Using HIP to Improve Aluminum Castings
With more design opportunities arising every year for aluminum casting, commercial demand for engineered cast components continues to increase. One large market niche falls with high-integrity aluminum castings for aerospace and automotive markets.
However, a major limitation of aluminum castings for these applications is the potential presence of porosity as the melt solidifies and contracts. Internal shrinkage porosity is a common fatigue life-limiting factor, and this can be true for very small amounts of porosity in high-quality castings and for more prevalent internal voids in parts of lower quality castings.
One solution to eliminate internal porosity is hot isostatic pressing (HIP), which bonds materials together in cast components. The process applies high pressure to the exterior of a casting or metal powder billet via a high temperature inert gas. Sand, permanent mold and investment casting are the most common casting methods that respond well to HIP.
HIP can eliminate internal porosity in castings, leading to improved fatigue life. Equipment and processing advancements in HIP have made it more attractive to both aerospace and automotive industries as an economical means to improve casting quality and reduce scrap rates. Further, a HIP process has been designed specifically for aluminum alloys to maximize cost effectiveness.
Investigations were conducted to evaluate how HIP methods (both aluminum and non-aluminum specific) can help improve the overall internal structure of aluminum castings.
This article details how HIP affects the fatigue life of different types of both aerospace and automotive aluminum alloy castings.
Testing the Market
The first investigation into HIP involved an aerospace OEM that manufactures helicopters. This firm sought to compare the fatigue properties of as-cast components to those that had undergone the HIP process. Investment cast A356 and sand cast A357 aluminum test plates (measuring 1.25 x 8 x 11 in. [3 x 20 x 28 cm]) underwent T6 heat treatment and were analyzed. The OEM decided to compare data of both aluminum-specific and traditional HIP test plates to the as-cast ones. In the test matrix, three or four plates were used for each condition and all plates underwent the final T6 heat treatment step together. All as-cast components underwent x-ray inspection with all but one receiving a Grade A rating.
The other investigation involved an automotive OEM, which took a different approach than the aerospace firm. The automotive firm selected several A356 aluminum steering knuckles produced via sand and permanent mold casting processes. These castings failed x-ray inspection; however, the firm wanted to determine if the aluminum-specific HIP process could salvage any of them. Test bars were extracted from each casting, and all as-cast and post-HIP parts underwent x-ray inspection. The HIP plates also underwent T6 heat treatment before fatigue testing (Table 1). Overall casting dimensions were 9 x 11 in. (23 x 28 cm) for sand and 11 x 17 in. (28 x 43 cm) for permanent mold, with varying wall thicknesses.
Table. 1. Heat Treatment Parameters for the Aerospace and Automotive Castings
|Al alloy casting type||Al-specific HIP||Traditional HIP||T6|
|Aerospace--A356, A357||Proprietary||961F/102 Mpa/2 hrs.||Hold at 900F for 1 hr., solutionize at 1,009F for 19 hrs., water quench, 320F age for 9 hrs.|
|Automotive--A356||Proprietary||N/A||Solutionize at 1,000F for 8 hrs., 120-140F water quench, 325F age for 4 hrs.|
The initial findings for the aerospace firm revealed that no measurable changes were discovered between the as-cast and post-HIP conditions via ultrasonic testing. For the automotive OEM, the x-ray inspection of the permanent mold castings revealed that HIP processing reduced x-ray defect indications to one-eighth of their original volume, with 20 of the 25 steering knuckles having no indications at all after HIP. All of the test bars then underwent fatigue testing (Table 2).
Table 2. Fatigue Testing Parameters for the Aerospace and Automotive Castings
|Al Alloy casting type||Gage diameter (mm)||Gage length (mm)||Test Frequency (Hz)||R-Ratio||Test Waveform|
|Aerospace--A356, A357||6.4||19.1||30||Variable||Sinusoidal, load control|
|Automotive--A356||8||16||50||-1||Sinusoidal, load control|
Finding the Source
A fractography study was conducted to determine the origin of fatigue failure. After being randomly renumbered to hide their prior processing path from the investigator, both fracture surfaces of each broken fatigue test bar were initially scanned under a stereomicroscope. In nearly all cases, the quarter- to half-moon shape of the fatigue (slow crack growth) zone was visible to the naked eye. If the fracture cause could not be identified at low-power magnification (up to 40x), then one of the two fracture surfaces was readied for insertion into a scanning microscope (SEM).
