Journal Information
Vol. 8. Issue 3.
Pages 2930-2943 (May - June 2019)
Share
Share
Download PDF
More article options
Visits
439
Vol. 8. Issue 3.
Pages 2930-2943 (May - June 2019)
Original Article
DOI: 10.1016/j.jmrt.2019.05.001
Open Access
Simultaneous refinement and modification of the eutectic Si in hypoeutectic Al–Si alloys achieved via the addition of SiC nanoparticles
Visits
439
Dapeng Jiang
Corresponding author
jdp@mail.nwpu.edu.cn

Corresponding author.
, Jiakang Yu
State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an 710072, PR China
This item has received
439
Visits

Under a Creative Commons license
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (12)
Show moreShow less
Tables (3)
Table 1. Chemical composition of A356.
Table 2. Parameters for the calculation of the critical nanoparticle radius based on Eq. (1)[28,31,32].
Table 3. Parameters for the calculation of the critical nanoparticle velocity based on Eqs. (4) and (5) [31,32].
Show moreShow less
Abstract

This paper reports that the addition of SiCnps is demonstrated to induce the simultaneous modification and refinement of the eutectic Si in the SiCnps/A356 samples, which becomes increasingly significant with increasing SiCnps addition. The effect of SiCnps addition is primarily investigated via the results obtained from differential scanning calorimetry and high resolution transmission electron microscopy. The modification and refinement of the eutectic Si are closely related to the interaction between the SiCnps and the solidification front (SF). When the particle size of the SiCnps is greater than the critical size, the SiCnps tend to move ahead of the SF, while smaller SiCnps tend to be engulfed. Here, SiCnps distributed at the interface between the eutectic Si/Al phase hinder the growth of the eutectic Si. Meanwhile, SiCnps are observed to act as heterogeneous nucleation sites for the eutectic Si. Hence, the simultaneous modification and refinement of the eutectic Si is mainly facilitated by a combination of the above two individual effects of SiCnps addition. Furthermore, SiCnps distributed in the eutectic Si leads to the formation of multiple Si crystal twin variants.

Keywords:
Refinement
Modification
The eutectic Si
Distribution
Full Text
1Introduction

The low density, good castability, and excellent mechanical properties of the A356 Al–Si alloy make it well suited for wide use as a structural material in various fields such as transportation and aerospace [1,2]. The A356 alloy is characterized by Al dendrites surrounded by an Al–Si eutectic system (i.e., a eutectic Si). However, the microstructure of the eutectic Si phase has a substantial influence on the mechanical properties of A356. Here, ensuring good mechanical properties for the alloy requires that the eutectic Si phase be modified from its originally coarse flake-like microstructure and refined to obtain a finer fibrous microstructure. Therefore, developing strategies for the effective modification and refinement of the eutectic Si microstructure in Al–Si alloys is of significant importance.

The conventional strategy for modifying the microstructure of the eutectic Si phase in Al–Si alloys employs the addition of a variety of alloying elements, such as Na, Sr, and Eu [3–5]. Several studies have observed that high density twin crystal growth is related to the modification of eutectic Si microstructures in Al–Si alloys [6–8]. Presently, a few widely accepted mechanisms have been adopted for explaining the modification of eutectic Si microstructures in Al–Si alloys, such as the impurity induced twinning (IIT) and twin plane reentrant edge (TPRE) growth mechanisms, as well as poisoning of the TPRE [9,10]. The formation of Al2Si2Sr within the eutectic Si has a negligible effect on the modification of the eutectic Si microstructure in Al–Si alloys [11]. However, recent investigation has shown that Al2Si2Sr and NaAlSi clusters formed at the Si/liquid interface exert a significant influence on the modification of the eutectic Si microstructure by altering the eutectic Si growth process [12]. In addition, other elements, such as Yb, Ca, and P, have been added into Al–Si alloys for refining the eutectic Si microstructure [13–15]. For example, the addition of Yb was shown to lead to the refinement of the eutectic Si microstructure rather than the modification of the eutectic Si microstructure during the eutectic Si growth process [10]. In addition, the combined addition of Ca and P was shown to lead to a deactivation of the AlP impurity particles, resulting in the poisoning of AlP nucleation sites for the eutectic Si, which promoted refinement via an increased recalescence undercooling of the eutectic Si [16,17].

Compared with the conventional modification and refinement strategy discussed above, the modification and refinement of not only the eutectic Si phase in Al–Si alloys can be achieved simultaneously but also the failure of the modification is avoided via the addition of nanoparticles. Numerous investigations have shown that nanoparticles can control the eutectic Si phase growth during solidification [18–20]. Nanoparticles, such as Al2O3 and TiCN, have been shown to hinder the growth of the eutectic Si phase by distributing on the Al/Si interface, and thereby lead to the refinement of the eutectic phase [21,22]. In addition, the refinement of the eutectic Si phase in Al–Si alloys has also been shown to be related to the heterogeneous nucleation of TiCN nanoparticles and pre-nucleation clusters during solidification [23,24]. It has also been reported that the addition of AlN nanoparticles in Al–Si alloys modified the eutectic Si from a flakelike microstructure to a block-like microstructure [25]. However, the role played by SiC nanoparticles in the modification and refinement of the eutectic Si in Al–Si alloys remains poorly understood. Therefore, additional research focused on this very important strategy is required.

