2023 Volume 64 Issue 8 Pages 1959-1968
Nanoindentation tests were conducted near the grain boundary (GB) of the Al–Mg–Si alloy, and the influence of GB character on the aging precipitation behavior and the mechanical properties was confirmed. After obtaining the GB characters by electron back scattered diffraction (EBSD) analysis and nanoindentation tests were carried out on under-aged, peak-aged, and over-aged samples. And then, the indentation areas were observed by back scattered electrons imaging (BSE) in order to identify indentation positions to the GB. In this study, for the GB character, focusing on the rotation angle, the high-angle GB (HAGB) and the low-angle GB (LAGB) were selected. In addition, coincidence site lattice GBs (CSL) were selected as the special GB. In the 180°C under-aged sample, the nano-hardness near GB is higher than that far from GB, while 180°C peak-aged and 250°C aged samples, the nano-hardness is lower than that far from GB. Then the amount of change in hardness of HAGB was larger than that of the LAGB. This suggests that the GB character affects the aging precipitation behavior and mechanical properties.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 85 (2021) 7–16.
Fig. 3 Nano-hardness around grain boundaries at aging temperature of 180°C; (a) at under-aged around HAGB, (b) at peak-aged around HAGB, (c) at under-aged around LAGB, (d) at peak-aged around LAGB, (e) at under-aged around CSL3, (f) at peak-aged around CSL3, (g) at under-aged around CSL5 and (h) at peak-aged around CSL5.
Al–Mg–Si alloys are used in automobile body panels because of their superior paint bake hardenability (age-hardenability), in addition to their workability and corrosion resistance compared to other aluminum (Al) alloys. Paint-bake hardenability means that the hardness increases by paint baking heat treatment after forming the car body, and the aging precipitates formed during heat treatment are believed to contribute to this hardenability. Therefore, it is necessary to understand the effect of aging precipitates on strength when developing products that use the age hardenability of Al–Mg–Si alloys.
When Al–Mg–Si alloys are subjected to aging, the aging products transition from a supersaturated solid solution (SSSS) to precipitates in the following process.1–3)
| \begin{align*} \mathrm{SSSS}&\Rightarrow \text{Mg–Si cluster}\Rightarrow \text{Guinier–Preston (GP) zone}\\ &\Rightarrow \beta''\Rightarrow \beta'\Rightarrow \beta \end{align*} |
It is well known that the dominant products in an alloy vary with the aging temperature and holding time. The aging products involved in the automobile body manufacturing process are the β′′ precipitates. β′′ (Mg5Si6) forms as coherent needle-like precipitates along the ⟨100⟩ direction of the Al matrix. β′′ precipitates are coherent with the matrix but behave as an obstacle to dislocation movement, which are considered to be responsible for the Orowan mechanism and significantly contribute to the strength of the alloy after aging. With continued heat treatment, the β′′ is eventually transformed by β′, a rod-like semi-coherent precipitate, and the final stable phase, plate-like incoherent precipitate β (Mg2Si). These precipitates contribute less to the strength than β′′ and are rarely used as the main strengthening phase. However, the β′ and β precipitates are considered to precipitate on the grain boundary depending on the heat treatment condition. It is necessary to clarify the nature of the intergranular precipitates to understand their effects on the strength and phenomena of intergranular fracture during tensile tests.
