2022 Volume 63 Issue 9 Pages 1258-1265
To elucidate the segregation behavior of solutes in Al–Si and Al–Cu binary alloys, specimens of several hypo-eutectic Al–Si alloys water-quenched at different stages of solidification or air-cooled from melt to room temperature were prepared and the distributions of Si concentration were analyzed across the primary dendrites. The Si distribution in the dendrites of the specimens water-quenched during primary solidification showed that Si concentration declined from the surface toward the center of the dendrite. In contrast, the specimens water-quenched after eutectic solidification or air-cooled from melt to room temperature showed that Si concentration increased from the surface toward the center of the dendrite. The diffusion distance of Si in the dendrites during the cooling process from the finish of solidification to room temperature was calculated according to Fick’s second law and showed good agreement with the measured value. Therefore, the segregation behavior with high Si concentration in the center of the dendrite was attributed to the diffusion of Si from the center to the surface of the dendrite, resulting in the precipitation on the adjacent Si phase during eutectic solidification.
This Paper was Originally Published in Japanese in J. JFS 93 (2021) 169–175. Reference 6) is deleted and the number of References is moved up from then.
JIS AC4CH alloy, because of its excellent balance of strength and ductility, is used as a casting material for the suspension parts of automobiles which require high strength as well as high reliability. These excellent mechanical properties are achieved through JIS T6 treatment (hereinafter, referred to as T6 treatment), i.e., through a process comprising solution treatment, quenching, and aging in sequence. This alloy, with a typical composition of Al–7 mass%Si–0.3 mass%Mg (hereinafter, referred to as Al–7Si–0.3Mg), is first solution-treated to homogenize the strengthening elements, Mg and Si, and granulate the eutectic Si phase, and is then quenched and aged to strengthen it via precipitation of homogenized Mg and Si. However, the duration of T6 treatment is as long as 5–8 h and needs to be shortened to reduce the energy consumption and manufacturing cost. Although rapid heating and high temperature solution treatments have been studied as techniques for reducing the time required for solution treatment, previous works merely focused on hardness changes and microstructural observations.1) Studies have also been reported on the behavior of solute elements during solidification and cooling, but they only verified the differences in the concentration distribution of solute elements for different cooling rates.2–5) Thus far, there are few reports on the transition processes of the concentration distributions of solute elements during solidification and cooling, or the influences of element species and their contents.
In this study, the segregation behavior of Si was clarified through structural observation and concentration analysis via an electron probe micro analyzer (EPMA) for Al–7Si alloy specimens quenched from various temperatures during solidification and cooling processes. Moreover, in order to observe the differences in the segregation behavior of Si as a function of content, concentration analyses were performed after casting hypo-eutectic Al–Si alloys with different Si contents. Further, the influence of the diffusion coefficients for solute elements on segregation behavior was elucidated through performing the same treatments and analyses on hypo-eutectic Al–Cu alloys, considering the fact that Cu has a lower diffusion coefficient in the α-Al phase than Si.
Al–(1.4, 3, 7, 10) Si and Al–(3.5, 15, 30) Cu alloys were prepared from pure Al with a purity of 99.8% and Al–25Si or Al–40Cu master alloys. For each composition of the alloys, a 1500 g sample was melted at 1023 K in a graphite crucible 120 mm in diameter and 158 mm in height. Subsequently, a slag-removing flux with a ratio of 0.3 mass% (hereinafter, referred to as %) relative to the weight of the alloy was added, stirred, and allowed to settle for 10 min to remove the inclusions, and the melt was poured into a room temperature steel mold 20 mm in diameter and 230 mm in height to obtain rod-shaped ingots. The rod-shaped ingots were cut into samples with an approximate length of 15 mm for investigation and analyses of segregation behavior during solidification and cooling processes. Table 1 shows the chemical compositions of the samples obtained by ICP analysis.
The Al–7Si alloy samples prepared in section 2.1 were remelted at 1023 K in a mullite crucible with an outer diameter of 25 mm and a height of 25 mm, then the crucible was moved into a square hollow 50 mm by 50 mm cut out in a refractory brick, and the temperature of the melt was monitored with a CA thermocouple installed in the center of the melt. When the temperatures of the melt reached the temperatures on the cooling curve shown in Fig. 1, the melt, together with the crucible, was put into water and cooled to room temperature.
