INTRODUCTION

Investigators have for several decades been interested in refining the structure of various materials up to the nanoscale level via severe plastic deformation [1–3]. The main method of such a deformation is torsion under a pressure (or high-pressure torsion (HPT)) in Bridgman anvils. New techniques for the deformation preparation of nano-structured materials have been developed in recent years, which are based on the improvement of the Bridgman installation [4], the application of complex deformation by various methods [2], and varying deformation conditions (such as the pressure, temperature, and rate) [5–10]. The Zener–Hollomon parameter is usually used to simultaneously take into account the temperature and rate of deformation [5, 6, 8, 9, 11–13].

Materials such copper and aluminum, which have a low melting temperature, undergo dynamic recrystallization (DR) under severe deformation by HPT at room temperature [7, 9, 10, 14–16]. It is impossible in this case to suppress subsequent static grain-growth processes (post-dynamic recrystallization (PDR)), which change the structure and make the finding of regularities of DR difficult [7–9, 13–15]. The deformation in Bridgman anvils upon simultaneous heating, which results in the DR of materials such as nickel and iron having the high melting temperatures [5, 6, 10, 17]: has been of interest in recent years. The rapid cooling allows one to retain unchanged the structure formed upon deformation.

The aim of the present study is to investigate the development of DR in pure single-phase iron in the course of HPT at 250°С.

EXPERIMENTAL

Pure iron (99.97%) is studied; it was prepared by zone melting with five runs. The deformation at 250°С was performed at a pressure of 6 GPa using Bridgman anvils (without the lateral support of the sample). The rate of the anvil rotation was 1 rpm. The angle of rotation of the anvil was varied from 0 (upsetting) to 5 revolutions. Before deformation, disk samples were 5 mm in diameter and 0.3 mm thick. The block of the anvils and the sample placed between them were heated using a special resistance furnace after placing onto the press. The temperature was measured using a thermocouple attached to the stationary anvil. The deformation was performed after reaching the predetermined temperature. The time from the end of deformation to the extraction of the sample into air was no more than 100 s. To realize the static recrystallization of iron deformed at room temperature by HPT, it should be annealed at 250°С for 4 h [18]. Therefore, the cooling for the aforementioned time does not lead to the marked development of PDR.

The true strain (е) was calculated taking into account the angle of anvil rotation, the distance from the center of sample, and its upsetting, which in turn depends on the angle of anvil rotation and sample radius [9]. The true strain (е) at a distance of 1.0–1.5 mm from the sample center was varied from 0.9 (after upsetting) to 9.3 after deformation with five revolutions of anvil.

The structure of deformed iron was studied at the fixed distance (1.5 ± 0.2 mm) from the sample center by transmission electron microscopy (TEM) using a JEM 200CX microscope. The size of structural components (dislocation cells, microcrystallites and grains 5 µm and less in size) was determined using bright- and dark-field images and the results of more than 400 measurements; this ensures the error of less than 10%. The results of electron-microscopic studies were used to construct histograms of size distribution of structural components; the average size (dav), maximum size (dmax), standard deviation (σ), and coefficient of variation (К = σ/dav) of linear sizes were determined. The sizes of recrystallized grains, their orientation and range of grain-boundary misorientation angle were estimated by scanning electron microscopy (SEM) using a QUANTA_200 Philips scanning electron microscope equipped with an EDAX/TSL analyzer for the electron back-scattered diffraction (EBSD) analysis. The EBSD analysis was performed at an accelerating voltage of 20 kV and a tilt angle of sample of 70°; in accordance with the size of structural components, the scanning step was varied from 0.1 to 0.5 µm. Samples were subjected to abrasion and electrolytic polishing before the EBSD analysis. Points with a confidence index (CI) of less than 0.1 and grains containing four and less than four pixels were excluded from consideration. Upon taking images, the magnification was selected so that the analyzed area includes no less than 200 grains. The EBSD analysis was performed for the area at a distance of 1.0–1.5 mm from the sample center. The average grain size was determined using the linear-intercept method. In analyzing the grain misorientation, boundaries with a misorientation angle of less than 2° were excluded from consideration; boundaries with a misorientation angle of more than >15° were classified as high-angle boundaries (HAB) [19].

The Vickers hardness was measured at room temperature along the sample radius at a step of 0.25 mm using a PMT-3 tester and a 0.5 N load. To plot the dependences of the hardness on the true strain, values obtained for different samples were averaged for intervals Δе = 0.4.

RESULTS AND DISCUSSION

Figure 1 shows the microstructure of starting iron. The average grain size in the initial state was 200 µm. The material is an ensemble of randomly oriented grains; the average misorientation angle is 39°. According to the inverse pole figure (IPF) obtained by EBSD, the texture in the initial state is absent.

