2014 Volume 54 Issue 7 Pages 1729-1734
To examine effects of the grain boundary (GB) and dislocation on the deformation mechanism for ultrafine-grained (UFG) and coarse-grained (CG) interstitial-free (IF) steels at room temperature, tensile tests and several types of microscopy were conducted for each steel. Atomic force microscopy revealed that the contribution of grain-boundary sliding (GBS) on deformation increased more prominently in UFG region than in CG region. Moreover, transmission electron microscopy revealed that dislocation motion was dominant in CG steel, where cell structure was formed with increasing strain. On the other hand, although dislocations moved in UFG steel, they did not tangle and piled up at GB, where interaction between GB and dislocation occurred markedly, causing significant GBS. Therefore, the dominant deformation mechanism changed from dislocation motion to GBS by decreasing grain size in IF steel.
Ultrafine-grained (UFG) metals with grain sizes of <1 μm show superior mechanical properties to those of conventional coarse-grained (CG) metals.1) UFG metals are fabricated by severe plastic deformation (SPD) processes such as equal channel angular pressing,2,3,4) high pressure torsion5,6,7) and accumulative roll-bonding (ARB).8,9,10) Among these processes, only ARB can be applied to continuous production of large bulky materials.
However, ductility and fracture strain are often reduced,11) which results in extremely limited application of the UFG metals. For further improvement of the material properties, the deformation mechanisms must be understood. Therefore, several deformation models have been proposed to predict the deformation behavior for UFG metals, such as deformation based on dislocation motion suggested by Kato et al.,12) or grain-boundary (GB) deformation described in earlier reports.13,14,15) Blum et al.16) reported that lattice dislocation piled up at GB without tangles in UFG metals, meaning that the interaction between dislocation and GB might have some importance. Moreover, the interaction is expected to be activated because UFG metals fabricated by SPD processes have non-equilibrium GB,17) which has excess GB energy.18,19)
In fact, the authors have proposed that grain-boundary sliding (GBS) was activated with decreasing strain rate ( ) in UFG interstitial-free (IF) steel,20) but that mechanism remains unclear. Therefore, this study was conducted to ascertain the GBS mechanism and discuss the effect of dislocation − GB interaction on deformation by comparing deformation in CG and UFG metals.
Ti-added ultra-low carbon IF steel with grain sizes of 16 μm was used as a virgin material for this study. The chemical composition of the material is shown in Table 1. UFG samples were fabricated by 5-cycled ARB process at 823 K, which introduced the equivalent strain8) of about four. Then, samples were annealed at 723 K for 0.5 h, where the grain thickness, width, and length were 0.39, 0.75, and 1.1 μm, respectively, for UFG steel 1, 0.42, 0.93, and 1.5 μm, respectively, for UFG steel 2, 0.43, 0.86, and 1.0 μm, respectively, for UFG steel 3, and 0.53, 1.3, and 1.7 μm, respectively, for UFG steel 4. It is noteworthy that the variety of grain sizes results from the differences of plate batches even after the same annealing condition. In addition, samples with several grain sizes were obtained by annealing using ARBed and virgin samples to acquire the Hall-Petch relation of the steel. Detailed sample’s conditions are summarized in Table 2. However, virgin sample as CG steel 4 is used mainly in the paper.
Tensile tests were performed with constant cross-head speeds corresponding to =10–6−100 s–1 at room temperature to investigate the mechanical properties, especially the strain-rate sensitivity exponent (m) of 0.2% proof stress (σ0.2). The strain was measured using strain gauges mounted directly on the specimen surface. Moreover, the specimens were prepared with 6 mm gauge length with the loading direction corresponded to the rolling direction in ARB using an electric discharge machine.
Therein, εgb is the strain generated by GBS, v is the heights of surface steps, f is the geometrical factor of 1.1, and d1 is the grain length parallel to the loading direction which coincides with dRD in this study. At least 350 grain-boundary steps were measured before and after tensile tests stopped at around σ0.2 and tensile strength (σB).
Subsequently, X-ray diffraction (XRD) measurements, electron back-scatter diffraction (EBSD) analyses, and transmission electron microscopic (TEM) observations were conducted with the samples deformed by tensile tests to reveal the intragranular dislocation motion. First, XRD measurements were performed to evaluate the change of dislocation density (ρ) using a Williamson-Hall plot.23)
In that equation, D is the average crystallite size, λ is the wavelength of 0.154 nm (CuKα) in this measurement, β is the full width at half maximum of diffraction peaks, ε is the strain induced inside the samples, and θ is the Bragg’s angle. Then, ρ was calculated from ε by using following equation;24)
Second, EBSD analyses were conducted to observe strain distribution, i.e., local misorientation, using Karnel Average Misorientation. Especially, the orientation gradient dMave/dh25,26) was evaluated as the magnitude of plastic strain because the parameter correlates well with the degree of damage caused by plastic strain.25,26) Samples were prepared using electropolishing with a solution of 5% perchloric acid and 95% acetic acid at 288 K at a voltage of 30 V.
