Evaluation Of Two New Cast Aluminum Alloys For High Performance Cylinder Heads
Cast aluminum alloys have been increasingly used in internal combustion engine cylinder head applications because of their light weight and high thermal conductivity. Increasing demand for fuel economy and high-power density has posed a significant challenge on existing cast aluminum alloys for high temperature performance. A recent study evaluated two new high temperature cast aluminum alloys Al-Q (AlSiCuMg) and ACMZ (AlCuMnZr) using semi-permanent mold (SPM) cast cylinder heads. While the new cast aluminum alloys, especially the ACMZ alloy, show a significant improvement in high temperature tensile properties, they exhibit similar fatigue performance as the commonly used cylinder head aluminum alloy A356+0.5%Cu.
In today’s cylinder head designs, the most commonly used cast aluminum alloys are A356, 319 and AS7GU (A356+0.5%Cu).1 The A356 alloy is a primary aluminum alloy with good ductility and fatigue properties at low to intermediate temperatures (RT-150C/302F). However, above approximately 200C (392F), creep resistance and tensile strength of this alloy rapidly degrade due to the fast coarsening of Mg/Si precipitates in the alloy.
The 319 alloy is a secondary aluminum alloy representing a lower cost alternative to A356. The copper-bearing 319 alloy has the advantage of better tensile and creep strength at intermediate temperatures (150-200C/302-392F) because the Al/Cu precipitates are more stable (higher resistance to coarsening) at higher temperatures than the Mg/Si precipitates in A356. However, this alloy is prone to shrinkage porosity due to the high Fe and Cu content and low ductility at room temperature.
The AS7GU alloy is a variant of A356, strengthened with 0.5%Cu. Like A356, the AS7GU alloy has good castability while the small Cu addition improves creep resistance and tensile strength at intermediate temperatures. Both Mg/Si and Al/Cu precipitates are thermally unstable, thus all three alloys have poor mechanical properties above 250C (482F) (482F) due to rapid coarsening of these precipitates. As reported by Engler-Pinto et al., the AS7GU-T64 alloy at room temperature, is clearly superior to “W” and “E” 319-T7 alloys. However, at 250C (482F), all the alloys evaluated show equivalent and significantly reduced fatigue properties compared to the room temperature data. This indicates that the beneficial effects of precipitation hardening on fatigue resistance completely disappears in the typical operating temperature range desired for engine efficiency.
With increasing demand for fuel economy and power density, peak engine pressures have increased significantly, Figure 1. Over the past 20 years, peak cylinder pressure has increased from about 40 bars in 1999 to near 150 bars, while the maximum operating temperature of cylinder heads has also increased from approximately 170C (338F) to temperatures near 250C (482F). Therefore, high temperature properties including tensile and fatigue strengths of the cast aluminum alloys become critical.
Over 200 cylinder heads were made with the Al-Q and ACMZ alloys using the semi-permanent mold casting process. Table 1 shows the chemical composition ranges of the Al-Q and the ACMZ alloys. The AS7GU (A356+0.5%Cu, baseline alloy) is included for comparison.
Prior to casting, the melt was degassed using argon with a rotary degasser to achieve a hydrogen level below 0.15ml/100g Al. For the Al-Q alloy, the melt was also modified using Al-10% Sr to obtain a final 0.01% Sr level. For the ACMZ alloy, the melt was grain refined using a TiBor (5:1) alloy with an addition of 200ppm B. After casting, the cylinder heads made of the Al-Q alloy were subjected to a T7 heat treatment consisting of a solution treatment (five hours @ 530C/986F), quench into warm water, and aging at 200C (392F) for five hours. The ACMZ heads were subject to a solution treatment at 540C (10004F) for five hours and aging at 240C(464F) for five hours to maximize Cu dissolution for over-aged precipitation hardening.
After heat treatment, samples were taken from the cylinder heads in both deck face and head bolt boss for mechanical property evaluation. Mechanical testing included both tensile and fatigue. Tensile properties were measured at room temperature (25C/77F), 150C/302F, 200C/392F, 250C/482F, and 300C/572F. Fatigue testing was carried out under fully reversed uniaxial loading (R=-1) at room temperature (25C/77F), 150C/302F, 200C/392F, and 250C/482F. In the case of testing at elevated temperatures above 150C (302F), the specimens were isothermally preconditioned at the specific testing temperature for 100 hours prior to testing. The reported tensile data in this study is the average property of at least five specimens. The fatigue strengths at both low cycle fatigue (10^4 cycles) and high cycle fatigue (10^7 cycles) were obtained from staircase fatigue testing of at least 30 samples for each data point.