Quenching temperatures and cooling curves of Al–7Si alloy specimens.
The solidification structures were observed with an optical microscope on the longitudinal section passing through the center line (hereinafter, referred to as the central cross-section) of the sample. Further, the concentration distribution of Si in the same cross-section was measured via EPMA surface and line analyses. The acceleration voltage, illumination current, and electron beam diameter were set to 15 kV, 100 nA, and 0 µm (the minimum setting value and equivalent to an actual diameter of 0.1 µm), respectively, for the EPMA analyses.
2.3 Effects of alloy composition and diffusion coefficients for solute elements on concentration distributions of solute elementsTo examine the effects of alloy compositions and the diffusion coefficients for solute elements in the α-Al phase on the concentration distributions of solute elements after cooling to room temperature, Al–(1.4, 3, 7, 10) Si and Al–(3.5, 15, 30) Cu alloy samples were also remelted and cooled to room temperature using the method described in section 2.2. The Si or Cu concentrations in the central cross-sections were measured using EPMA.
2.4 Preparation of standard samples for EPMA analysesAl–7Si and Al–15Cu alloy samples with a diameter of 20 mm and a length of 15 mm cut out from the rod-shaped ingots were homogenized and used as standard samples for the EPMA analyses. The homogenization conditions for Al–7Si and Al–15Cu alloy samples were chosen as 833 K for 3 h and 803 K for 24 h, respectively, and the homogenized concentrations of Si or Cu in primary α-Al dendrites (hereinafter, referred to as dendrites) shown in Fig. 2 were measured. Further, the solid solubilities of Si or Cu in Al–Si or Al–Cu binary alloys at different temperatures were calculated using the thermodynamic calculation software Thermo-Calc, as shown in Fig. 3. The equilibrium concentrations of Si and Cu in Al–Si and Al–Cu alloy samples after the homogenization treatment were calculated as 1.35% and 4.96%, respectively.
Distribution of solute concentration in homogenized Al–7Si and Al–15Cu alloy specimens.
Al–Si and Al–Cu binary phase diagrams calculated by Thermo-Calc.
Of the Al–7Si alloy samples A to O shown in Fig. 1, the cross-sectional structures of samples A to F, H, I, and O are shown in Fig. 4. Samples A to O exhibited significant structural changes. Samples H and I were quenched from temperatures above and below the eutectic temperature, respectively, while sample O was naturally cooled to room temperature. In sample A, which was quenched from 885 K immediately below the temperature of the beginning of primary solidification, fine dendrites were observed as shown in the outlined areas in Fig. 4. These fine dendrites are considered to have been generated during quenching and were also observed in some areas of sample B, which was quenched from 881 K. Acicular eutectic Si particles formed in sample E, which was quenched during supercooling below the eutectic temperature, 848 K. The acicular eutectic Si particles increased with the progress of solidification as observed in samples F and H, both of which were quenched from 850 K but had different solid fractions before quenching. No significant differences were observed by an optical microscope between the microstructures of samples I and O, where sample I was quenched from 841 K immediately after the completion of solidification and sample O was naturally cooled to room temperature in air. The above results indicate that the solidification structures obtained via the method of the present study can reflect the progress of solidification before quenching, thus the cooling rate in the present quenching method is considered to be sufficiently high to freeze the distributions of Si during solidification and cooling processes that existed just before quenching.
Microstructures of Al–7Si alloy specimens quenched at various temperatures indicated in the micrographs or cooled to room temperature.
Figure 5 shows the Si concentration distributions in representative samples from A to O in Fig. 1 obtained via EPMA line analyses. Samples A, D, E, H, I, J, K, L, and O were chosen as representative samples in order to observe the transition process of the Si concentration distribution during and after solidification. Samples A, D, E, and H were quenched from temperatures near the beginning or end of primary or eutectic solidification, and samples I, J, K, L, and O were quenched after the end of eutectic solidification. The dashed line is the maximum solid solubility of Si in the α-Al phase, 1.55%, and the dash-dotted line is the solid solubility of Si in the α-Al phase at the temperature from which the sample was quenched. On the right side of the figure, the positions of EPMA line analyses are indicated by arrows, along with the results of the EPMA surface analyses.