To address these issues, the present study prepares A356 Al–Si alloys with the addition of 0.5–2.0wt% SiC nanoparticles (SiCnps) by a solidification process combined with ultrasonic treatment. The effect of SiCnps addition on the modification and refinement of the eutectic Si in Al–Si alloys is systematically studied primarily via differential scanning calorimetry (DSC) and high resolution transmission electron microscopy (HRTEM). The simultaneous modification and refinement of the eutectic Si is shown to be mainly facilitated by a combination of the two effects of SiCnps addition that are closely related to the distribution of the SiCnps. These include the following. (1) The effect of SiCnps distributed at the interface between the eutectic Si and Al phase for hindering the growth and promoting the fragmentation of the eutectic Si. (2) The effect of SiCnps to act as heterogeneous nucleation sites for the eutectic Si. Furthermore, the effect of SiCnps distributed in the eutectic Si for promoting the formation of multiple Si twin crystal variants.

2Experimental procedure

The chemical composition of A356 is listed in Table 1. The A356 samples were melted in an alumina crucible using an electric resistance furnace, and SiCnps (purchased from Shanghai Yao Tian Nano Material Co. Ltd.) were added into the molten A356 with varying additions of 0.5, 1.0, and 2.0wt%. The particle sizes of the added SiCnps are presented in detail in Section 3. The apparatus and solidification process combined with ultrasonic treatment employed for preparing SiCnps/A356 materials are described in detail elsewhere [26]. The melt was cast using a cast iron cylindrical mold preheated to 400°C. For comparison, A356 samples without the addition of SiCnps were also prepared by an equivalent process.

Table 1.

Chemical composition of A356.

Alloy  Elements (wt%)
  Si  Mg  Al 
A356  6.5–7.5  0.3–0.45  BAL 

Metallographic samples were sectioned 51mm from the bottom of the casting, and then ground with 400 grit, 800 grit, and 1200 grit emery papers in turn. Subsequently, the samples were polished and lightly etched using a 0.5% aqueous HF solution. In addition, samples were also deeply etched for 1, 2, or 3h to reveal the three-dimensional (3D) morphology of the eutectic Si.

The morphology of the eutectic Si was characterized using optical microscopy (OM; Olympus MPG4, Tokyo, Japan) and scanning electron microscopy (SEM; FEI Nova NanoSEM 450, Hillsboro, USA) with energy dispersive X-ray spectroscopy (EDS; INCA X-Max, Oxford, UK). The dimensions of the distributed eutectic Si phase and the particle sizes of the SiCnps were evaluated from micrographs using Image-Pro Plus (6.0, Media Cybernetics, Rockville, MD, USA). X-ray diffraction (XRD; PANalytical X’Pert PRO, Almelo, The Netherlands) was used to analyze the crystalline phases of A356 and SiCnps/A356. The interfaces between the SiCnps and the eutectic Si, and the distributions of SiCnps in the eutectic Si and primary α-Al regions were investigated using HRTEM (Tecani F30 G’, FEI, USA). Finally, DSC (STA 449F3, NETZSCH, Germany) was used to investigate the nucleation behavior of SiCnps/A356 samples processed by the droplet emulsion technique [27].

3Results

Transmission electron microscopy (TEM) bright field images and the size distribution of the SiCnps are shown in Fig. 1(a) and (b), respectively. As such, the diameters of the SiCnps mainly ranged from 10nm to 80nm with an average diameter of about 40nm.

Fig. 1.

TEM bright field image of SiCnps (a) and the particle size distribution of SiCnps (b).

(0.33MB).

The XRD patterns obtained for A356 and 2.0wt% SiCnps/A356 samples are shown in Fig. 2(a) and (b), respectively. As can be seen from Fig. 2(a), the SiCnps/A356 sample includes three crystalline phases, namely Al, Si, and SiC. From Fig. 2(b), we note that A356 is composed of only Al and Si phases. Comparing Fig. 2(a) and (b) indicates that the SiCnps were effectively incorporated into the A356 alloy, and no other intermetallic crystalline materials were detected. As such, the SiCnps did not perceptibly react with the matrix alloy.

Fig. 2.

XRD patterns of the 2.0wt% SiCnps/A356 sample (a) and A356 alloy (b).

(0.1MB).

Fig. 3 presents optical micrographs of A356 and SiCnps/A356 samples with SiCnps additions of 0.5, 1.0, and 2.0wt%. As can be seen from the marked regions in Fig. 3, the addition of SiCnps has an obvious effect on the microstructure of the eutectic Si. Here, Fig. 3(a) presents a eutectic Si phase for the A356 sample as a long flake-like microstructure, with coarse particle sizes of several micrometers. However, the refinement in the microstructure of the eutectic Si becomes increasingly significant with increasing SiCnps content. For the 0.5wt% SiCnps/A356 sample, we note that, while the sizes of the eutectic Si microstructures are smaller than those in the A356 sample, the microstructure remains flake-like. Finally, the eutectic Si microstructure becomes greatly refined for the 1.0 and 2.0wt% SiCnps/A356 samples.

Fig. 3.

Optical micrographs of A356 alloy (a), the 0.5wt% SiCnps/A356 sample (b), the 1.0wt% SiCnps/A356 sample (c), and the 2.0wt% SiCnps/A356 sample (d).