The balance between strength and ductility is important for structural materials. In material design, strength is increased by increasing the intragranular strength; however, the stress is concentrated at the grain boundary, which is the starting point of fracture, causing a decrease in ductility. Therefore, the influence of the grain boundaries cannot be ignored because of the balance between strength and ductility. In age-hardened Al–Mg–Si alloys, which are increasingly used as lightweight structural materials, there have been reports investigating the effects of grain boundaries on plastic deformation.4–11) It has also been reported that a precipitate-free zone (PFZ) and a region of heterogeneously dispersed coarse precipitates adjacent to the PFZ are formed near the grain boundaries of artificially aged Al alloys.12,13) The PFZ is considered to have a direct effect on the mechanical properties of alloys because the effect of precipitation strengthening is low in the PFZ. Therefore, studies have been conducted to relate PFZ to material properties such as strength, ductility, and corrosion resistance.13–16) Fine grain strengthening (grain boundary hardening) is another useful strengthening method for Al alloys. However, it has been reported that when the grain size is refined to the submicron level by severe plastic deformation, the precipitation strengthening ability decreases because of the competition between grain boundary precipitation and intragranular precipitation as the volume fraction of grain boundaries in the alloy increases, and the precipitation behavior in coarse grains is different from that in ultrafine grains. Therefore, it is necessary to deepen our understanding of the precipitation behavior near grain boundaries to provide fine grain strengthening without loss of strengthening ability, which has been studied in various Al alloys, especially in 7000 series Al alloys.12,17–22)
Cai et al. systematically classified the PFZ widths of Al–Zn–Mg–Cu alloys by grain boundary character and reported that the PFZ widths of Σ1a, Σ3, and Σ5, which are low Σ coincidence site lattice (CSL) grain boundaries, are narrower than those of random grain boundaries.23) Grain boundary character is a property of grain boundaries classified by the misorientation of neighboring grains (rotation angle/axis) and grain boundary plane orientation (inclination). This property is known to have a significant effect on various properties of materials. Random grain boundaries, which are general grain boundaries, have a high-energy structure and easily cause intergranular degradation, whereas low Σ CSL grain boundaries, which have a low-energy structure with regularly arranged atoms, are known to hardly cause intergranular degradation. Therefore, grain boundary engineering, the idea of introducing many CSL grain boundaries into a polycrystalline material to improve its properties, has been developed.24) Shimada et al. reported that the corrosion rate of SUS304L was reduced to 1/4 of that of the base metal by reducing the number of random grain boundaries, which can be the starting point of corrosion and fracture, and introducing many CSL grain boundaries.25) Thus, studies on the relationship between grain boundary character and material properties have attracted significant attention.26,27) For age-hardened Al–Mg–Si alloys in particular, microstructural analysis including the presence or absence of precipitates at and near grain boundaries is essential to improve strength and ductility, and studies have focused on precipitation behavior within and just above grain boundaries to increase strength and ductility.28,29) However, there have been few comprehensive studies of local mechanical properties and aging precipitation behavior near the grain boundary when the aging temperature and aging time are varied, and there have been no reports classifying the obtained results by grain boundary character.
In this study, grain boundaries were identified by microstructural observation and grain boundary character distribution (GBCD) of samples obtained from rolled Al–Mg–Si alloys after recrystallization, solution treatment, and aging. The effects of the grain boundaries on the aging precipitation and local mechanical behaviors of different grain boundaries were then investigated, and the results were classified in terms of the grain boundary character.
The composition of the Al–Mg–Si alloy used in this study is listed in Table 1. Cold-rolled Al–Mg–Si alloy sheets (1 mm thick) were cut into 15 mm × 10 mm pieces, and the specimens were prepared for the Vickers hardness and the nanoindentation tests. The specimens were subjected to recrystallization and solution treatment at 550°C for 1 h, and then quenched in iced water to produce an SSSS sample. The SSSS samples were artificially aged at 180 and 250°C in an oil bath and at 350°C in air using an electric furnace.
After the artificially aged specimens were quickly polished and buffed, they were subjected to the Vickers hardness test with a load of 29.4 N and loading time of 15 s to investigate the age-hardening behavior. The indenter was driven to seven points, and the average of five points, excluding the maximum and minimum values, was determined as the Vickers hardness value.
2.3 Nanoindentation testThe 180°C and 250°C under-aged samples (180U and 250U) were heat-treated under the conditions described below, then immediately subjected to mechanical polishing, and electropolished for 90 s at a voltage of 15 V. The electrolyte solution comprised a mixture of perchloric acid, ethylene glycol, and ethanol (1:2:7).
A Vickers hardness tester was used to indent each under-aged sample surface as a 2 × 1 mm rectangular marker. Using the indentation as a marker, crystal orientation analysis within a rectangular region was performed using an electron back scattered diffraction (EBSD) analysis system (TSL Solutions K. K. OIM D.C.) attached to a field emission scanning electron microscope (FE-SEM: JEOL JSM-6500F) to obtain the GBCD. The acceleration voltage was 15 kV and the step size was set to 10 or 3 µm. Back scattered electron (BSE) images were then obtained using FE-SEM to confirm the microstructure. The grain boundaries selected in this study were high-angle grain boundaries (HAGB) and low-angle grain boundaries (LAGB), focusing on the rotation angle and rotation axis, and Σ3 and Σ5 CSL grain boundaries as special grain boundaries. After confirming that the common rotation axis of the selected grain boundaries is close to the common rotation axis based on the CSL theory, CSL grain boundaries were defined as grain boundaries with angles within the maximum allowable misorientation angle, Δθ = 15/Σ1/2 (degrees), which can maintain the grain boundary structure proposed by Brandon.30) Tables 2 and 3 summarize the classification of grain boundaries selected in this study and the information on each grain boundary actually investigated, respectively.