Distribution of Si concentration in Al–7Si alloy specimens quenched or cooled to room temperature.
From Fig. 5, the Si concentration in the dendrites of sample A, which was quenched from 885 K immediately after the beginning of primary solidification, was 1.0%. In sample D, which was quenched from 859 K near the end of primary solidification, the Si concentration in the outer peripheral areas of dendrites increased to approximately 1.5%, because the solid solubility of Si in the α-Al phase grows with decreasing temperature in the temperature range from the beginning of primary solidification to eutectic solidification, as shown in Fig. 3. Further, the Si concentrations in the central areas of the dendrites rose to approximately 1.2%, because Si diffusion from the peripheral areas to the central areas was enhanced by the increased Si concentration gradient, owing to the increase in the Si concentration in the outer peripheral areas. Such a distribution of Si concentration, i.e., higher Si concentrations in the outer peripheral areas than those in the central areas, was observed in samples up to E, which was quenched immediately after the end of primary solidification at 848 K.
The Si concentrations in the outer peripheral areas were stable at approximately 1.55% for the samples quenched during eutectic solidification, i.e., samples E and H, which were quenched from 848 K and 850 K, respectively, because the Si solid solubility in the α-Al phase remains at 1.55% (the maximum solid solubility) during eutectic solidification. On the other hand, Si continued to diffuse from the outer peripheral areas to the central areas, as observed in the samples quenched during primary solidification. Thus, the Si concentration in the central areas of sample H, which was quenched from 850 K, and sample I, which was quenched from 841 K immediately after the end of eutectic solidification, rose to approximately 1.4% or almost equal to that in the outer peripheral areas, respectively. In the samples quenched immediately after the end of eutectic solidification, the Si concentrations in the outer peripheral areas decreased slightly and became lower than those in the central areas. This type of Si concentration distribution was observed in some areas of samples I and J, which were quenched from 841 K and 828 K respectively. Subsequently, as the quenching temperature was lowered, the Si concentration in the peripheral areas decreased to approximately 1.25% and 1.0% for samples K and O, which were quenched from 793 K and naturally cooled to room temperature, respectively. As shown in sample L, which was quenched from 765 K, the decrease in Si concentration occurred within approximately 10 µm of the outer peripheries of the dendrites, and the range exhibiting a decrease in the Si concentration hardly changed during cooling from 765 K to room temperature.
In samples quenched from temperatures equal to or below 841 K, the above-stated decreases in Si concentrations in the areas near the outer peripheries of the dendrites were observed, but the Si concentrations in the central areas of the dendrites hardly changed from the maximum solid solubility of 1.55% even in the sample naturally cooled to room temperature. In some dendrites of samples K and O, small areas having a size of approximate 1 µm showed Si concentrations as high as approximately 2%, suggesting the precipitation of Si in some dendrites. However, this only occurred in very limited areas. Therefore, during the cooling process following eutectic solidification, most of the supersaturated solid solution of Si in the dendrites was presumed to have precipitated onto eutectic Si phases contacting the dendrites rather than precipitating inside the dendrites.
3.3 Effects of alloy composition and diffusion coefficient on Si concentration distribution in primary α-Al dendritesFigure 6 shows cooling curves for the Al–(1.4, 3, 7, 10) Si and Al–(3.5, 15, 30) Cu alloy samples. The time for local solidification from the liquidus to the solidus temperature and the rate of cooling by about 100 K after the end of solidification ranged from 120 to 140 s and from 2.0 to 2.4 K/s, respectively, and both showed no significant differences between alloy compositions. Primary solidification took more time than eutectic solidification in the Al–1.4Si and Al–3.5Cu alloy samples, while the situation was reversed in the Al–10Si and Al–30Cu alloy samples.
Cooling curves of Al–(1.4, 3, 7, 10) Si and Al–(3.5, 15, 30) Cu alloy specimens.