(0.7MB).

Cross-sectional SEM micrographs were captured to more clearly reveal the morphology of the eutectic Si, as shown in Fig. 4. Here, Fig. 4 presents the overall cross-sectional morphologies of the A356 sample and the SiCnps/A356 samples with 0.5, 1.0, and 2.0wt% SiCnps. These figures clearly reveal that the lengths of the eutectic Si microstructures are greatly decreased with increasing SiCnps content and the shapes of the eutectic Si are also modified significantly and approach an equiaxed microstructure. We note that the lengths of the eutectic Si in the 0.5wt% SiCnps/A356 sample are reduced relative to those of the A356 sample, and the morphology of some of the eutectic Si have changed from a flake-like microstructure to a polygonal microstructure. However, most of the eutectic Si in the 0.5wt% SiCnps/A356 sample retain a flake-like microstructure. With increasing SiCnps content from 0.5wt% to 1.0wt%, we note that the eutectic Si microstructure lengths become substantially reduced, and their length-to-width ratios (i.e., the aspect ratios) are also reduced. With further increase in the SiCnps content to 2.0wt%, the lengths of eutectic Si microstructures are further reduced, and the shapes become nearly equiaxed. As such, the addition of SiCnps both refines and modifies the microstructure of the eutectic Si in A356.

Fig. 4.

SEM micrographs of A356 with different SiCnps additions: (a) A356; (c) 0.5wt% SiCnps/A356; (e) 1.0wt% SiCnps/A356; (g) 2.0wt% SiCnps/A356. In addition, (b), (d), (f), and (h) represent the areas marked in subfigures (a), (c), (e), and (g), respectively, at greater magnification.

(1.63MB).

The eutectic Si microstructure length and aspect ratio distributions of the A356 and SiCnps/A356 samples are presented in Fig. 5. We note that the average lengths of the eutectic Si microstructures are reduced slightly from 15μm to 13.7μm when introducing 0.5wt% SiCnps into the A356 alloy, and that the change in the aspect ratio distribution of the eutectic Si microstructures is also slight. However, the addition of 1.0wt% SiCnps into the A356 alloy substantially decreases the average eutectic Si microstructure length to 1.36μm (Fig. 5(e)), and the aspect ratio distribution (Fig. 5(f)) is skewed to smaller values characteristic of a polygonal microstructure. Finally, we note from Fig. 5(g) that the refinement of the eutectic Si is very significant with the addition of 2.0wt% SiCnps, where approximately 97% of the measured eutectic Si microstructures are less than 2μm in length and their average length is 0.85μm. Furthermore, the aspect ratio distribution (Fig. 5(h)) is further skewed to smaller values characteristic of a near-equiaxed microstructure.

Fig. 5.

Eutectic Si particle length distributions for A356 with different SiCnps additions: (a) A356; (c) 0.5wt% SiCnps/A356; (e) 1.0wt% SiCnps/A356; (g) 2.0wt% SiCnps/A356. In addition, (b), (d), (f), and (h) represent the aspect ratio distributions of A356, 0.5wt% SiCnps/A356, 1.0wt% SiCnps/A356, and 2.0wt% SiCnps/A356, respectively.

(0.5MB).

SEM micrographs of samples etched for 3h are presented in Fig. 6. We note that the 3D morphology of the eutectic Si in A356 is characterized by a large plate-like microstructure. Furthermore, some step-like formations are observed on the parallel platelets of the eutectic Si. These formations are representative of the layer growth behavior of the eutectic Si in A356 [6]. Comparing Fig. 6(a) and (b), we note that the morphology of a portion of the eutectic Si has transformed from platelets to block-like microstructures. However, the addition of 1.0wt% SiCnps into the A356 alloy substantially alters the 3D morphology of the eutectic Si from relatively coarse platelets and blocks to a dendritic microstructure with some fine platelets (Fig. 6(c)). Finally, as can be seen from Fig. 6(d), the addition of 2.0wt% SiCnps into the A356 alloy produces the eutectic Si phase with a strongly dendritic microstructure that is nearly characteristic of a fibrous morphology. The effect of the addition of SiCnps on the eutectic Si morphology may be related to a modification of the local diffusion behavior of the solute during crystal growth with the increasing addition of SiCnps. Hence, SiCnps make a difference in the growth of the eutectic Si.

Fig. 6.

SEM micrographs of A356 samples with different SiCnps contents that were etched using a 0.5% HF-water solution for 3h: (a) A356; (b) 0.5wt% SiCnps/A356; (c) 1.0wt% SiCnps/A356; (d) 2.0wt% SiCnps/A356.

(1.38MB).

