Low yield, tensile and hardness in A356 2

[trytrue]

Dear colleagues:
We make aluminum parts by low pressure permanent mold, LPPM. Some times after heat treat the A356 castings, have low hardness, low yield and/or low tensile strength.
We sent samples to two Universities in order to determine in which step of the heat treat (solution, quench, or age) we have the failure but the results were not clear.
We sent two samples a bad one (soft and low mechanical properties) and another good one (hard and good mechanical properties).
One report told us that the good one has precipitates in the matrix and the bad one does not have. Then they concluded that the precipitates are most likely to be Mg2Al4Si5O18 cordierite. They use TEM and work only with the matrix. This statement does not agree with the technical information’s that establish that the precipitates only are Mg2Si.
The second reported that the bad one has the Mg2Si small and well distributed in the matrix and the good one has large Mg2Si distributed in the matrix and in the eutectic. Then they conclude that large Mg2Si is due to an incomplete solution that left Mg2Si in the eutectic and the Mg2Si agglomerated and grew during the age. They used a special etching technique to reveal the Mg2Si precipitates in the matrix. The report has SEM microphotographs showing white spots visible at 1600X for both samples. The technical information told that the Mg2Si precipitates are very small around 200 nanometers, this size needs TEM with higher magnification than SEM.

Our specifications are:
Solution 9 hours at 1005 F + /- 10 F. Quench water temperature 150 F +/- 10 F; quench delay 15 seconds max. Age 3.5 hours at 320 F +/- 10 F.
Mechanical properties 217 MPa yield, 281 MPa tensile and 7% elongation. All these values are minimum values. Minimum hardness 85 BHN.

Are there labs or University that can identified the Mg2Si precipitates in the matrix? The matrix has a 30 microns width.
Can some body tell me how to etch in order to glow white the Mg2Si precipitates?
Is there another way to identified where part (solution, quench or age) of the heat treat is failing when we have low hardness, low yield, and or low tensile?

Any questions or comments I will be more then happy to know.

Thank you for your time

 

[TVP]

I'm going to tackle your problem from a couple of different perspectives. First, I would like to cover some of the basics regarding part configuration, heat treating process, etc. Then I will address the university lab results.

First, what type of part is this? Nominal wall thickness? What type of basic microstructural characterization do you typically perform? I am asking these questions because I am wondering if sufficient analysis has been performed to rule out a basic problem with grain size. Grain size has a significant affect on mechanical properties, and this is something you should absolutely rule out before looking into TEM analysis of precipitates, etc.

Next, I have some questions about heat treating practice. Have you ever measured metal temperature as opposed to furnace settings? What type of spacing do you maintain between parts during the heat treating process? How often is quench temperature monitored? Are the parts agitated during the quench process?

All of this leads into my analysis of your heat treating parameters. All of the parameters you listed fall within the nominal ranges outlined by MIL-H-6088 (now SAE-AMS-H-6088) and the various ASM reference books (ASM HANDBOOK Volume 4 Heat Treating, etc.). The quench delay is the maximum allowed, so have you investigated the effect of reducing this? The solutionizing time is on the low side, especially if you have thick parts, complex geometry, and the parts are jammed together with no space among them. MIL-H-6088 allows for up to 24 hours solution time. The aging time is somewhat low, with 6 hours being a maximum in the MIL spec. Also, the aging temperature should probably be more like 310 F +/- 10 F, as 300-320 F is most appropriate range.

All of this leads me to the following recommendation:

Take the parts that have low mechanical properties/hardness, and re-heat-treat them.

1. Solution at 1005-1025 F for 24 hours. Quench into water at 150 F, with less than 15 second delay (as fast as possible). Artificially age at 300-320 F for 6 hours.

2. Characterize the parts for a.) mechanical properties, b.) hardness, and c.) microstructure (grain size, precipitate distribution, etc.)

Compare the results of this with i.) previously identified good parts and ii.) previously identified bad parts. These parts should exhibit the best homogeneity, mechanical properties, etc. I realize this does not address specifically your concern about which part of the process is likely to be at fault. However, this would provide you with the best baseline for subsequent comparisons.