From the obtained GBCD and BSE images, the relative distance from the grain boundary was confirmed, and the nano hardness at that location was measured using a nanoindenter (Bruker TI950). The load, measurement point interval, loading rate, and holding time, were set to 500 µN, every 1–3 µm, 50 µN/s, and 10 s, respectively. After the tests, the indentation position was confirmed using BSE images. The distance from the grain boundary at each indentation point was used as the vertical distance from the grain boundary.
After the nanoindentation tests on 180U and 250U, peak-aged samples (180P and 250P) were prepared by aging at the same temperature, and nanoindentation tests were performed on the same grain boundaries as those tested during under-aging.
The over-aged sample (350O) was prepared by aging at 350°C and then mechanically polished and buffed under the same conditions as above, after which the GBCD was obtained and nanoindentation tests were performed.
2.4 Microstructural observationUnder-aged and peak-aged samples were prepared at 180°C and 250°C, respectively, and then roughly polished to a thickness of 0.20 mm or less. Subsequently, φ3 mm discs were punched out. The samples were then polished to a thickness of 0.10 mm or less and subjected to twin-jet electropolishing to prepare samples for transmission electron microscopy (TEM). TEM (JEM-2000FX, JEOL, Ltd.) and energy-dispersive X-ray spectroscopy (EDS) analyses were performed on each sample to confirm the intra- and intergranular microstructure and composition.
Table 4 shows the GBCD and average grain size evaluated from EBSD analysis for the samples aged at 180°C, 250°C, and 350°C, and the SSSS sample. Figure 1 shows the inverse pole figure (IPF) map and grain boundary character map for the 180°C aged sample as an example. The crystallographic orientation distribution in the figure corresponds to the color key of a unit stereographic triangle. There was no regularity in the distribution of the grain boundaries, which were randomly distributed. The CSL grain boundaries of Σ3 and Σ5 contain only 1–3 traces within the observation area. Table 4 also shows that there were no significant differences in the GBCD and grain sizes of the samples after each aging process. This indicated that the GBCD did not significantly change during aging in this study.
Crystal orientation distribution maps of 180°C aged sample; (a) IPF map and (b) GBCD map.
Figure 2 shows the relationship between aging time and Vickers hardness for samples aged at 180°C and 250°C. The Vickers hardness of the SSSS sample is approximately 41 HV. For 180°C aging, hardening begins at 7.5 min after the start of aging and reaches a peak hardness of approximately 90 HV at 8 h. For 250°C aging, a peak hardness of 66 HV is reached at 30 min. Based on these results, 180°C-8 h and 250°C-30 min were selected as the peak aging conditions for each aging temperature in this study, and 180°C-1 h and 250°C-7.5 min were selected as under aging conditions. The aging conditions and names of the samples are listed in Table 5.
Age-hardening curves of samples aged at 180°C and 250°C. “U” and “P” in the figure indicate under-aged and peak-aged.
The average nano hardness distribution around the grain boundary during the 180°C aging is shown in Fig. 3, and the average hardness within the grain is shown in Table 6. Table 6 shows that the average nano hardness in the grains was 1.04 GPa for the under-aged sample (180U) and 1.25 GPa for the peak-aged sample (180P). Figures 3(a), (c), (e), and (g), show that the hardness near the grain boundaries was higher than that within the grains in the under-aged samples. However, Figs. 3(b), (d), (f), and (h), show that the hardness near the grain boundaries was smaller than that within the grains for the peak-aged samples. The range of hardness differences (indicated by the arrows in the figure) varied depending on the grain boundary character, and the range was larger for HAGBs than for LAGBs, Σ3 CSL grain boundaries (CSL3), and Σ5 CSL grain boundaries (CSL5).
