}-{\mathrm{S}}_{\mathrm{I},2\%}$).

The softening ratio reflects the competition between recrystallization and precipitation at the holding temperature for a holding time. If the steel completely recrystallizes, the second stress–strain curve will be approximately the same as the first one, i.e., the value of S_II,2% will be close to that of ${\mathrm{S}}_{\mathrm{I},2\%}$, and the value of the softening ratio will reach 1. If the steel does not recrystallize, the second stress–strain curve will continuously follow the first stress–strain curve, i.e., the value of ${\mathrm{S}}_{\mathrm{II},2\%}$ will be close to that of ${\mathrm{S}}_{\mathrm{m}}$, and the softening ratio will reach 0. Since, in this study, the precipitates could delay recrystallization or prevent new grains from growing, the precipitation time could be evaluated by analyzing the evolution of the softening ratio for different holding times at each deformation temperature. After the acquisition of the softening ratio curve and determination of the precipitation time for each holding temperature, the samples treated with the most favorable precipitation temperature for softening resistance were selected for subsequent metallography and transmission electron microscopy (TEM) analyses. Here, the samples used in the analyses were taken from steels treated with the first process, held at that temperature for different periods ranging from 1 to 800 s, and quenched directly to room temperature.

Samples for optical metallography (OM) were prepared from the corresponding dilatometer specimens, and the observation area was at least 1 mm away from the rod surface to exclude the influence of decarbonization. The picric acid powder was dissolved in batches in 70 °C hot water in a beaker container, until some picric acid particles precipitated at the bottom of the beaker, whereby a supersaturated picric acid solution was obtained. The polished OM samples were immersed in the supersaturated picric acid solution at 70 °C for 100 s and then rinsed with clean water. Newly formed grain boundaries and prior austenite grain boundaries were revealed. Samples for TEM were prepared by twin-jet electro-polishing in an electrolyte solution containing 5% perchloric acid, 15% glycerol, and 80% ethanol at −5 °C with a working current of 20 mA. Afterward, the TEM samples were observed with a thermal-emission-gun TEM (FEI Tecnai G2 T20, Thermo Fisher Scientific, Waltham, MA, USA) and a field-emission-gun TEM (FEI Tecnai G2 F20, Thermo Fisher Scientific, Waltham, MA, USA), both operated at 200kV.

3. Results and Discussion

3.1. Softening Ratio Curves and Precipitation-Time-Temperature Curves

The softening ratio versus time curves and the corresponding precipitation-time-temperature (PTT) curves for the Steels 2N and 3N at temperatures of 900, 925, 950, and 1000 °C are presented in Figure 2. As can be seen in Figure 2a, the softening ratio curves have similar shapes and upward trends at each temperature. However, at higher holding temperatures, the recrystallization of the steel is more pronounced, so as the holding time increases, the difference in the curve behavior at different temperatures becomes more obvious. The curve at 1000 °C starts to rise rapidly at the very beginning of the holding period, while the curves of 900, 925, and 950 °C rise slowly and display nearly the same softening rate. This phenomenon can be explained by the fact that the common recrystallization process has an S-shaped curve depicting the reaction rate, which is characterized by a slow start, a fast middle period, and a slow end, when plotted in terms of recrystallization degree and time. At higher recrystallization temperatures, the recrystallization will occur earlier; that is, the high softening rate of 1000 °C in the early stage is the result of the high temperature causing prevalently intense recrystallization, and the low softening ratio of 900 °C~950 °C is due to the absence of recrystallization. At the same strain rate with the same amount of strain for both steels, the recrystallization abilities for the Steels 2N and 3N in the early stage at the same temperature are not significantly different. The softening ratios of the Steels 2N and 3N show only a slight difference in the early stage because the precipitation has not occurred. As precipitation and coarsening occur during the isothermal holding after the first compressive deformation, the effects of the compositions of the two steels are evident. After holding for a longer time, Steel 3N experiences a more serious softening phenomenon than that of Steel 2N, i.e., the value of the soft ration for the former is obviously larger than that for the latter, indicating that a high niobium content may not always improve softening resistance. Both steels exhibit good softening resistance at the beginning of the 900 °C holding (Steel 2N within 300 s and Steel 3N within 100 s). However, as the temperature rises, the duration of the low softening ratio (X < 0.1) is rapidly shortened. At 1000 °C, neither Steel 2N nor Steel 3N can maintain low softening because of the rapid recrystallization.

In the curves of the softening ratio (Figure 2a), the precipitation time can be determined from the point at which the value of the softening ratio starts to decrease. To obtain the nose tips of the precipitation-start-time versus temperature curves, the precipitation-start-times at 875 °C and 850 °C must also be determined. In Figure 2b, it can be seen that the precipitation-time-temperature (PTT) curves of Steels 2N and 3N are similar, and the nose tips of the curves are at about 900 °C. Both precipitation and recrystallization tend to occur at the high-energy defects; however, in order to retard the latter, the former must occur in advance. Therefore, the optimum hot-working temperature should be close to the nose tip of the PTT curve.