To investigate the effect of SiCnps on the microstructure of the eutectic Si, we present SEM micrographs of 2.0wt% SiCnps/A356 samples etched for 1h and 2h in Fig. 7(a) and (c), respectively. From Fig. 7(a) and (b), we note that layers of SiCnps are observable on the surfaces of the eutectic Si. The SiCnps on the surface of the eutectic Si can be distinguished clearly in Fig. 7(b), and the average particle size is about 100nm. As such, the average particle size of SiCnps distributed on the surface of the eutectic Si is much greater than the average particle size of the original SiCnps (i.e., 40nm [Fig. 1(b)]). Therefore, we confirmed the composition of the SiCnps layer by applying EDS analysis to the area marked “+” in Fig. 7(b), and the results are shown in Fig. 7(d). The EDS results indicate the presence of C, Mg, Si, and Al. Here, we note that Al and Mg are detected because the A356 alloy material is not completely dissolved during etching and C mainly derives from SiCnps. It confirms that the nanoparticles are SiC. This indicates that the SiCnps are not homogeneously distributed according to particle size throughout the matrix, and that SiCnps with a much larger average particle size tend to reside at the interface between the eutectic Al and Si phase. This is an important finding that is discussed at much greater length in Section 4. Fig. 7(e) presents a high-magnification SEM micrograph of the edge of a eutectic Si, and the vicinity of the eutectic Si in the sample etched in short time. Here, we note that SiCnps are adhered to the eutectic Si, and the cross-sectional morphology of the eutectic Si becomes nearly equiaxed. In addition, some SiCnps are distributed in the vicinity of the eutectic Si. These results further suggest that SiCnps distributed at the surface of the eutectic Si block the local diffusion of the solute, and lead to the modification and refinement of the eutectic Si.

Fig. 7.

(a) and (c) are SEM images of the eutectic microstructures in 2wt% SiCnps/A356 samples deep-etched for 1h and 2h respectively. (b) Magnified from the area marked as the rectangle in (a) shows the morphology of the eutectic microstructures. (d) The EDS results of the area marked with “+” in (b). (e) The SEM image of the eutectic Si edge and the vicinity of the eutectic Si in 2wt% SiCnps/Al composites etched in short time.

(1.29MB).

Fig. 8 presents DSC solidification curves obtained at a cooling rate of 10°C/min for the A356 and 2.0wt% SiCnps/A356 samples. Here, the observed extrema for the A356 sample are marked A, B, and C and those for the 2.0wt% SiCnps/A356 sample are marked A1, B1, and C1. As can be seen from the figure, no obvious endothermic and exothermic extrema occurred in two DSC curves below 540°C. For the A356 sample, the onset temperature of the first exothermic peak A was 610°C, the onset temperature, peak temperature, and ending temperature of peak B were 569.9°C, 563°C, and 552.1°C, respectively, and the onset temperature, peak temperature, and ending temperature of peak C were 543°C, 540.5°C, and 538°C, respectively. For the 2.0wt% SiCnps/A356 sample, the first exothermic peak A1 occurred with an onset temperature of 607.2°C, which was followed by peaks B1 and C1, with peak temperatures of 555.6°C and 535.5°C, respectively. The onset and ending temperatures of peak B1 were 562.5°C and 549.4°C, respectively, while the onset and ending temperature of peak C1 was 540.4°C and 533°C, respectively.

Fig. 8.

DSC solidification curves obtained at a cooling rate of 10°C/min for the A356 alloy (with extrema marked A, B, and C) and the A356 alloy with the addition of 2.0wt% SiCnps (with extrema marked A1, B1, and C1). The extrema marked A and A1 represent the solidification of primary Al dendrites. The extrema marked B and B1 represent the solidification of the binary Al–Si eutectic at the α-Al grain boundaries. The extrema marked C and C1 represent the solidification of binary Al–Si eutectic droplets in the interior of α-Al grains.

(0.09MB).

The temperature ranges and peak features for the two samples given in Fig. 8 indicate that the extrema given by A and A1 correspond with the solidification of α-Al, while the extrema given by B, C and B1, C1 correspond with the solidification of the Al–Si eutectic. Here, the extrema given by B and B1 correspond with the solidification of the binary Al–Si eutectic at the α-Al grain boundaries, while the extrema given by C and C1 correspond with the solidification of Al–Si eutectic droplets in the interiors of the α-Al grains. A comparison of the DSC curves of the two samples indicate that the onset, peak, and ending temperatures of extremum B1 for the 2.0wt% SiCnps/A356 sample were less than those of extremum B for the A356 sample. In addition, the onset, peak, and ending temperatures of extremum C1 for the 2.0wt% SiCnps/A356 sample were also less than those of extremum C for the A356 sample. Undercooling (ΔT) is defined as the difference between the onset temperatures corresponding with the solidification of the Al–Si eutectic at the α-Al grain boundaries and the solidification of Al–Si eutectic droplets in the interiors of the α-Al grains. As such, ΔT represents the extent of undercooling required to nucleate the eutectic Si microstructure during solidification. Accordingly, we obtain ΔT values of 26.9°C for the A356 alloy and 22.1°C for the 2.0wt% SiCnps/A356 sample. As such, the value of ΔT for the SiCnps/A356 sample is considerably less than that of the A356 alloy. This indicates that the addition of SiCnps leads to a decreased extent of undercooling required to nucleate the eutectic Si microstructure in the 2.0wt% SiCnps/A356 sample.