Nano-hardness around grain boundaries at aging temperature of 180°C; (a) at under-aged around HAGB, (b) at peak-aged around HAGB, (c) at under-aged around LAGB, (d) at peak-aged around LAGB, (e) at under-aged around CSL3, (f) at peak-aged around CSL3, (g) at under-aged around CSL5 and (h) at peak-aged around CSL5.
The nano hardness distribution around the grain boundaries for the 250°C and 350°C aging treatments is shown in Fig. 4. Figures 4(a)–(d) and Table 6 show that the under-aged (250U) and peak-aged (250P) samples exhibited a decrease in hardness near the grain boundary. The average nano hardness value within the grain was greater for 250U (1.08 GPa) than for 250P (0.98 GPa), which was the opposite of the hardness relationship shown in Fig. 2 for the Vickers hardness test. The range of the hardness difference differed depending on the grain boundary character, which was similar to that observed for the 180°C-aged sample.
Nano-hardness around grain boundaries; (a) around HAGB in 250U, (b) around HAGB in 250P, (c) around CSL5 in 250U, (d) around CSL5 in 250P, (e) around LAGB in 350O and (f) around CSL5 in 350O.
Figure 4(e) shows that there is a monotonic increase in nano hardness from within the grain to the grain boundary in the vicinity of LAGB in the over-aged sample at 350°C (350O). However, the hardness near CSL5 decreased with increasing distance from the grain boundary and increased within 2 µm from the grain boundary, indicating that the nano hardness was greater than the intragranular hardness. Figure 5 shows a BSE image of the vicinity of CSL5 after nanoindentation tests in 350O. The presence of PFZ along the grain boundaries was also confirmed. PFZs were also observed in HAGB, LAGB, and CSL3, and the width of the PFZs varied depending on the grain boundary character. The PFZ width, wPFZ was measured from the BSE image of each grain boundary. The relationship between the hardness of the PFZ and DGB/wPFZ, where DGB is the distance from the grain boundary normalized by wPFZ, is shown in Fig. 6. In range (I), the hardness is smaller than the average hardness within the grain, and the smallest nano hardness, 0.50 GPa, was obtained at the point near the center of the PFZ to the grain boundary (0.6). In range (II), the hardness increases as the PFZ approaches the grain boundary, whereas in range (III), the opposite trend of decreasing hardness is observed.
Images of after nanoindentation test around CSL5 in 350O; (a) IPF map, (b) GBCD map and (c) BSE image.
The relation between the nano-hardness on PFZ and the distance from GB normalized by PFZ width in 350O.
Figure 7 and Table 7 show TEM bright field images obtained under 180°C and 250°C aging conditions, and the shape, average radius and length, and number density of precipitates measured through microstructural observations. From Figs. 7(a) and 7(b), needle-like and dot-like contrasts along the ⟨100⟩ direction can be observed in the matrix at 180°C aging, which suggest β′′ precipitation. The number density measured from the needle- and dot-like contrasts was higher for 180P than for 180U. Figure 7(c) and (d) show that PFZs of approximately 100 nm were observed along the grain boundaries. In addition, precipitates larger than those within the grains were sparsely dispersed in the 100 nm region adjacent to the outer PFZ. On the grain boundary, precipitates were observed in 180U and 180P, and the needle-like contrast in 180U indicated by the white arrow suggests that the precipitates on the grain boundary are β′′. In 180P, a spherical contrast was observed. Figure 8 shows the TEM bright-field image of 180P in a different field of view from that in Fig. 7. The EDS analysis results of the areas (i) and (ii) in the TEM image are also shown in Fig. 8. Intergranular precipitates of different shapes are observed at the grain boundaries, as shown in Fig. 7(d). Figure 8(a) shows the average composition of region (i), which includes the composition of the surrounding Al matrix. Since the Mg/Si ratio is 2:1, the grain boundary precipitates are β (Mg2Si).
BF-TEM images of the aged samples; (a) grain interior in 180U, (b) grain interior in 180P, (c) around grain boundary in 180U, (d) around grain boundary in 180P, (e) in 250U and (f) in 250P. White arrow indicate precipitates and black arrow indicate grain boundaries.
BF-TEM images and TEM/EDS analyses in 180P; (a) on grain boundary precipitate and (b) on grain interior.
Figure 9 shows the TEM images of the grain boundary at 180U, the results of the EBSD analysis after TEM observation, and the changes in the number density of precipitates near the grain boundary. The number density values are averaged over a range of 0.50 µm, including 0.25 µm on either side of the measurement point. The results show that the number density of β′′ increased from approximately 3.5 µm to 1 µm from the low-angle grain boundary.