Fig. 1.
figure 1

Microstructure of iron (a) before deformation and (b) IPF corresponding to the initial state.

Full size image

The deformation by upsetting at 250°С (e = 0.9) and torsion with an angle of anvil rotation of 15° (e = 1.7) leads to the formation of a cellular dislocation structure (Fig. 2a). Moreover, locally misoriented areas ~1 µm in size are observed, i.e., their sizes are substantially lower than the initial grain size (Fig. 2b); in this case, the local misorientation of neighboring areas does not exceed 8° (bottom inset in Fig. 2c, curve 2). At the same time, within an initial grain, the gradual increase in the misorientation with respect to the staring scanning point occurs (bottom inset in Fig. 2c, curve 1). This is typical of the slightly misoriented cellular structure. Boundaries of initial grains remain; at such boundaries, the misorientation changes abruptly by tens degrees (top inset in Fig. 2c). After deformation with e = 2.4 (the angle of anvil rotation is 45°), the structure can be classified in accordance with two types. Firstly, the relatively weakly misoriented structure with the gradually accumulated misorientation remains (bottom area in Fig. 3a and corresponding inset).

Fig. 2.
figure 2

Microstructure of iron deformed at 250°С by (a, b) upsetting and (b) torsion with an angle of anvil rotation of 15°: (a, b) dark-field image (TEM) taken in (110)α reflection; (c) orientation map displayed in inverse-pole-figure coloring (SEM). Insets show variations in the misorientation along the scanning lines (1) relatively to the starting point and (2) between neighboring points.

Full size image
Fig. 3.
figure 3

Microstructure of iron deformed at 250°С with an angle of anvil rotation of 45°: (a) orientation map displayed in inverse-pole-figure coloring (SEM); (b) bright-field image; and (c) dark-field image (TEM) taken in (110)α reflection. Insets show variations in the misorientation along the scanning lines (1) relatively to the starting point and (2) between neighboring points.

Full size image

Secondly, the fine-grained structure with high-angle boundaries between neighboring grains is observed; their misorientation reaches 60°. Such a structure is not characterized by the accumulation of misorientation within grains (top area in Fig. 3a and corresponding inset). The TEM data show the presence of elastically distorted grains, whose boundaries are characterized by both abrupt and gradual changes of the contrast (Fig. 3b, 3c). The electron diffraction pattern becomes a ring one. At the background of reflection diffused in the azimuthal direction, several closely spaced point reflections are present. The noted peculiarities of this structure allow us to conclude that it results from the DR.

As the deformation increases in a range of 4.6–9.3 (the angle of anvil rotation is from 90° to five complete revolutions, respectively), the structure observed by TEM remains highly nonuniform and typical of DR (Fig. 4). Along with grains having relatively regular shape and low density of defects, grains with high and nonuniformly distributed density of dislocations are present. Many of the boundaries exhibit fine recrystallization nuclei. The cellular dislocation structure is not found. The electron diffraction patterns are ring and formed by point reflections (Fig. 4c).

Fig 4.
figure 4

Fine structure of iron deformed at 250°С with different angles of anvil rotation: (a) 90° (е = 4.6); (b) 180° (е = 6.0); (c) 1 revolution (е = 6.6); and (d) 5 revolutions (е = 9.3). (c) Dark-field image taken in (110)α reflection.

Full size image

However, the SEM data show that, in a range of true strain of 4.6–6.0, the fractions of low-angle and high-angle boundaries change (Figs. 5, 6). After deformation with е = 4.6, the fraction of low-angle boundaries is significant; after deformation with е = 6.0, the high-angle boundaries dominate (Fig. 5b). Because of this, the average misorientation angle increases from 26° to 39°, respectively; it remains unchanged after subsequent deformation (Fig. 6).

Fig. 5.
figure 5

Scheme of boundaries (EBSD) in iron deformed at 250°С with at angle of anvil rotation of (a) 90° and (b) 180°. Black and bright-grey lines correspond to high-angle and low-angle boundaries, respectively.

Full size image
Fig. 6.
figure 6

(◻) Variations of the average angle of grain-boundary misorientation and (●) fraction of recrystallized grains upon hot deformation of iron.