Finally, TEM observations were done to elucidate the dislocation structure after deformation. The observations were operated with voltage of 200 kV using LaB6 cascade. TEM samples were prepared using twin-jet electro-polishing with a solution of 5% perchloric acid and 95% acetic acid at 288 K at a voltage of 50 V.
Figure 1 shows a double logarithmic plot of σ0.2 and for each sample. The m value decreased with decreasing grain size and strain rate. In UFG steel, although the m value was about 0.02 at the high strain rates ( >10–3 s–1), it equaled to –0.01 at the low strain rates ( <10–3 s–1). Similarly, the value decreased from 0.07 to 0.02 with decreasing strain rate in CG steel.
Double logarithmic plot of σ0.2 and . The m value was about 0.02 at the high strain rates ( >10–3 s–1) while it equaled to –0.01 at the low strain rates ( <10–3 s–1) in UFG steel. Similarly, the m decreased from 0.07 to 0.02 with decreasing strain rate in CG steel 4.
Thereby, AFM observations were performed to investigate correspondence of the transition of m value to GB deformation. Figure 2 shows the surface profiles and differential contrast images taken after the tensile tests at ≈10–5 s–1. In this figure, because slip bands and few grain boundary steps by GBS were observed, intragranular deformation is significant in CG steel. On the other hand, because the steps appear significantly, intergranular deformation becomes dominant in UFG one.20) Then, Fig. 3 shows the contribution of GBS (εgb/εpl) functioned as plastic strain (εpl). In Fig. 3(b), εgb/εpl at around εpl=0.002 is increased from 0.06 to 0.4 with decreasing grain size at ≈10–3 s–1. It increased from 0.12 to 0.76 with decreasing strain rate in UFG steel. These changes of εgb/εpl were consistent with the transition of the m value.
Surface profiles and differential contrast images taken after the tests at ≈10–5 s–1 for (a) CG steel 4 and (b) UFG steel 2.20) Because slip bands and few grain boundary steps by GBS were observed, intragranular deformation is significant in CG steel. On the other hand, because the steps appear significantly, intergranular deformation becomes dominant in UFG one.20)
Relation between εgb/εpl and εpl at different strain rates in CG steel 4 and UFG steel 2, and (b) is enlargement region of (a). εgb/εpl is increased from 0.06 to 0.4 with decreasing grain size at ≈10–3 s–1. Moreover, εgb/εpl is increased from 0.12 to 0.76 with decreasing strain rate in UFG steel. These changes of εgb/εpl were consistent with the transition of m value.
Moreover, the effect of grain size on the GBS represents particularly in Fig. 4 which portrays the Hall-Petch relation and εgb/εpl functioned as the average grain size (d) of the grain thickness, width, and length at ≈10–5 s–1. The figure shows that σ0.2 increases dramatically in CG-UFG transient region (d of several μm) and becomes stable in UFG region (d<1 μm), but the contribution of GBS increases in UFG region.
Hall-Petch relation of the samples and relation between εgb/εpl and the average grain size of the grain thickness, width, and length. The strength and the contribution of GBS increases with decreasing grain size obviously.
Next, dislocation structures in CG and UFG steels were investigated using microscopic observations. XRD measurements were performed to evaluate the change of ρ with increasing strain in the primary deformation region. Figure 5 shows the relation between ρ and εpl. Although ρ increased with increasing εpl in CG steel, it decreased in UFG steel.
Relation between ρ and εpl in (a) CG steel 4 and (b) UFG steel 5. ρ increased with increasing εpl in CG steel; it decreased in UFG steel.
The results of EBSD analyses revealed the same trend as that observed in XRD measurements. Figures 6 and 7 show local misorientation maps for CG and UFG steel, respectively. In CG steel, the distribution of the misorientation at εpl=0.002 was similar to the cell structure and appeared more prominently with increasing εpl. However, the distribution was not observed in UFG steel. Additionally, intragranular misorientation decreased and distributed near GB after deformation in UFG steel. Figure 8 shows the relation between dMave/dh and εpl, where the value increased with increasing εpl in CG steel, whereas it decreased in the primary deformation region in UFG steel.
Local misorientation maps in CG steel 4 with various εpl of (a) 0, (b) 0.002, (c) 0.04, and (d) 0.27. Distributions of the misorientation were similar to cell structure at εpl=0.002 and appeared more prominently with increasing εpl.
Local misorientation maps in UFG steel 3 with various εpl of (a) 0, (b) 0.002, (c) 0.004, (d) 0.008, and (e) 0.012. Distributions such as cell structure were not observed. Intragranular misorientation decreased and distributed near GB after deformation.
Relation between dMave/dh and εpl in CG steel 4 and UFG steel 3. The value increased with increasing εpl in CG steel, although it decreased in UFG steel.