Correspondence between TEM observation area and EBSD analysis area in 180U; (a) BF-TEM image around LAGB (13.8°) and (b) GBCD map. The circle of dots line indicates TEM observation area. And (c) number density of precipitates around LAGB in 180U.
Figure 7(e) and (f) show that longer rod contrasts were observed in the grains of the 250°C-aged sample compared to the precipitates in the 180°C-aged sample, suggesting that β′ precipitates. An elliptical contrast is also observed at the grain boundaries. Precipitates were observed outside the PFZ in the 250°C-aged sample and the sample aged at 180°C.
In Al–Mg–Si alloys, β′′ is the dominant precipitate in Al–Mg–Si alloys during aging at 180°C. It is known that β′′ improves the intragranular strength via dense dispersion. Figures 7(a) and (c) show that some β′′ precipitates are observed in the under-aged sample, which can be confirmed by the diffraction contrast. However, the number density of precipitates is smaller than that of the peak-aged sample, suggesting that Mg–Si clusters and GP zones, which are not visible, exist densely within the grains. As the aging time is increased, the dominant Mg–Si clusters and GP zones transform to β′′ and the number density of these precipitates increases. The average hardness of the peak-aged samples increased compared to the under-aged samples.
In Figs. 7(b) and (d), precipitates larger than the intragranular precipitates are observed inside and adjacent to the PFZ. This precipitation behavior is consistent with the precipitate sparse zone (PSZ) or transition zone reported by Ma et al. and Ogura et al. in the Al–Zn–Mg system alloys.12,13) Ogura et al. reported that the Zn and Mg concentrations (degree of supersaturation) in SSSS in the matrix phase decrease close to the grain boundary during aging and that the average Zn and Mg concentrations in the PFZ also decrease as the aging time is increased. Miyazaki et al. also systematically investigated microstructural changes with composition using “the composition gradient method”.31–33) For example, in Ni–Al alloys, when an Ni–15 at% Al alloy and pure Ni were joined using arc melting, the solute concentration in the matrix phase decreased as it approached the interface between the different phases. In addition, it has been reported that the precipitates in the low-concentration region near the interface are much more sparsely distributed than those in the high-concentration region away from the interface.33) This is because the precipitate size is dependent on the equilibrium composition at the interface between the matrix and precipitate, and precipitation on the low-concentration side follows the Gibbs-Thomson equation:34)
| \begin{equation} C_{e}(r) = C(\infty)\left[1 + \left(\frac{2\gamma_{s}V_{m}}{rRT}\right)\right] \end{equation} | (1) |
Here, Ce(r) is the equilibrium solute concentration as a function of the precipitate size, γs is the interfacial energy between the precipitate and the matrix, Vm is the molar volume of the precipitate, r is the critical radius, R is the gas constant, T is temperature, and C(∞) is the equilibrium solute concentration when the precipitate size is infinite. According to this equation, the critical radius for precipitation was large; and therefore, the precipitate size was larger at low concentrations.
Two theories have been proposed for the mechanism of PFZ formation: 1) vacancy depletion, which is caused by depletion near grain boundaries because quench excess vacancies required for precipitation diffuse to grain boundaries and disappear on them, and 2) solute depletion, which is caused by solute elements diffusing to grain boundaries to form intergranular precipitates. Ogura et al. explained that the wide (>200 nm) PFZ formed during under-aging in Al–Zn–Mg alloys was because of vacancy depletion by obtaining atomic maps near the grain boundaries using the 3DAP method.20) From this result, it is believed that the number density of the precipitates decreases with a reduced vacancy concentration near the grain boundary.
Figure 9 shows that in the vicinity of the 180U LAGB, in addition to the PFZ and PSZ, which are both approximately 100 nm in width, a “growth enhancement area” with a high number density is formed over approximately 2 µm.
Given that the above phenomena occurred, the PSZ suggested by the present experimental results was formed by the following process:
In 180U, owing to the short aging time, Mg–Si clusters and GP zones were dispersed in the grains without sufficient formation of the reinforcing phase, β′′. In the “growth enhancement area”, many of the clusters and GP zones transform to β′′, which may result in an increase in the number density of β′′ compared to the grains. On the PSZ located approximately 200 nm from the grain boundary, large-sized β′′ is considered to have precipitated because of the decrease in solute concentration. This suggests that the hardness near the grain boundary at 180U was higher than that within the grains.