HRTEM analyses were applied to further evaluate the effect of SiCnps on the microstructure of the eutectic Si. Fig. 9 shows a series of HRTEM images obtained from A356 with the addition of 2.0wt% SiCnps. Fig. 9(a) shows that SiCnps are distributed uniformly in the α-Al and eutectic area, and that the particle sizes of SiCnps distributed in the α-Al are smaller than those in the eutectic area. These results contrast sharply with average particle size of SiCnps at the eutectic Si/Al interface, which serves as another important finding that is discussed at much greater length in Section 4. As can be seen from Fig. 9(b), a significant number of multiple Si twin crystals are formed within the eutectic Si. And nanoparticles are distributed at the eutectic Si/Al interface and in the eutectic Si. This is further illustrated by the magnified image of the eutectic Si shown in Fig. 9(c), where a number of SiCnps and Si micro-twin crystals are observed. The high-density Si micro-twin crystals are parallel to each other and grow along the <112> growth direction of Si, which is marked with white arrows in Fig. 9(c). The SiCnps are distributed along the <112> growth direction of Si, which is marked by white dashed lines with the growth direction indicated by the arrow in Fig. 9(c). Furthermore, some SiCnps give rise to line defects within the eutectic Si. The role played by SiCnps in the eutectic Si can be further evaluated by the HRTEM image in Fig. 9(d) of the area marked by the white box in Fig. 9(c). Here, we note that the Si micro-twin crystals originate at, or very near, the interface between the nanoparticle and the eutectic Si. These micro-twins are parallel to each other and grow along the <112> growth direction of Si. The EDS analysis of SiCnps in the eutectic Si is shown in Fig. 9(e). Accordingly, we note the presence of only C and Si in the area. In combination with the HRTEM images, these results indicate that the C signal mainly derives from SiCnps in the eutectic Si.

Fig. 9.

HRTEM bright field images of the 2.0wt% SiCnps/A356 sample showing the distribution of SiCnps on the α-Al and eutectic area (a), multiple twinned Si crystals and SiCnps (b), and SiCnps within a micro-twinned Si crystal (c). (d) HRTEM image and (e) EDS analysis of the area marked by the white box in (c).

(0.8MB).
4Discussion

The results presented indicate that the addition of SiCnps has a vital effect on the nucleation and growth of the eutectic Si. However, to understand this effect, we must first discuss the factors affecting the distribution of SiCnps based on the interaction mechanism between SiCnps and the solidification front (SF) of the alloy.

The impact of Brownian motion on the distribution of particles becomes increasingly significant as the size of the particles approaches the nanoscale, and a critical size is eventually reached where nanoparticles are dominated by Brownian motion, which then maintains a uniform distribution of nanoparticles within a molten alloy. Therefore, determining the critical size of nanoparticles is essential for evaluating the distribution of SiCnps. Based on a consideration of both Stokes’ law and Brownian motion in the settling of nanoparticles in molten alloys, Schultz et al. [28] developed the following expression for the critical radius of nanoparticles (R*):

Here, C is a shape-dependent constant that is assumed typically to be C=1, ρp and ρf are the densities of the SiCnps and the molten alloy respectively, μ is the viscosity of the molten alloy, mfp is the mass of the molten alloy particles, vfp is the velocity of the molten alloy particles based on the temperature T and Boltzmann's constant kB (i.e., vfp=2kBT/mfp), and g is the constant of gravitational acceleration. These parameters for the SiCnps/A356 system are listed in Table 2, and the value of R* can be calculated accordingly from Eq. (1). For SiCnps in molten Al–Si alloy, R*=61.3nm, and, for SiCnps in molten Si, R*=49.5nm.

Table 2.

Parameters for the calculation of the critical nanoparticle radius based on Eq. (1)[28,31,32].

Parameter  Unit  Value 
μ  Pa1×10−3(Al); 5×10−4(Si) 
mfp  kg  4.48×10−26(Al); 4.68×10−26(Si) 
ρp  kg/m3  3220 
ρf  kg/m3  2700 (Al);2320 (Si) 
kB  J/K  1.38054×10−23 
g  kgm/s2  9.8 
T  1023 

Furthermore, nanoparticles are also subject to a drag force. The drag force can be determined from Stokes’ law adjusted for Brownian motion as follows [29]:

where v and R are the velocity and radius of the nanoparticle, respectively. The repulsive force exerted by the SF on nanoparticles can be expressed as follows [30]:
where A is the Hamaker constant and D is the distance between the nanoparticle and the SF. Here, defining the diameter of a molten metal atom as Ds, D adheres to the range 2DsD50nm. Assuming that Fd=Fvdw, the critical velocity below which nanoparticles are captured by the SF is obtained as follows.

However, because Brownian motion does not play a significant role in nanoparticle motion when the particle size is greater than R*, the critical velocity below which nanoparticles with a radius greater than R* are captured by the SF can be expressed as follows [20]:

where Δγ0 is the interfacial energy between the solid and liquid phase in the molten alloy, a0 is the atomic diameter of the matrix, and α is the thermal conductivity ratio between the nanoparticle and the solid-liquid interface. These parameters for the SiCnps/A356 system are listed in Table 3, and the values of vcr can be calculated accordingly from Eqs. (4) and (5).

Table 3.

Parameters for the calculation of the critical nanoparticle velocity based on Eqs. (4) and (5) [31,32].