Fatigue crack initiation defect results were matched against the processing conditions (Table 3). Four different types of fatigue crack origins were found: pores, oxides, non-oxide inclusions and slip planes. The oxides and non-oxide inclusions were identified by their characteristic x-rays. Porosity was determined to be the predominant fatigue life-limiting factor in the as-cast components, regardless of quality. HIP eliminated porosity in the other castings, and oxides were then the most common fatigue crack initiation defect, followed by a few instances of non-oxide inclusions and microstructural features. However, oxide size, frequency and characteristic x-ray strength varied widely. In addition, once a fatigue crack is initiated, it follows the path of least resistance, so the fracture surfaces were likely depicting an abnormally high concentration of oxides. It also should be noted that three of the as-cast A356 aluminum aerospace investment castings failed at oxides rather than pores. This indicates that oxides can overcome porosity as the fatigue life-limiting factor for aluminum alloy castings depending on the relative defect sizes.
Table 3. Frequency of Fatigue Crack Initiation Defect Types in the Test Bars
|Casting Category||Casting Type||Alloy||Condition Prior to T6||Test Bars w/ Pores at FIS||Test Bars w/ Oxides at FIS||Test Bars w/ Non-Oxide Inclusions at FIS||Test Bars w/ Slip Planes at FIS||Number of Unbroken Test Bars||Total Number of Test Bars|
|Automotive||Permanent Mold||A356||Al-specific HIP||2||12||0||0||4||18|
The bars that did not break before the number of fatigue cycles (30 million for aerospace and 10 million for automotive) were tested with lower values of maximum stress (i.e. 124.1 MPa and below for the A356 aluminum).
Two of the automotive permanent mold castings that had undergone the HIP process were found to fail at pores. After viewing the samples in a SEM, it was discovered that the fatigue fracture surfaces indicated massive interconnected porosity across their faces. The investigators concluded that surface-connected porosity existed in the castings from which these fatigue test bars were machined.
The two types of SEM electron detectors played critical roles in identifying fatigue fracture initiation defects. The fracture initiation origin was normally located using secondary electron mode. Switching to the contrast mechanism in backscattered electron mode usually pinpointed the origin defect, then an energy dispersive spectroscopy (EDS) spectrum identified what elements were present in that defect.
After the fatigue fracture initiation defects were correlated with the samples’ histories, the cycles to failure versus maximum applied stress values could be compared for the various aluminum casting sources, alloys and processing condition. Because the aerospace and automotive firms utilized different fatigue testing methods, their fatigue properties could not be directly compared. In addition, the sample lot size for each condition of the aerospace castings was too small to incorporate statistical analysis.
In 50% of the automotive castings, the pores that initiated fatigue failure were visible by low-power stereomicroscope, so no further evaluation was required. That percentage was cut in half with the high-quality aerospace castings.
Once the analyses of each casting group were finalized, it was found that higher-quality aerospace casting test plates appeared to improve slightly more from traditional HIP versus the aluminum-specific HIP process. The apparent added benefit from HIP must be weighed against the cost savings of the aluminum casting-specific process. Both the A356 and A357 aluminum aerospace castings appeared to have decreased fatigue life variation after HIP, but the improvement was most dramatic in the A357 sand castings. Additional test plates currently are undergoing fatigue testing, which will provide more data for statistical evaluation.
For the automotive castings, all but one indicated significant improvement after HIP to a 95% confidence level (T-test). The test bars here were tested at a maximum stress level of 131 MPa, which included one premature failure due to probable surface-connected porosity. However, this analysis combined the data from both the permanent mold and sand castings, and the permanent mold castings exhibited a great degree of scatter in fatigue life. Comparing the two automotive S-N curves indicated that the sand casting process produced steering knuckles with less scatter in the fatigue property data than those cast in permanent molds, and in turn, the sand castings obtained a greater improvement from the aluminum-specific HIP process. However, even with the range of fatigue lifetimes, the permanent mold castings showed an 88% decrease in x-ray defect indications after HIP. It should be noted that all of these automotive components were x-ray rejects before this investigation.
This article was adapted from a paper presented at the AFS International Conference on High Integrity Light Metal Castings, Oct. 31-Nov. 1, 2005, Indianapolis.