Taking Mg-4La-2Al-0.3Mn (LA42) alloy as research object, OM, XCT, numerical simulation and other methods were used to compare conventional die-casting process of AZ91D to optimize die-casting process suitable for LA42 alloy filter housing. Study shows that optimal die-casting process for filter housing is a pouring temperature of 720 ℃, a mold temperature of 250 ℃, and a boost pressure of 90 MPa. Microstructure and performance of back and heat dissipation teeth of filter housing formed under this process are quite different. Cooling rate of back of housing is slow, grain size is large, and there is a large amount of pre-crystallization structure, while cooling rate of heat dissipation teeth is fast, grain size is small, there are cold shut and hole defects. Thermal conductivity of heat sink teeth [107.7 W/(m·K)] is lower than that of back of shell [112.3 W/(m·K)]. Yield strength of heat sink teeth (170.1 MPa) is much higher than that of back of shell (138.4 MPa). However, heat sink area contains obvious cold shut and hole defects, which significantly affect elongation.
Traditional die-cast magnesium alloy AZ91D has excellent casting properties (fluidity, thermal cracking resistance, etc.) and is widely used in the field of electronic communications. However, with improvement of equipment integration, volume is reduced, and power consumption of equipment increases. AZ91D alloy contains a large number of solid solution atoms, which causes scattering during electron transportation, resulting in a thermal conductivity of only 51 W/(m·K), which cannot meet heat dissipation requirements of equipment. To address this problem, author team has developed a high thermal conductivity magnesium alloy LA42 (Mg-4La-2Al-0.3Mn, wt.%) suitable for die casting, with a thermal conductivity of up to 110 W/(m·K), a die casting test tensile rod yield strength of 140 MPa, a tensile strength of 240 MPa, and an elongation of 12%. However, in actual application, most products have complex structures, large sizes, uneven wall thickness, and thin walls, which require more stringent die casting filling capabilities of alloy. Effect of die casting filling process ultimately affects quality of die casting.
As core component of radio frequency of communication base station, filter needs to solve heat dissipation problem of filter housing in order to ensure its efficient operation. There is an urgent need for filter housing materials with excellent thermal conductivity. As shown in Figure 1, size of filter housing is as high as 456 mm*301 mm*76 mm, and the thinnest part of heat dissipation teeth is only 1 mm. Casting performance of LA42 alloy is slightly worse than that of AZ91D alloy. Conventional die casting process cannot produce high-quality parts for filter housing. In this work, die-casting process suitable for large thin-walled LA42 alloy filter housings was obtained through numerical simulation optimization of die-casting process, microstructure and properties of casting body were analyzed to explore feasibility of promoting application of LA42 magnesium alloy in filter housings.

 

Figure 1 Die-casting of filter housing

Test material is LA42 alloy, and commercial AZ91D alloy is used for control test. Chemical composition of LA42 and AZ91D alloys was determined by ICP-OES, and results are shown in Table 1. Filter heat dissipation housing was die-casted by a die-casting machine Bulher SC140L, with a pressure chamber length of 760 mm and a cooling time of 15 s. Die-casting process parameters were optimized and selected through numerical simulation of die-casting process on Zhizhu Super Cloud Platform (Supreium®). Specific die-casting process is shown in Table 2. Finally, XCT was used to compare and evaluate the porosity, hot cracks and defect distribution of LA42 and AZ91D die-castings to evaluate forming ability of LA42 alloy.

Table 1 Chemical composition of LA42 and AZ91D alloys wB/%

Table 2 Die-casting process parameters used in this experiment
A 12.7 mm diameter round piece was cut from die-casting body, a 15 mm * 3.5 mm plate-shaped tensile piece was cut from gauge section, corresponding samples were cut from heat dissipation teeth and back of shell. Specific sampling positions are shown in Figure 2. Room temperature thermal diffusion coefficient α was measured using an LFA467 laser thermal conductivity meter, and room temperature density ρ of sample was tested using a densitometer (Sartorius Quintix124-1CN). Specific heat capacity was calculated based on Neumann-Kopp law.

 

Where: Cp,i (T) is constant pressure specific heat capacity of each element in alloy, and xi is mass fraction of element in alloy.