In 180P, clusters and GP zones transform to β′′ within the grains, resulting in an increase in precipitation strengthening. Conversely, the hardness decrease near the grain boundary is believed to be because of a relative decrease in precipitation strengthening, as the number density decreases because of more precipitate growth in the “growth enhancement area” and on the PSZ.
In the 250°C aged sample, the increase in aging temperature decreases the supersaturation, resulting in a decrease in precipitation strengthening because of the sparse dispersion of β′, which is larger than β′′, in the grains. In addition, as with the 180°C aged sample, the decrease in solute and vacancy concentrations resulted in a sparse distribution of large precipitates and the formation of a PSZ. Furthermore, unlike in 180U, β′ precipitates which are not the strengthening phase grew in the “growth enhancement area” and were more sparsely distributed than β′ of the grains, suggesting that even the under-aged sample exhibited lower hardness near the grain boundaries than in the grains. The inverse relationship between the nano hardness of 250U and 250P within the grain and the Vickers hardness may be related to the effect of supersaturated solute atoms during under-aging and the size of precipitates in the plastic deformation area of the nanoindentation, which requires further investigation.
Based on the results of this study, Fig. 10 presents a schematic diagram showing the relationship between nano hardness and precipitated microstructure near the grain boundary for each aging condition. The plastic deformation area when an indenter was pressed into the sample was considered to be 1.5 times the indentation diameter reported by Itokazu et al.35) Nano hardness is determined by the interaction between the dislocations introduced during indentation and factors such as precipitates in the plastic deformation area formed during indentation. In 180U, the hardness increases with increased proximity to the grain boundary because the plastic deformation area contains more of the strengthening phase β′′, which is densely dispersed within the “growth enhancement area”. In the extreme vicinity of the grain boundary, the PSZ and PFZ, which were less precipitation-strengthened, were included in the region, and the increase in hardness became slow (Fig. 10(a)). In the “growth enhancement area”, the β′′ in 180P and β′ in 250U and 250P decrease in number density as their growth progresses, and the decrease in the fraction of precipitates in the plastic deformation area is considered to be the cause of the decrease in hardness (Fig. 10(b)). To the best of our knowledge, no similar reports exist on the “growth enhancement area” in age-hardening alloys such as Al–Mg–Si alloys. However, the detailed mechanism of formation requires further investigation.
Schematic diagram showing the effect of each factor near the grain boundary on the local mechanical behavior, which varies with aging conditions; (a) in 180U and (b) in 180P, 250U and 250P.
The distribution of nano hardness under each aging condition showed differences in the mechanical properties depending on the grain boundary character.
It has been suggested that LAGBs with a relative misorientation of 11° or less have the largest diffusion activation energy of solute atoms among all grain boundaries, suggesting that LAGBs are much more difficult to diffuse than HAGBs.23) In addition, among HAGB, the CSL3 and the CSL5 are less likely to cause solute atom diffusion than random grain boundaries because the two crystals composed of the grain boundary are well matched and the atomic arrangement is less disordered. Conversely, random grain boundaries have a low atomic density near the grain boundary, which allows solute atoms to diffuse easily, and they are believed to contain more nucleation sites than LAGBs and low Σ CSL grain boundaries. Because the differences in atomic density near grain boundaries for each grain boundary character have been reported to affect the segregation of solute atoms and vacancy concentration during quenching,36,37) the solute and vacancy concentration gradients near grain boundaries are expected to be large near random grain boundaries and small at LAGBs and low Σ CSL grain boundaries. Therefore, considering that the grain boundary precipitates formed by grain boundary diffusion during aging grow by the diffusion of solute atoms near the grain boundary to the grain boundary, they are more easily diffused at random grain boundaries; therefore, they form a large solute concentration gradient as aging proceeds. This results in the formation of PFZ, PSZ, and the “growth enhancement area”. Figure 3 shows that the range of hardness change observed near the LAGB, CSL3, and CSL5, was smaller than that of the HAGB. It is considered that the small amount of solute segregation at these grain boundaries during quenching and the difficulty of grain boundary diffusion reduced the volume fractions of the precipitates.