3.2. Prior Austenite Grains

The 2N and 3N steel samples, which were treated at a 900 °C holding temperature, exhibited the shortest precipitation time, and were expected to maintain a smaller size of their precipitates for a longer time, as compared to those treated at 925, 950, and 1000 °C. Both the holding temperature and time contribute to the softening resistance; therefore, the experimental temperature of 900 °C and the holding times of 10 s and 1000 s were chosen for the subsequent analysis to compare the pre- and post-precipitation periods of related samples in the steels. Figure 3 presents the optical metallographs of the prior austenite grains in the samples of both steels treated by the first compressive-deformation process, held at 900 °C for 10 s and 1000 s, respectively, and quenched directly to room temperature. The compression strain is oriented horizontally (in Figure 3a–d), so the pancake structure where the grains are stretched up and down can be seen (in Figure 3a,c). In the 2N and 3N steel samples treated with 10 s holding at 900 °C (Figure 3a,c), the prior austenite grain boundaries remain intact. The grains can be clearly distinguished, indicating that no recrystallization has occurred. Since recrystallization has not occurred, the band structures in the grains (as seen in Figure 3a) are presumably annealing twins of austenite produced during the austenitization at 1300 °C. In Figure 3b,d, the newly-formed fine grains of austenite in the samples of both steels treated with holding times of 1000 s can be identified. These new equiaxial grains can be seen at the grain boundaries of the pancake structures, whose original boundaries cannot be well distinguished. The reason why new equiaxial grains first formed at the grain boundary of the pancake structure is that the region where the slip band and grain boundary intersected stored more strain energy, which provided a greater driving force for recrystallization.

The features of the OM metallographs in Figure 3 is consistent with the trend of the softening ratio curves. The value of the softening ratio did not increase within 10 s at 900 °C (Figure 2a), indicating that recrystallization did not occur. As the holding time at 900 °C was increased, the value of the softening ratio of Steel 3N gradually surpassed that of Steel 2N (Figure 2a), and the OM metallographs also show a more obvious recrystallization in Steel 3N than in Steel 2N (Figure 3b,d). However, although the softening ratios of both Steels 2N and 3N increased after 300 s of holding at 900 °C, the increases were not significant. Therefore, although evidence of recrystallization can be seen in Figure 3b,d, the original pancake structure can still be roughly distinguished, i.e., a recrystallization reaction was not yet complete.

3.3. Precipitate Observation

Both Steels 2N and 3N had fully lath martensite structures after quenching. Owing to the low alloy and carbon contents, the martensite start temperature was high (approximately 360 °C), and it also resulted in many auto-tempered structures of martensite within the steels.

Figure 4 presents the TEM images of the precipitates that formed in the auto-tempered martensite matrix. The precipitates in the images have two kinds of morphology: one strip-like (indicated by blue arrows) and the other spherical (indicated by red arrows). Strip-like precipitates can only be found within the laths, whereas spherical precipitates have no specific area of occurrence. The reason for the difference is their distinct formation temperatures. The spherical precipitates are presumed to be niobium precipitates, which formed in austenite at high temperatures; since the steels were subjected to strain and precipitates subsequently formed throughout the whole austenite matrix at high temperatures, the precipitates were then evenly dispersed in the martensite due to the transformation of austenite to martensite in the quenched samples. As to the strip-like precipitates, they are presumed to be cementite platelets, which formed within the martensite laths due to the auto-tempering effect during the cooling of the martensite to ambient temperature. The auto-tempering cementite platelets are easily generated from early-formed martensite in the temperature range of 300°C ~400 °C near the martensite’s start temperature. Laths with similar orientations would be coalesced together and discharge the oversaturated carbon; immediately, the carbon would be combined directly with the iron to form cementite instead of forming alloy carbides with other alloying elements, i.e., NbC carbides.

A representative TEM image in Figure 5a shows a spherical niobium precipitate (marked by a red circle) and its corresponding energy-dispersive X-ray spectroscopy (EDS) is presented in Figure 5b. As the precipitate was embedded in the matrix, the EDS displays a high level of Nb, also accompanied by a much higher content of Fe and minor contents of Cr and Mn. The TEM sample of Steel 3N, which was subjected to a 1300 °C treatment for 3 min and then quenched to room temperature without any other treatment, was examined as shown in Figure 6; Figure 6a presents a TEM dark field image of strip-like cementite platelets with the corresponding diffraction pattern (Figure 6b) and a high-resolution transmission electron microscopy (HR-TEM) image (Figure 6c) with the corresponding fast Fourier transform (FFT) diffractogram (Figure 6d). Cementite platelets are easily visible in the sample, which contains no spherical precipitates. Figs. 6c and 6d more clearly show that the strip-like precipitates were cementite [45], instead of the high-temperature NbC carbide precipitate.