Parameter  Unit  Value (Al)  Value (Si) 
A  −1.09×10−21  1.22×10−21 
Ds  nm  0.3  0.3 
D  nm 
a0  nm  0.286  0.117 
α    0.35  0.56 
Δγ0  J/m2  0.85  2.02 

Fig. 10(a) and (b) shows the dependence of vcr on R for SiCnps based on the parameters listed in Table 3 for α-Al and the eutectic Si, respectively. For R<R*, vcr increases with increasing R. However, for RR*, vcr decreases with increasing R. For the current work, the velocity of a dendrite (Vt) is determined to be 0.38mm/s [20], which is shown by the dashed line in Fig. 10(a). Accordingly, SiCnps with radii resulting in vcr values less than Vt tend to be trapped within the α-Al area by the SF. From the HRTEM results shown in Fig. 9, we know that the value of R for SiCnps in the α-Al about 10nm, which is consistent with this theoretical analysis. In contrast, SiCnps with larger radii resulting in vcr values greater than Vt tend to move ahead of the SF, and therefore tend to move into the eutectic area, owing to its lower solidification temperature, or become trapped at the α-Al/eutectic interface upon complete solidification. This is consistent with the observation in Fig. 9 that the average particle size of SiCnps in the eutectic area was greater than that in the α-Al. We must also consider the interface velocity during eutectic solidification, which can be calculated for unmodified hypoeutectic Al–Si alloys based on the extent of undercooling (ΔT) as follows [33]:

Fig. 10.

The critical velocity vs the radius of particles: (a) for α-Al; (b) for the eutectic Si.

(0.16MB).

According to the results in Fig. 8, the measured value of ΔT for the 2.0wt% SiCnps/A356 sample in this study was 4.8K, and Ve is accordingly 21.76μm/s. Wang [22] reported that the interactions between nanoparticles and the SF in Al–Si alloys are equally applicable in the case of eutectic growth. Therefore, Fig. 10(b) shows the dependence of vcr on R for the eutectic Si, where Ve is represented by the dashed line in the figure. The basic trends of the curve are similar to that shown in Fig. 10(a) for the α-Al. According to the inset in Fig. 10(b) showing the variation in vcr over small R, nanoparticles with a value of R less than about 25nm have a value of vcr that is less than or equal to Ve. Therefore, these nanoparticles will tend to be engulfed in the eutectic Si. From the HRTEM results shown in Fig. 9, we know that the value of R for SiCnps in the eutectic Si area ranges from 5nm to 20nm, which is consistent with this theoretical analysis. Finally, Fig. 10(a) and (b) indicate that nanoparticles with R values around 30–60nm tend to move ahead of both the SF in the α-Al phase and the eutectic Si phase, and therefore become trapped at the eutectic Si/Al interface after solidification, which is consistent with the results presented in Fig. 7.

According to the results given and the above discussion, we can analyze the process by which SiCnps bring about the modification and refinement of the eutectic Si microstructure by inhibiting the growth and stimulating the nucleation of the eutectic Si according to the illustration given in Fig. 11. The distribution of SiCnps in the illustration is closely related to the evolution of the morphology of the eutectic Si. Beginning with the fully distributed SiCnps/the eutectic Si system in Fig. 11(a), the effects of SiCnps on the nucleation and growth of the eutectic Si are accordingly discussed separately as follows.

Fig. 11.

Schematic diagram illustrating the process by which SiCnps bring about the modification and refinement of the eutectic Si microstructure by inhibiting the growth and stimulating the nucleation of the eutectic Si. Here, (a) represents the fully distributed system, (b) represents the initial nucleation of Si atoms around select SiCnps, and (c) describes the growth of the eutectic Si.

(0.21MB).
4.1Effect of SiC nanoparticles on the nucleation of the eutectic Si

Considering the reduced nucleation undercooling of the eutectic after the addition of SiCnps (Fig. 8) and the distribution of SiCnps in the eutectic Si (Fig. 9), SiCnps at the center of the eutectic Si are likely to act as nucleation sites, as shown in Fig. 11(b). According to the free growth model [23], the undercooling required for grain initiation can be given as follows:

where the solid–liquid interfacial energy is given as σls=0.352J/m2, the entropy of fusion per unit volume is given as ΔSv=7.3×106J/Km3[23], and d is the particle diameter. Therefore, ΔTfg=4.8K when d=40nm. This result is consistent with the value of ΔT measured by DSC in the present work. Furthermore, good crystallographic matching between nanoparticles and the eutectic Si is a necessary condition for nanoparticles serving as nucleation sites. The crystallographic matching at an Si/SiC interface of the 2.0wt% SiCnps/A356 sample is examined in Fig. 12(a). In addition, the selected area diffraction pattern of the interface is shown in the inset of Fig. 12(a), where the SiCnps and eutectic Si lattices are identified using white and red lines, respectively. The results demonstrate a cube-on-cube orientation relationship between the eutectic Si and SiCnps lattices, with an orientation relationship of (1 1¯ 1)Si[0 1 1]Si//(1 1¯ 1)SiC[2¯ 1 3]SiC. The lattice spacing mismatch between <011>Si on {111}Si and <213>SiC on {111}SiC is δ=10.4%. This small value of δ demonstrates that SiCnps have excellent potential for serving as heterogeneous nucleation sites for the eutectic Si during eutectic solidification. In addition, the absence of an interfacial product in Fig. 12(a) indicates that very good bonding occurs between SiCnps and the eutectic Si. Finally, Fig. 12(b) indicates that the lattice mismatch between SiCnps and the eutectic Si is alleviated by the generation of dislocations at the Si/SiCnps interface.

Fig. 12.