Full size image

The EBSD method allows us to analyze the gradient of intragranular orientation (the grain orientation spread (GOS)). According to the method, grains with a GOS of less than 2°–2.5° are considered to be recrystallized (free of substructure) [19, 20]. Figure 6 shows data on the fraction of recrystallized grains, which were obtained based on the analysis of GOS. It is seen that the fraction of recrystallized grains increases as the strain increases and reaches the maximum value at е = 4.6–6.6; after that, the fraction decreases. After deformation with е = 1.7, the orientation spread within a grain of studied iron can reach 15° and can be the same for a group of neighboring initial grains. After deformation with е = 2.4, the fraction of nonrecrystallized structure remains marked; the GOS for this structure reaches 10°. After such a deformation, the fraction of recrystallized grains is ~30%. As the true strain increases to е = 4.6, the structure consists of recrystallized grains and micron-sized subgrains. The GOS reaches 5°–7° in some grains; this is related to the accumulation of defects in recrystallized grains in the course of the hot deformation. As the strain increases, the fraction of grains with an orientation gradient of more than 2.5° decreases to 2–5%; as the strain reaches е = 9.3, all grains resolved by EBSD method have an orientation gradient of less than 2.5°. In this case, the fraction of recrystallized structure in Fig. 6 corresponds to the fraction of the structure resolved by EBSD.

The TEM data indicate the presence of a great number of “new” submicron-sized recrystallization nuclei (Fig. 4d), which were not found by SEM. As the true strain increases from е = 6.6 to е = 9.3, the substantial decrease in the fraction of relatively coarse (1–2 µm) grains takes place which is clearly illustrated in Fig. 7. It is likely that the transition to the new cycle of recrystallization occurs, which leads to some refining of the grains and a slight increase in the hardness (Fig. 8).

Fig. 7.
figure 7

Scheme of grain structure (EBSD) of iron deformed at 250°С with (a) 1 and (b) 5 revolutions of anvil (е = 6.6 and е = 9.3, respectively).

Full size image
Fig. 8.
figure 8

Variations of the hardness of iron deformed at (●) 250 and (dash-and-dot line [21]) 20°С. Variations of the average size of structural component in iron deformed at (◻ and △) 250 (TEM and SEM data, respectively) and (dashed line) 20°С [21].

Full size image

A comparison of the hardness of iron deformed at different temperatures shows that, at the starting stage, the hot deformation leads to more substantial hardening; in the course of further deformation, the hardness increases slightly (Fig. 8).

Upon hot deformation, the formation of misoriented structure accelerates and, subsequently, the hardness goes to the steady stage [13]. Figure 8 shows that the hardness of iron deformed at 250°С correlates with the average size of structural components, which was determined from TEM data. The SEM provides information about coarse structural components, such as blocks formed by slightly misoriented dislocation cells or subgrains (Figs. 3a, 5a) formed after relatively low deformation and coarse recrystallized grains (Figs. 5b, 7), which are formed after biigger deformation. As the true strain increases, the differences in sizes obtained by different methods decrease and, after е > 6, the sizes almost coincide (Fig. 8); the average grain size is 0.4–0.5 µm. It is likely that this is related to the steady stage of DR.

It was shown in [22] that, upon annealing of iron with the submicrocrystalline structure prepared at 20°С by HPT with five revolutions (е = 8), an axial {110} recrystallization texture forms. For example, after annealing at 350°С for 1 hour, the fraction of area occupied by grains, in which the {110} plans are parallel to the sample surface, is 40% (Fig. 9). After the hot deformation performed in the present study, the fraction of the area occupied by grains with the three orientations, such as {110}, {321}, {201}, which dominate in the structure of the studied iron, changes nonmonotonically. The maximum fraction of area, which corresponds to grains having any orientation, does not exceed 25%. After deformation at е > 6, the ratio of areas ocuppied by grains of different orientations almost corresponds to the starting textureless state (Fig. 9).

Fig. 9.
figure 9

Variations of the fraction of area of grains with certain orientations in iron in the course of its deformation at (open symbols) 250°С: (○, ●) {110}, (◻, ◼) {321}, (△, ▲) {201}. (filled symbols) Data [22] obtained for iron deformed at 20°С and annealed at 350°С for 1 h are given for comparison.

Full size image

CONCLUSIONS

Upon HPT at 250°С, the dynamic recrystallization in iron of 99.97% purity begins on reaching the true strain е = 2.4. It was found by SEM and TEM that a structure with an average grain size of 0.4–0.5 µm forms. Results of the analysis of orientation spread within a grain allowed us to estimate the fraction of recrystallized structure, which reaches the maximum value, 60%, at 4.6 < е < 6.6. As the true strain increases, the fraction of recrystallized structure decreases. In this case, the grain refinement and a slight increase in the hardness are observed. The combination of these data indicates the cyclic character of dynamic recrystallization.

In contrast to the static recrystallization of iron deformed at room temperature with the same true strains, the dynamic recrystallization of iron studied in the present work does not lead to the formation of any preferred grain orientation.