Figures 9 and 10 show TEM bright-field images for CG and UFG steels before and after tensile tests at strain rate of ≈10–5 s–1. In the image of CG steel, the dislocation tangled and cell structure was generated in grains. The cell structure sizes were similar to that observed by EBSD analysis. In contrast, dislocations piled up at GB without tangles and the cell structure was not generated after deformation in UFG steel. It means that dislocations move in grain interior and is absorbed into it. These results were consistent with results of XRD and EBSD analyses.
TEM bright-field images in CG steel 4 (a) before and after tensile tests at of 1×10–5 s–1 stopped at (b) σ0.2 and (c) σB. Dislocations tangled and cell structure was generated in grains with increasing strain.
TEM bright-field images in UFG steel 2 (a) before and after tensile tests at of 8.3×10–6 s–1 stopped at (b) σ0.220) and (c) σB. Dislocations stored at GB without tangles and cell structure was not generated.
In this paper, an effect of GB and dislocation on deformation in CG and UFG steels at room temperature was studied. CG steel is deformed by dislocation motion, where the dislocation piles up and cell structure is formed with increasing εpl. However, UFG steel shows that dislocation moves to GB without tangles and is absorbed into it27,28) because GB in UFG steel has higher GB energy than that in CG steel. The absorption of dislocation activates GBS to relax intragranular strain, i.e., “slip-induced GBS”,29,30) whose amount increases with increasing GB energy.31) Moreover, because the activation energy of the GBS was evaluated as about 20 kJ/mol,29,31) it might activate at room temperature where diffusion processes do not activate. Therefore, it claimed that deformation by slip-induced GBS becomes the dominant deformation mode in UFG region for IF steel.
Furthermore, the amount of GBS is influenced by the dislocation absorption rate. Since the dislocations arriving at GB are too numerous to absorb at high strain rates, many dislocations might not be absorbed but pile up at GB. On the other hand, because the dislocation has sufficient time to be absorbed into GB at low strain rates, the amount of absorbed dislocations increases, leading to slip-induced GBS obviously (Fig. 3). Moreover, because the amount of GBS is changed by the strain rate, the absorption rate is lower than the pile up rate at the primary deformation.
These results are explainable by the following deformation mechanisms. In CG region, the increase of ρ and local misorientation results from formation of the cell structure (Figs. 5, 6, 8, and 9). In contrast, because dislocations moves toward GB and are absorbed into it during deformation, ρ and the misorientation decrease in UFG region (Figs. 5, 7, 8, and 10). Moreover, the phenomenon might be activated because non-equilibrium GB with excess GB energy can absorb dislocations easier than equilibrium GB. Therefore, the increase of strength in CG and the transient regions is caused by intragranular deformation, i.e., dislocation motion, whereas the saturation of it in UFG region is thought to be due to the dislocation absorption caused by the slip-induced GBS.
Finally, our results and discussion were compared with earlier studies.32,33,34,35) They reported that UFG metals have the activation volume of <100 b3 at room temperature. In these study, Conrad et al.32) mentioned that the dislocation slip is dominant deformation mode but cell structure is not constructed, which coincides with our results of microscope observations. In addition, depinning process of dislocation from GB might rate-control at >10–4 s–1.12,34) Although the model showed good agreement with our experimental data at relatively high strain rates, it does not explain the strain-rate sensitivity of the amount of GBS. On the other hand, Blum et al.16) mentioned that dislocation annihilation at GB rate-controls at <10–4 s–1, and our results and discussion claimed that the GBS activates in the condition (Figs. 2 and 3). Moreover, it revealed that the amount of GBS has the strain-rate sensitivity because the number of dislocations absorbed into GB increases with decreasing strain rate. Therefore, it considers that GBS becomes dominant with decreasing grain size and strain rate obviously at room temperature.
To examine the relation between GB and dislocation on deformation mechanism for CG and UFG IF steels at room temperature, tensile tests and several types of microscopy were used for these samples. Detailed conclusions are presented below.
(1) The m values of σ0.2 decreased with decreasing grain size or strain rate, i.e. m decreased from 0.02 to –0.01 in UFG steel and from 0.07 to 0.02 in CG steel. In addition, the contribution of GBS increased with decreasing grain size and increased with decreasing strain rate in UFG steel. These results suggest that GBS activated with grain refinement and dominant deformation mechanism was changed by the strain rate from dislocation motion to GBS across ≈10–3 s–1 in UFG steel.
(2) From results of XRD measurements, EBSD analyses, and TEM observations, it was found that the dislocations tangled and cell structure was generated with increasing strain in CG steel, whereas UFG steel showed they piled up at GB without tangles. In addition, intragranular dislocation increased with increasing strain in CG steel but it decreased in UFG steel. These results demonstrate that absorption of dislocation and slip-induced GBS were activated by the decreasing grain size or strain rate.
The authors sincerely appreciate technical support of AFM observations from Prof. Yutaka Watanabe (Tohoku University) and funding support from 21st ISIJ Research Promotion Grant (incl. Ishihara/Asada Grant) of the Iron and Steel Institute of Japan.