HRTEM image of the interface between Si and SiC (a), and its corresponding selected area diffraction pattern in the inset. (b) The processed HRTEM images for the Si/SiC interface, after processing by the fast Fourier transform (FFT) and the Fourier mask filtering technique.

(0.43MB).
4.2Effect of SiC nanoparticles on the growth of the eutectic Si

As discussed previously, relatively large SiCnps that move ahead of the SF tend to be distributed in either the eutectic Al or at the eutectic Si/Al interface during the solidification of the eutectic microstructure. When the SiCnps are distributed in the eutectic Al, the Gibbs free energy is given as:

where S is the surface area of the SiCnps and σSiC–Al is the interfacial energy between SiCnps and Al, which is given as 0.81J/m2[32]. When the SiCnps are distributed at the eutectic Si/Al interface, the Gibbs free energy is given as:
where σSiC–Al and σSiC–Si are the contact areas between the SiCnps and Al and between the SiCnps and Si, respectively, and σSiC–Si and σSi–Al are the interfacial energies between SiCnps and Si (i.e., 1.26J/m2[33]) and between Si and Al (i.e., 0.479J/m2[34]), respectively. Based on the above parameters, we note that the calculated value of ΔG2 is less than that of ΔG1, indicating that the system is more stable if the SiCnps are distributed at the eutectic Si/Al interface. As such, the relatively large SiCnps that move ahead of the SF would tend to distribute at the eutectic Si/Al interface, and thereby hinder the diffusion of the solute atoms and inhibit the growth of the eutectic Si, as illustrated in Fig. 11(c). The factor would then lead to a transformation of the microstructure of the eutectic Si from a flake-like to a near-equiaxed microstructure (Fig. 4) and the refinement of the eutectic Si (Fig. 5).

Accordingly, SiCnps lead to the modification and refinement of the eutectic Si by the processes of eutectic Si growth blockage and the nucleation of the eutectic Si on SiC nanoparticles. In addition, some SiCnps are engulfed by the SF and distributed in the eutectic Si, as shown in Fig. 11(c). The eutectic Si crystals are extended in the <1 1 2¯> direction only. However, SiCnps engulfed by the SF alter the stacking sequence of Si atoms during crystal growth, and a new stacking fault is formed with an equivalent <112¯> direction to restore the original stacking sequence. Hence, multiple Si micro-twin variants are observed, which are parallel to each other in the growth direction of the eutectic Si. This analysis is consistent with the restricted TPRE growth mechanism. These Si micro-twin variants may be related to the morphology of the eutectic Si shifting from flake-like to block-like, and finally to a fibrous microstructure.

5Conclusions

The present work demonstrated that the addition of 0.5–2.0wt% SiCnps can induce the modification and refinement of the eutectic Si in Al–Si alloys, and that the effect becomes increasingly significant with an increasing concentration of SiCnps. The results and the analysis presented verified that the modification and refinement of the eutectic Si is closely related to the interaction between the SiCnps and the SF. When the particle size of the SiCnps is greater than the critical size, the SiCnps tend to move ahead of the SF, while smaller SiCnps tend to be engulfed. The SiCnps moving ahead of the SF hinder the growth of the eutectic Si by the formation of an SiCnps layer coating its surface. In addition, the SiCnps can act as heterogeneous nucleation sites for the eutectic Si. Ultimately, the modification and refinement of the eutectic Si is caused collaboratively by a combination of the above two effects of SiCnps addition. Furthermore, the SiCnps engulfed by the SF cause the formation of Si crystal twins in the eutectic Si.