3.4. Precipitate Statistics

Two kinds of precipitates in the samples of Steels 2N and 3N have been confirmed. The strip-like ones proved to be cementite, which formed within the auto-tempered martensite during the quenching to room temperature and, therefore, did not contribute to the softening resistance at high temperatures. Thermo-Calc software (2022b, Thermo-Calc Software, Stockholm, Sweden) was also used to evaluate the newly-formed phase and its equilibrium amounts in Steels 2N and 3N at 1300 °C and 900 °C. The results are listed in

Table 2. The stable phase and its equilibrium amount at 1300 °C and 900 °C for Steels 2N and 3N, calculated via Thermo-Calc software.

Temperature		2N	3N
1300 °C	stable phase	fully solid solution	fully solid solution
	amount (in at%)	no precipitation	no precipitation
900 °C	stable phase	NbC	NbC
	amount (in at%)	0.0247	0.0372

. At 1300 °C, both Steels 2N and 3N were fully FCC austenite solid solution, and at 900 °C, stable niobium carbides formed in both steels. It is appropriate to conclude that the spherical NbC precipitates provide the main contribution to softening resistance at high temperatures.

Figure 7 presents the size distributions of the precipitates in both steels treated with holding times of 10 s and 1000 s, respectively, after the first compressive deformation at 900 °C. Since only the precipitates produced at high temperatures in the austenite region contributed to the softening resistance, only spherical precipitates (NbC) were counted in the statistics. The general precipitation process includes three stages: nucleation, growth, and coarsening. During the growth stage, the precipitates simultaneously nucleate and grow, and the size distribution follows a unimodal Gaussian distribution. The peak gradually moves to a larger size as the process continues. When the precipitation process enters the coarsening stage, large precipitates begin to incorporate smaller ones, and then the size distribution gradually changes from unimodal to bimodal. Figure 7a,b (for Steel 2N) show that as the holding time increases, the peak size of the precipitate size distribution curve becomes larger. Although the bimodal size distribution is not observable in Figure 7b (for Steel 2N), it shows that the numbers of precipitates with sizes smaller than ~25 nm decreased and those of larger than 35~40 nm increased, indicating that for Steel 2N held at 900 °C for 1000 s, the coarsening of NbC carbides took place. In Figure 7c,d (for Steel 3N), the bimodal size distribution can be clearly observed, implying that the precipitation process quickly enters the coarsening stage even if the holding time is only 10 s at 900 °C. As the holding time increases to 1000 s for Steel 3N, the number of precipitates in the smaller size peak region decreases and the size of precipitates in the larger size peak region increases, indicating the occurrence of strong coarsening in Steel 3N, which displays the faster evolution of the precipitation of NbC as opposed to Steel 2N. The high fraction of large precipitates in Steel 3N shows that the high niobium content was consumed to form larger precipitates rather than more small precipitates. All these findings suggest that a high niobium content of 0.03 wt% may lead to the coarsening of precipitates and affect the softening resistance of steels subjected to thermomechanical processes in the present work.

4. Conclusions

Two steels with the same base composition of Fe-0.19C-0.82Mn-1.16Cr (wt%) but different niobium (Nb) contents of 0.02 and 0.03 wt%, accordingly denoted Steel 2N and 3N, were investigated with respect to softening resistance in the austenite region (850~1000 °C) under elevated temperature deformation. A deformation-dilatometer was employed to conduct double deformation processes, from which two corresponding stress–strain curves brought about the softening ratio versus time curve to evaluate the softening resistance. The softening ratio curves reflected the precipitation start time; when precipitation occurred, the precipitates delayed the recrystallization and strengthened the steels such that the values of the softening ratio no longer rose, or even showed a downward trend. The results showed that the softening ratio of the 0.03 wt% Nb-containing steel rose after 100 s at 900 °C, while the 0.02 wt% Nb-containing steel maintained a low softening ratio within 300 s at 900 °C. The microstructural characterization showed that the recrystallization of the 0.03 wt% Nb-containing steel occurred more strongly than the 0.02 wt% Nb-containing one after being deformed at 900 °C and a 1000 s holding time. In addition, the precipitates in Steel 3N revealed a tendency to coarsen even in the early stage of the isothermal holding. It should be noted that steels containing higher Nb content would generally be expected to exhibit a better softening resistance. However, under our designed experimental conditions, the results indicate that an excess niobium content may be unconducive to the softening resistance and accelerate the coarsening of precipitates in steels subjected to the thermo-mechanical process. Dynamics factors, such as the amount of strain and the strain rate, should also be important factors to effect softening resistance.