Fractographic analysis was conducted under scanning electron microscopy (JEOL JSM-IT300 SEM) to identify and quantify the fatigue crack initiators for every fatigue specimen. The defects that initiated the fatigue cracks were statistically analyzed and compared.
Figure 2 compares the average tensile properties between baseline (A356+0.5%Cu) alloy and Al-Q alloy in two locations taken from the cast cylinder heads. At room temperature, the Al-Q alloy provides better strengths than A356+0.5%Cu alloy. The increase in intermetallics by the addition of additional copper and the minor elements in the Al-Q alloy did result in slightly lower ductility compared to the baseline alloy. At 300C (572F), the Al-Q alloy, strengthened with Q-phase, also showed better strength in comparison with the A356+0.5%Cu. With regards to both the total elongation and ultimate tensile strength (UTS) at 300C (572F), any significant yielding of the material will result in engine distortion and would be unacceptable. As the greatest stresses and temperatures are in the bridge areas between valves, head distortion could lead to valve seat distortion resulting in poor valve seating, loss of compression, lower power, and higher emissions.
Although the minor elements, zirconium and vanadium have stabilized the Q-phase and produced an increase in tensile strength at 300C (572F) for the Al-Q alloy, property differences between the two alloys at the midrange temperatures (200-250C/392-482F) are insignificant. The steep drop in strength at intermediate temperatures for both alloys indicate that the Q-phase may not necessarily be a more effective high temperature strengthener than other precipitates like 0-phase or ß-phase. This has been confirmed with an excellent transmission electron microscopy (TEM) study (Fig. 3). Loss of orientation relationship was observed at 300C (572F), with highly misoriented Q-phase visible in both alloys. At lower temperatures (<250C/482F), the Q-phase rods are orientated mainly along certain crystal orientations as shown as two main orientations in Figures 3a-3d. Additionally, the Q-phase rods are observed to coarsen significantly in both alloys at 300C (572F). The loss of orientation of the Q-phase rods, in addition to the coarsening of the precipitates, is the likely cause for the drastic drop in strength at higher temperatures.
Table 2 compares the measurements of precipitate length and radius of both alloys at temperatures between 200-300C (572F). The Al-Q alloy is more coarsening resistant at all the temperatures, which explains its superior performance. The extent of misoriented Q-phase precipitates was the lowest in the case of the Al-Q alloy.
Tables 3 and 4 compare the yield strength (YS) and ultimate tensile strength (UTS) of the ACMZ, Al-Q, and baseline A356+0.5%Cu alloys. At temperatures below 200°C, the yield strength of the ACMZ alloy is lower than that of the Al-Q alloy and baseline A356+0.5%Cu alloy. At temperatures above 200C (392F); however, the ACMZ alloy is much better than the other two alloys. The yield strength of ACMZ alloy at 300C (572F) is more than doubled in comparison with the YS of the other two alloys.
The UTS of the ACMZ alloy is comparable and even slightly better than that of the Al-Q and Al-356+0.5%Cu alloys at temperatures below 200C/392F (Table 4). At temperatures above 200C (392F), the ACMZ alloy is much better than the other two alloys. Like YS, the UTS of the ACMZ alloy at 300C (572F) has more than doubled compared with the other two alloys.
The enhanced tensile strengths of the ACMZ alloy particularly at high temperatures are attributed to the limited coarsening of Ø’-AlxCu precipitates, which is affected by both thermodynamic and kinetic factors. As shown in Figure 4, the segregation of Mn and Zr elements at the interfaces between Ø’-AlxCu precipitates and the aluminum matrix reduces the interfacial energy, which makes the precipitate thermodynamically more stable. Furthermore, the presence of Mn and Zr at the precipitate interfaces act as diffusion barriers reducing the coarsening rate of Ø’-AlxCu precipitates at elevated temperatures. Figure 5 shows a TEM image of Ø’-AlxCu precipitates at 300C (572F). It is seen that the Ø’-AlxCu precipitates at 300C (572F) still maintain high aspect ratios, which explains why the alloy shows high tensile strengths at elevated temperatures.