Figure 7 shows the Si concentration distributions in the central cross-sections of the Al–(1.4, 3, 7, 10) Si alloy samples after cooling to room temperature. In the Al–1.4Si and Al–3Si alloy samples, the segregation phenomenon showing a monotonic increase of Si concentration from the outer periphery to the center of the dendrites, as observed for the Al–7Si alloy sample, was not observed. Instead, the Si concentration in the central areas remained low and the Si concentration in the outer peripheral areas decreased by approximately 0.2 to 0.3%. In the Al–10Si alloy sample, segregation similar to that for the Al–7Si alloy sample was observed, i.e., the Si concentration in the central area of dendrites was higher than that in the peripheral areas. Therefore, it can be inferred that in hypo-eutectic Al–Si alloys, the higher the Si content or the closer to the eutectic composition the Si content is, the more likely that segregation occurs with a higher Si concentration in the central area of dendrites than in the outer peripheral area.
Distribution of Si concentration in Al–(1.4, 3, 7, 10) Si alloy specimens cooled to room temperature.
Figure 8 shows the Cu concentration distributions in the central cross-sections of the Al–(3.5, 15, 30) Cu alloy samples treated by the above method. Although segregation with higher Cu concentration in the central area than in the outer peripheral area was not observed for the Al–3.5Cu and Al–15Cu alloy samples, it did appear in the Al–30Cu alloy sample. In the hypo-eutectic Al–Cu alloy samples, it was also indicated that the higher the Cu content or the closer to the eutectic composition the Cu content, the more likely that segregation with a higher Cu concentration in the central area of dendrites than in the outer peripheral area occurs, as observed in the hypo-eutectic Al–Si alloy samples, despite the diffusion coefficient for Cu in Al being one-third that of Si.
Distribution of Cu concentration in Al–(3.5, 15, 30) Cu alloy specimens cooled to room temperature.
The above results suggest that segregation with Si concentrations higher in the central areas than in the outer peripheral areas occurs through a route comprising the following processes A and B.
Process A: temporary homogenization of Si or Cu via diffusion from the outer peripheral areas to the central areas of dendrites.
Process B: most of the supersaturated solid solution Si or Cu in the dendrites precipitate on the eutectic Si phase or the Al2Cu phase, resulting in lower Si or Cu concentrations in the peripheral areas.
Process A and B are presumed to mainly proceed during solidification and cooling after solidification, respectively. Moreover, process A is considered to possess the following 3 features.
These conditions become easily satisfied with hypo-eutectic alloys approaching the eutectic composition. In contrast, the progress of process B is almost independent of the composition of alloys, because the temperature drop during cooling is hardly affected by the composition, as shown in Fig. 6. However, it should be noted that even for Al–(7∼7.5)Si alloys, if the cooling rate is extremely high, segregation with Si concentrations in the central areas higher than those in the outer peripheral areas cannot occur, as observed in samples A to J in Fig. 5 and reported in previous reasearch.2–5)
From the above discussion, the progress of process A is critical to the occurrence of segregation with higher Si concentration in the central areas than in the outer peripheral areas, i.e., the time at the eutectic temperature, the gradient of Si concentration between the outer peripheral areas and the central areas of dendrites, and the arm spacings of dendrites are important factors for the segregation phenomenon. In addition, even if process A progresses, if process B does not proceed sufficiently, the above segregation cannot occur. That is to say, a comparatively low cooling rate after solidification is also important for the segregation to take place.
3.4 Theoretical examination with respect to diffusion distancesTo confirm that the above-stated temporary homogenization process (process A) and the segregation process (process B) were caused by the diffusion of solute elements during the eutectic solidification or the cooling processes after solidification, a theoretical examination of the Si concentration distributions was performed on Al–7Si alloy as an example.
3.4.1 Diffusion during eutectic solidificationWhether the temporary homogenization (process A) arose from the diffusion of solute elements during the eutectic solidification was first verified by theoretically computing the Si concentration distributions from the time immediately after the beginning of eutectic solidification (sample E) to the final stage of eutectic solidification (sample H) in Al–7Si alloy.