 

Figure 2 Sampling position of thermal conductivity disc sample and tensile sample
Specific heat capacity of each element used in this paper varies with temperature as follows:

 

Thermal conductivity of final sample can be obtained by formula (6):
λ=αρCp (6)
Where: α is thermal diffusion coefficient, ρ is density, and Cp is specific heat capacity.
Above tensile sheet was subjected to room temperature tensile test at a tensile rate of 0.5 mm/min on an MTS tensile testing machine. After mechanical grinding, polishing and etching with a special etchant for magnesium alloy (4 mL nitric acid and 96 mL ethanol) for 5 s, microstructure was observed using a Zeiss optical microscope AXIO SCOPE 5. Backscattered (BSE) photos were taken using a scanning electron microscope equipped with EDS, and solid solution atomic concentration was measured in at least ten matrix areas using EDS point scanning mode.

Based on thermophysical property data of material, pressure chamber filling mode is used to calculate die casting filling process to assist die casting process design. Figure 3 shows numerical simulation results of LA42 alloy under different die casting process parameters at the same slow injection filling time (2.88 s). By observing melt filling in pressure chamber, it is found that process A is same as AZ91D process, with a pouring temperature of 680 ℃ and a mold temperature of 180 ℃. Under this condition, LA42 alloy is filled smoothly, but because solidification range of LA42 alloy is about 630~580 ℃, which is much higher than solidification range of AZ91D (595~437 ℃), temperature of most melts close to inner wall of pressure chamber is lower than liquidus temperature, α-Mg has begun to nucleate and solidify (position indicated by arrow of process A in Figure 3), viscosity increases, fluidity decreases, and it is easy to produce pre-crystallization structure, which causes insufficient shrinkage compensation in dendrite gap to produce shrinkage and significantly affects plasticity.

 

Figure 3 Numerical simulation results of LA42 alloy under different die-casting process parameters at the same slow injection filling time (2.88 s)
Therefore, it is necessary to increase pouring temperature and mold temperature to improve filling and shrinkage compensation capabilities of LA42 alloy. Processes B and C increase pouring temperature to 720 ℃ and mold temperature to 250 ℃ on the basis of process A. Melt temperature distribution in pressure chamber section is greatly improved, all above liquidus temperature. Process D further increases pouring temperature to 760 ℃ on the basis of process C, melt is overheated, and splashing occurs in slow injection stage (as shown in red box in Figure 3D). Therefore, pouring temperature should be controlled at 720 ℃.
As shown in Figure 4, numerical simulation results of entrainment pressure distribution under different die-casting process parameters are shown. It can be seen from figure that entrainment pressure of process B has not been significantly improved compared with process A, and is distributed in same area (as shown in red box in Figure 4B). This is because this position is in the area where two melts meet and is the thinnest part of casting wall. When melts meet here and front of interface cannot be exhausted, this area is prone to air entrainment and pores. Flow tracking results in Figure 5 also verify above conclusions. Flow intersection and air entrainment are easy to occur in black box. Therefore, on the basis of process B, it is necessary to further increase boost pressure to 90 MPa (process C). It is found that air entrainment pressure is reduced, and further increase of pouring temperature (process D) increases air entrainment pressure. This is because temperature is too high, which is easy to react with hydrogen to produce bubbles that are entrained into melt. According to above results, process C with a pouring temperature of 720 ℃ and a boost pressure of 90 MPa is the best choice.

 

Figure 4 Numerical simulation results of air entrainment pressure under different die-casting processes

 

Figure 5 Numerical simulation results of flow tracking of process B
As shown in Figure 6, numerical simulation results of shrinkage and shrinkage cavities under different die-casting processes are shown. It can be seen from figure that from process A to process B, tendency of shrinkage and shrinkage cavities is significantly reduced by increasing pouring temperature, then from process B to process C, boost pressure is increased, tendency of shrinkage and shrinkage cavities is further reduced. From process C to process D, pouring temperature is further increased to 760 ℃, melt is overheated, and when solidified to same volume at different temperatures, the greater solidification shrinkage, the easier it is to produce shrinkage cavities and shrinkage. Therefore, in process D, shrinkage tendency is increased. Numerical simulation results of shrinkage cavities verify rationality of process C.

 

Figure 6 Numerical simulation results of shrinkage cavities under different die-casting processes

Filter heat dissipation housing was trial-produced using process C. Die-casting samples of LA42 alloy and AZ91D alloy are shown in Figure 7. By comparison, it is found that surface quality of AZ91D alloy sample is excellent, thin-walled heat dissipation teeth near gate and far from gate are well filled. LA42 alloy sample is basically filled completely, but flow marks are visible on the surface, and there are unfilled cold shut defects in some parts. This is attributed to high viscosity of LA42 alloy, which makes it easy to stick to mold.