Data availability

The raw data required to reproduce these findings are available to download from https://doi.org/10.17632/nmj7d4j9d8.1.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References
[1]
J.G. Kaufman.
Handbook of materials selection.
John Wiley & Sons, (2002), pp. 116-117
[2]
M.M. Makhlouf, H.V. Guthy.
The aluminum–silicon eutectic reaction: mechanisms and crystallography.
J Light Met, 1 (2001), pp. 199-218
[3]
K. Nogita, M.S. David.
Eutectic solidification mode in sodium modified Al–7 mass% Si–3.5 mass% Cu–0.2 mass% Mg casting alloys.
Mater Trans, 42 (2001), pp. 1981-1986
[4]
J. Asensio-Lozano, B. Suarez-Peña.
Effect of the addition of refiners and/or modifiers on the microstructure of die cast Al–12Si alloys.
Scripta Mater, 54 (2006), pp. 943-947
[5]
J.H. Li, X.D. Wang.
Modification of eutectic Si in Al–Si alloys with Eu addition.
Acta Mater, 84 (2015), pp. 153-163
[6]
X. Liu, Y. Zhang.
Twin-controlled growth of eutectic Si in unmodified and Sr-modified Al–12.7%Si alloys investigated by SEM/EBSD.
Acta Mater, 97 (2015), pp. 338-347
[7]
C.H. Liu, J.H. Chen.
Multiple silicon nanotwins formed on the eutectic silicon particles in Al–Si alloys.
Scripta Mater, 64 (2011), pp. 339-342
[8]
A. Darlapudi, S.D. Mcdonald.
The influence of ternary alloying elements on the Al–Si eutectic microstructure and the Si morphology.
J Cryst Growth, 433 (2016), pp. 63-73
[9]
A.K. Dahle, K. Nogita.
Eutectic modification and microstructure development in Al–Si alloys.
Mater Sci Eng A, 413–414 (2005), pp. 243-248
[10]
J.H. Li, M.Z. Zarif.
Nucleation kinetics of entrained eutectic Si in Al–5Si alloys.
Acta Mater, 72 (2014), pp. 80-98
[11]
J.H. Li, M. Albu.
Solute adsorption and entrapment during eutectic Si growth in A–Si-based alloys.
Acta Mater, 83 (2015), pp. 187-202
[12]
J. Barrirero, J. Li.
Cluster formation at the Si/liquid interface in Sr and Na modified Al–Si alloys.
Scripta Mater, 117 (2016), pp. 16-19
[13]
J.H. Li, S. Suetsugu, Y. Tsunekawa, P. Schumacher.
Refinement of eutectic Si phase in Al–5Si alloys with Yb additions.
Metall Mater Trans A, 44 (2013), pp. 669-681
[14]
N. Fatahalla, M. Hafiz, M. Abdulkhalek.
Effect of microstructure on the mechanical properties and fracture of commercial hypoeutectic Al–Si alloy modified with Na, Sb and Sr.
J Mater Sci, 34 (1999), pp. 3555-3564
[15]
T.H. Ludwig, P.L. Schaffer, L. Arnberg.
Influence of some trace elements on solidification path and microstructure of Al–Si foundry alloys.
Metall Mater Trans A, 44 (2013), pp. 3783-3796
[16]
T.H. Ludwig, J. Li.
Refinement of eutectic Si in high purity Al–5Si alloys with combined Ca and P additions.
Metall Mater Trans A, 46 (2015), pp. 362-376
[17]
T.H. Ludwig, E.S. Dæhlen.
The effect of Ca and P interaction on the Al–Si eutectic in a hypoeutectic Al–Si alloy.
J Alloys Compd, 586 (2014), pp. 180-190
[18]
L.Y. Chen, J.Q. Xu, X.C. Li.
Controlling phase growth during solidification by nanoparticles.
Mater Res Lett, 3 (2015), pp. 43-49
[19]
L.Y. Chen, J.Q. Xu.
Rapid control of phase growth by nanoparticles.
Nat Commun, 5 (2014), pp. 3879
[20]
K. Wang, H.Y. Jiang.
Nanoparticle-inhibited growth of primary aluminum in Al–10Si alloys.
Acta Mater, 103 (2016), pp. 252-263
[21]
H. Choi, X. Li.
Refinement of primary Si and modification of eutectic Si for enhanced ductility of hypereutectic Al–20Si–4.5Cu alloy with addition of Al2O3 nanoparticles.
J Mater Sci, 47 (2012), pp. 3096-3102
[22]
K. Wang, H.Y. Jiang.
Microstructure and mechanical properties of hypoeutectic Al–Si composite reinforced with TiCN nanoparticles.
Mater Des, 95 (2016), pp. 545-554
[23]
K. Wang, H.Y. Jiang.
Nanoparticle-induced nucleation of eutectic silicon in hypoeutectic Al–Si alloy.
Mater Charact, 117 (2016), pp. 41-46
[24]
Y. Zhang, H. Zheng.
Enhanced nucleation of primary silicon in Al–20Si (wt%) alloy inoculated with Al–10Si–2Fe master alloy.
Mater Lett, 123 (2014), pp. 224-228
[25]
K. Hu, X. Ma.
Morphological transformation mechanism of eutectic Si phases in Al–Si alloys by nano-AlNp.
J Alloys Compd, (2018), pp. 765
[26]
D. Jiang, J. Yu.
Fabrication of Al2O3/SiC/Al hybrid nanocomposites through solidification process for improved mechanical properties.
Metals, 8 (2018), pp. 572
[27]
M.P.D. Cicco, L.S. Turng.
Nucleation catalysis in aluminum alloy A356 using nanoscale inoculants.
Metall Mater Trans A, 42 (2011), pp. 2323-2330
[28]
J.B. Ferguson, B.F. Schultz.
Impact of Brownian motion on the particle settling in molten metals.
Met Mater Int, 20 (2014), pp. 747-755
[29]
B.F. Schultz, J.B. Ferguson.
Microstructure and hardness of Al2O3, nanoparticle reinforced Al–Mg composites fabricated by reactive wetting and stir mixing.
Mater Sci Eng A, 530 (2011), pp. 87-97
[30]
I.B. Ozsoy, G. Li.
Shape effects on nanoparticle engulfment for metal matrix nanocomposites.
J Cryst Growth, 422 (2015), pp. 62-68
[31]
J.Q. Xu, L.Y. Chen.
Theoretical study and pathways for nanoparticle capture during solidification of metal melt.
J Phys Condens Matter, 24 (2012), pp. 255304
[32]
S.I. Chung, K. Izunome.
Estimation of surface tension of molten silicon using a dynamic hanging drop.
Jpn J Appl Phys, 34 (1995), pp. L631
[33]
S.D. Mcdonald, A.K. Dahle.
Eutectic grains in unmodified and strontium-modified hypoeutectic aluminum–silicon alloys.
Metall Mater Trans A, 35 (2004), pp. 1829-1837
[34]
E. Saiz, R.M. Cannon, A.P. Tomsia.
Acta Mater, 45 (1999), pp. 4209
Copyright © 2019. The Authors
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Tools
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.