This may have resulted in a narrowing of the region of solute/vacancy concentration gradient change and of the PFZ, PSZ, and “growth enhancement area” widths.
Figure 4(e) and (f) show that the hardness tendencies of the LAGB and CSL5 are different at 350O. Figure 11 shows a schematic diagram of the relationship between nano hardness and precipitate structure in 350O with a PFZ of a few micrometers. In the case of indentation just above the wide PFZ, the nano hardness was significantly reduced owing to the influence of only the solid-solution Mg and Si atoms in the matrix (Fig. 11(a)), as there were no PSZ or intergranular precipitates in the plastic deformation zone. In contrast, in the case of indentation just above the narrow PFZ, the precipitates on the PSZ and the grain boundary precipitates were included in the plastic deformation area, and their influence was also affected; thus, the decrease in nano hardness was suppressed (Fig. 11(b)). Therefore, there was no hardness reduction when indenting into a narrow-width PFZ near the LAGB. In region (I) of Fig. 6, the increase in hardness away from the grain boundary was because of the increased contribution of precipitation strengthening caused by precipitates in the PSZ. Next, the hardness increase behavior in the region (II) of Fig. 6 is similar to that observed during aging at 180°C. However, in 180U, the hardness is indented into the “growth enhancement area” located outside the PFZ, while in 350O, the hardness is indented just above the PFZ. This is considered to be a different cause of hardening than that observed at 180U. Tsurekawa et al. reported that the local mechanical behavior near the grain boundaries of an Fe–3 mass%Si alloy hardens regardless of the grain boundary character, and that the stress required for strain transfer to adjacent grains varies with the grain boundary character.38) In addition, Tokuda et al. reported that in pure Al, the critical stress required for the formation of dislocations decreases toward the grain boundary. The stress required to transfer the strain to adjacent grains was also calculated, and the effect of grain boundaries on hardness was quantitatively evaluated.39) Although a more detailed investigation is needed because the sample used in this study is different from the alloy, the increase in hardness approaching the grain boundary at 350O is believed to be related to the dislocation pile-up owing to the different slip planes in adjacent grains, which stop dislocations introduced during deformation at the grain boundary. Sekido et al. analyzed the morphological changes in plastic deformation by plotting (P/h)-h for load P and displacement h obtained from nanoindentation tests.40) To understand the decrease in hardness in the extreme vicinity of the grain boundary in region (III) of Fig. 6, it is necessary to investigate the formation of dislocations using the above-mentioned method.
Schematic diagram showing the influence of factors near grain boundaries on local mechanical behavior at 350O with a PFZ width of a few µm; (a) around HAGB and (b) around LAGB.
In addition, this experiment did not consider the effects of the grain boundary plane and the inclination angle of the grain boundary relative to the sample surface. Furthermore, to investigate the grain boundary character dependence of the effects of the PFZ, PSZ, and grain boundary precipitates on the local mechanical behavior, it is useful to normalize the PFZ width and organize the nano hardness, as shown in Fig. 6. However, it is difficult to use this method because the effects of the PFZ and PSZ cannot be considered separately under aging conditions with PFZ widths of several hundred nanometers. Therefore, as shown in Fig. 9, the relationship between the nano hardness change near the grain boundary and the precipitation behavior can be continuously organized into TEM images to a position far from the grain boundary, and the number density of precipitates can be calculated. Electron beam diffraction and EBSD analysis can also be conducted on the sample observed by TEM to obtain the grain boundary character. However, this investigation was only conducted for 180U LAGB, and it is necessary to continue the investigation of other aging conditions and grain boundaries in the future.
These results indicate that the aging precipitation behavior is dependent on the grain boundary character, which causes changes in the mechanical properties.
Nanoindentation and TEM observations were performed on Al–Mg–Si alloys to investigate the effect of grain boundaries on the aging precipitation behavior and local mechanical behavior of different grain boundaries, and the results were classified in terms of the grain boundary character. The results obtained are summarized as follows.
This work was financially supported by JSPS KAKENHI (Grant Numbers JP16H04502 and JP19K22036) and the Light Metal Educational Foundation, Inc. Parts of this work were also conducted at the Laboratory of Nano-Micro Materials Analysis and High-Voltage Electron Microscope Laboratory, Hokkaido University, supported by “Nanotechnology Platform Program” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