 

Figure 7 Surface quality photos of LA42 and AZ91D alloy filter housing castings
Internal quality of castings was further explored by XCT, and results are shown in Figure 8. XCT scanning resolution is 136 μm. At this resolution, porosity of LA42 alloy casting is 0.03%, and porosity of AZ91D alloy casting is 0.02%. The two are basically same, and location of holes is basically consistent with numerical simulation results of process C. Hole defects of die castings of both alloys are easily distributed at intersection of liquid flows (red box in Figure 8) and wall thickness area at far end of ingate (blue box in Figure 8). In addition to a small number of hole defects, no obvious hot cracks were found inside the LA42 alloy casting. This is because even though fluidity of LA42 alloy is worse than that of AZ91D, its solidification range is narrow and eutectic reaction temperature is high. Final solidification area can be fed at a higher temperature, so tendency to hot cracking is low. It shows that except for sticking phenomenon on the surface of LA42 die casting, under this process, there are no large number of holes larger than 100 μm inside die casting, LA42 alloy is suitable for filling large and complex thin-walled parts. High vacuum die casting can be further used to reduce internal pores, and mold oil circuit can be modified to increase surface temperature of heat dissipation tooth mold, thereby improving casting quality.

 

Figure 8 Three-dimensional defect distribution results of LA42 and AZ91D magnesium alloy filter housing castings obtained by XCT

Figures 9a and 9b are metallographic structures of back of LA42 alloy housing near mold surface and center of cavity, respectively. Figures 9c and 9d are metallographic structures of heat dissipation teeth of LA42 alloy housing near mold surface and center of cavity, respectively. It can be seen from figure that compared with Figures 9c and 9d, Figures 9a and 9b show coarser grains and contain a large amount of pre-crystallized structures. There are no obvious defects on the back of shell near mold surface (Figure 9a), while heat sink teeth contain a few cold shut cracks near mold surface (Figure 9c), which indicates that cooling rate of heat sink teeth is faster than that of back of shell, nucleation rate is high, grains are fine, and multiple strands of molten metal have solidified before heat sink teeth can fuse together. There are also a large number of small hole defects near center of cavity of heat sink teeth (Figure 9d), while almost no hole defects are observed near center of cavity on the back of shell (Figure 9b). As shown in Figure 10, numerical simulation results of solidification temperature field of LA42 filter shell obtained by process C verify above view.

 

Figure 9 Metallographic structure of LA42 alloy filter shell

 

Figure 10 Numerical simulation results of solidification temperature field of LA42 alloy filter shell
Table 3 shows thermal conductivity and mechanical properties of each area of LA42 filter shell, and Figure 11 shows corresponding tensile curve. By comparison, it is found that thermal conductivity of heat sink tooth area is lower than that of back of shell. Figure 12 and Table 4 are backscattering photos of heat sink tooth area and back of shell area and comparison of solid solution atom concentration. Concentration of dissolved atoms in heat dissipation tooth area is higher than that in the back of shell. This is attributed to faster cooling rate in heat dissipation tooth area. Dissolved atoms do not have time to diffuse and are dissolved in matrix. Dissolved atoms disrupt periodic arrangement of Mg lattice, which is main factor affecting thermal conductivity. At the same time, high solid solubility and finer grain size in heat dissipation tooth area contribute to higher solid solution strengthening and grain boundary strengthening effects, so yield strength is also significantly higher than that in the back of shell. However, elongation and tensile strength in heat dissipation tooth area are significantly lower than those in the back of shell, which is attributed to cold shut and hole defects in heat dissipation tooth area, which also corresponds to organizational photos of Figures 9c and 9d. This also reflects that compared with pre-crystallized structure, effects of cold shut and hole defects on elongation are more prominent.

Table 3 Thermal conductivity and mechanical properties of different parts of LA42 alloy filter housing

 

Figure 11 Tensile curves of LA42 alloy samples in different parts of filter housing

Table 4 Concentration of solid solution atoms in different regions of filter housing

 

Figure 12 Backscattering photos of different parts of LA42 alloy filter housing

(1) Numerical simulation results show that optimal die casting process for LA42 alloy filter housing is a pouring temperature of 720 ℃, a mold temperature of 250 ℃, and a boost pressure of 90 MPa. This process can successfully form filter housing castings.
(2) Microstructure and properties of back and heat dissipation tooth area of LA42 alloy filter housing are quite different. Due to high cooling rate, fine grains, and a large amount of solid solution atoms in heat dissipation tooth area, thermal conductivity is low [107.7 W/(m·K)] and yield strength is high (170.1 MPa), but elongation is significantly reduced due to cold shut and hole defects. In contrast, back of shell has high thermal conductivity [112.3 W/(m·K)] and low yield strength (138.4 MPa).