Tables 5 and 6 show the low-cycle fatigue and high-cycle fatigue strengths of the Al-Q and A356+0.5%Cu alloys at various temperatures and two locations in the cylinder heads. Compared to the baseline alloy, the Al-Q alloy does not show clear superiority in fatigue strength.
Additionally, Table 6 compares the high-cycle fatigue strengths of the ACMZ and Al-Q alloys with the baseline A356+0.5Cu alloy at various temperatures and two locations on the cylinder heads. It appears that on the deck faces, the ACMZ alloy is slightly better than the other two alloys. In the head bolt bosses, however, the ACMZ alloy does not show clear superiority over the A356+05%Cu and Al-Q alloys.
It is generally accepted that fatigue strength is controlled mainly by defect size while tensile properties are more related to volume fraction of defects. In aluminum castings, the size of defects depends more upon melt quality, hydrogen level, solidification rate, and other casting process variables than upon alloy composition.
Fractographic analysis and quantification of crack initiators of the fatigue samples was conducted under SEM. In general, the fatigue failure was caused by defects (porosity) in the samples (Figs. 6-7).
Looking at the Al-Q alloy, the high-cycle fatigue strengths of the alloy at various temperatures in the deck face area seem to be slightly higher than those of the baseline A356+0.5%Cu alloy, but the results are not consistent. This can be explained from the measured sizes of defects (porosity and oxides) that initiate fatigue cracks in the fatigue failure samples. As shown in Figure 8, the average defect size (B50) in the baseline A356+0.5%Cu alloy is 137μm, which is slightly smaller than 152μm for the Al-Q alloy. However, the 3-Sigma pore size (B99) of the A356+0.5%Cu alloy is 1233 μm, which is significantly larger than 494μm for the Al-Q alloy.
In the head bolt boss areas, the fatigue strengths are also quite comparable between the baseline and Al-Q alloys, although the Al-Q alloy seems more durable than the baseline alloy at 150C and 200C (302 and 392F). As expected, the average sizes (B50) of defects that initiated fatigue cracks are also very comparable (Figure 9). Similarly, the scatter of defect sizes in the baseline alloy is significantly larger than that in the Al-Q alloy. As a result, the 3-Sigma pore size (B99) of the baseline alloy is 1306μm, while the Al-Q alloy is only 456 μm.
For the ACMZ alloy, the fatigue strengths in the deck face area appear better than those of the other two alloys. However, the ACMZ alloy does not show any superiority in the head bolt boss area. This can also be explained with the sizes of defects that initiated the fatigue cracks in the material (Fig. 10). In the deck face area, the average defect size (B50) of the ACMZ alloy is 95μm, which is smaller than that of both the Al-Q (152μm) and baseline A356+0.5%Cu (137μm) alloys. However, in the head bolt boss areas, the average defect sizes (B50) are very comparable. The average defect sizes (B50) are 212μm, 221μm, and 216μm for the ACMZ, Al-Q, and baseline A356+0.5%Cu alloys, respectively. The 3-Sigma defect size (B99, 424μm) of the ACMZ alloy is similar to that of the Al-Q alloy (457μm), but it is much smaller than the 3-Sigma defect size of the baseline alloy (1306μm).
High tensile properties of an alloy do not guarantee high fatigue performance, as fatigue properties are very sensitive to defect sizes in the castings. The ACMZ alloy has excellent tensile strengths particularly at elevated temperatures; however, its fatigue properties are very similar to the other two head alloys (baseline alloy A356+0.5%Cu and Al-Q alloy). To achieve higher fatigue performance for the new higher strength head alloys, casting defects should be minimized.
Main Points Discovered
The microstructure and mechanical properties of two new cast aluminum alloys have been evaluated using cylinder head castings. The following conclusions can be drawn from this work:
1. In general, the castability of the Al-Q and ACMZ alloys is good. Sound cylinder head castings were made using the same tools as for the baseline A356+0.5%Cu (AS7GU) alloy without macro shrinkage defects. For the ACMZ alloy, a high content of grain refinement agent was added to avoid hot cracking.
2. Tensile strengths of the two new alloys, particularly the ACMZ alloy, are remarkably superior to the baseline alloy at elevated temperatures.
3. Fatigue performance of the two new alloys is quite similar to the baseline alloy as it is controlled by defect sizes and population in the cylinder head castings. The defect sizes are rather comparable in the heads made from the three alloys.