As shown in Fig. 9, the concentration distribution c in a plate with a thickness h is expressed by eq. (1),6) given that the initial concentration is cmin and the concentration on the surface (x = 0, h) at t > 0 is cmax.
| \begin{align} c &= \frac{-4(c_{\textit{max}} - c_{\textit{min}})}{\pi}\sum\nolimits_{j=0}^{\infty} \frac{1}{2j+1}\sin \frac{(2j+1)\pi x}{h}\\ &\quad \times \mathit{exp} \left[-\left\{ \frac{(2j+1)\pi}{h}\right\}^{2}Dt \right] + c_{\text{max}} \end{align} | (1) |
Schematic diagrams of concentration distribution in flat plate with constant surface concentration.
From the concentration distributions in Al–7Si alloy samples (samples E and H in Fig. 5) during eutectic solidification, the values of cmax, cmin, and h were obtained as 1.70%, 1.25%, and 26.5 µm, respectively. The diffusion coefficient D for Si in Al is given by eq. (2):
| \begin{equation} D = D_{0}\,\mathit{exp}\left(-\frac{Q}{RT}\right) \end{equation} | (2) |
Time change of Si concentration during eutectic solidification.
Although the calculated Si concentration distribution in sample E was slightly lower than the measured value, the difference was only approximately 0.05% to 0.1%. Therefore, the change in the Si concentration distribution from sample E to sample H can be ascribed to the diffusion of Si during eutectic solidification. The difference between the calculated and the measured values is attributable to the assumption in the calculation that the dendrites are flat plates. From the above examination, it was proved that even at a relatively low temperature of around 850 K, if the eutectic solidification time is long enough, Si can also diffuse sufficiently to bring about temporary homogenization; i.e., process A can occur.
3.4.2 Diffusion in the cooling process after solidificationIn order to confirm that segregation with higher Si concentrations in the central areas of dendrites than in the outer peripheral areas was caused by the diffusion of solute elements in solid phases during cooling after solidification, the diffusion distance of Si in the dendrites of Al–7Si alloy was calculated during the cooling process from the homogenization temperature 828 K to 765 K, at which the above segregation is almost finished, as shown in Fig. 5.
The solute concentration, C, in a solid phase is expressed by eq. (3) using the error function,8) when the solute diffuses from the surface where the solute concentration is CS into the solid phase with an initial concentration C, where, t, x, and D are time, the distance from the surface of the solid phase, and the diffusion coefficient for the solute in the solid phase, respectively.
| \begin{equation} \frac{C_{\text{s}} - C}{C_{\text{s}} - C'} = \mathop{\text{erf}}\nolimits \left(\frac{x}{2\sqrt{Dt}} \right) \end{equation} | (3) |
If the range within which the Si concentration decreases by 0.01 or more from the initial value due to diffusion of the solute to the surface is defined as the diffusion distance xmax, then xmax is obtained by setting CS = 1, C′ = 0, and C = 0.01 in eq. (3). In this case, the left side of eq. (1) is 0.99, and the maximum diffusion distance xmax can be expressed by eq. (4) by reading the value of the argument for the error function to equal 0.98994 from the error function table.9)
| \begin{equation} x_{\text{max}} \fallingdotseq 1.82 \times 2\sqrt{Dt} \end{equation} | (4) |
From the cooling curve shown in Fig. 6, the time required for cooling from 828 K to 765 K was t = 27.2 s. The average diffusion coefficient for Si from 828 to 765 K was calculated to be 2.69 × 10−13 m2/s from eq. (2) and was taken as the diffusion coefficient in this temperature range. Substituting the above values into eq. (4), the theoretical diffusion distance in the cooling process from 828 to 765 K is approximately 9.8 µm. This value agrees well with the range within which the Si concentration decreased near the outer peripheries of the dendrites in sample L, as shown in Fig. 5. Therefore, segregation with higher Si concentrations in the central areas of dendrites than in the outer peripheral areas can be considered to have been caused by Si diffusion in the solid phase after solidification.
Through examining the segregation of solutes in primary α-Al dendrites during the solidification and cooling processes of various Al–Si and Al–Cu alloys, the following conclusions were obtained.