Ok, so now lets talk about the university results. I am inclined to believe the lab that said your good parts had a more uniform distribution of Mg₂Si, and that the bad ones were due to an incomplete solution. My reasoning is as follows:

1. Mg₂Si should be small, and uniformly distributed throughtout the matrix, not agglomerated at the grain boundaries with the eutectic. Mg₂Si segregates upon solidification to the grain boundary along with the eutectic. Proper solutionizing dissolves this phase, and allows for a smaller, and more finely distributed structure of the precipitates after the aging process.
2. Mg₂Si does agglomerate and grow if not dissolved during the solution phase.
3. Cordierite is not a phase that is recognized as a strengthening precipitate. It may form due to reaction with oxygen during the melting process, but it certainly is not the reason why your "good" parts demonstrate superior strength, hardness, ec. It is more likely some Al₂O₃ that was still present was complexed with the Mg₂Si, not that a distinct cordierite phase was present. X-Ray Diffraction would be necessary to conclusively decide on which phase was actually there.

I am still investigating the part about revealing Mg₂Si with a special etchant. Mg₂Si is usually identified without the aid of an etchant on a uniformly polished sample at low magnification (~ 250x) due to it's distinct blue color. Use of a caustic etchant (1 g NaOH with 100 mL water, swab for 5-10 s) may enhance the blue color. Also, viewing with polarized light can aid in visual determination of Mg₂Si.

However, this is for the larger size Mg₂Si that forms is not subsequently dissolved during the solution treatment. 200 nanometers is the right size for strengthening precipitates, and would not be easily resolved with light or electron microscopy.

Anyway, I would like to hear some more about your process, and let me know if you have any questions. As I mentioned previously, I will get back to you regarding the etching procedure.

 

[TVP]

I found a reference for color etching Al-Si casting alloys, but I am not sure of the exact effect on the Mg₂Si phase. The generic name is Weck's reagent for aluminum, and the composition is as follows:

4 grams KMnO₄; 1 gram NaOH; 100 mL distilled water, slightly warmed.

Etch by swabbing.

 

[trytrue]

TVP,
Thank you for your answers.

The plant cast and heat treat suspension components, they have between 0.5 to 1.5 inches wall thickness.
The tensile test bars are cut from rib castings of 1 inch thickness (high stress location)

The typical microstructure is round silicon in the eutectic, very small Al5FeSi needles in the eutectic and a dendritic aluminum matrix; DAS 30 microns in the outer surface and 40 microns in the center. The grain size is 5AFS (10 ASTM). We use Sr to modified the silicon and Ti to refine the grain. In our LPPM process, the typical microstructure has not significant variation from load to load.

As I told you the low hardness, yield and/or tensile strength happens from time to time not very often.

Every month each furnace is surveyed with 20 embedded thermocouples and the entire probes need to reach the temperature lower limit in 3 hours, then in the remaining time or the dwell time, 6 hours, all the probes need to stay within the temperature limits.

The parts are placing in racks, they are very close each other. All the time we have in each racks the same amount of castings.

The quench tank has thermocouples, and a control that do not allow to quench casting out of the water temperature. The quench tank has four agitator.

The quench delay usually is under 15 seconds if some load has more or equal to 15 sec the load is place on hold to take additional samples (tensile, hardness, microstructure etc.) or reheat treat.

ASTM B917 for LPPM heat treat states: 1000 F +/- 10 F, 4 to 12 hours; age 310 F, 2 to 5 hours. We increase a bit these temperatures in order to reduce time.

In the automotive aluminum business most of the companies are using less a less time in solution and age in order to do that they are increasing the temperature. For example, we are using for another smaller castings 330 F +/- 10 F as aging temperature with very good result, but that process is gravity PM. It will be expensive to heat treat for 24 hours.

You are right in increasing our age time, we will use 6 hours. This will be our next experiment

About Mg2Si point 1 agree, 2 disagree, 3 agree.

The larger Mg2Si is 1000 nanometers still very difficult to see with optical or SEM (according with experts in Electron beam analysis and Optical analysis).

I will try the Weck’s reagent. Thank you

Again thanks a lot for your comments and opinions,

 

[TVP]

trytrue,

Thanks for the additional information. It is obvious that you have a well-controlled heat-treating process, which does not make it very easy to diagnose problems such as the one that you are having. As I mentioned in my previous email, the MIL spec allows for up to 24 hours for the solution treatment phase, but it is quite clear that automotive production schedules and volumes do not allow for this much time.

I wanted to add some comments regarding Mg₂Si. The following is taken from Table 5, page 356 of ASM HANDBOOK Volume 9Metallography and Microstructures:

Mg₂Si external shape: cubic habit; eutectic forms script that easily coalesces on heating.

Also, pages 374-375 have a number of figures showing the microstructure of various aluminum casting alloys, including 356. Mg₂Si is easily identified in several of the optical micrographs, taken with low magnfication and only a 0.% HF etch. Figure 132 on page 375 shows a 500x micrograph of A357 that has not been properly solution treated and artificially aged. Mg₂Si is present as an undissolved phase along the grain boundaries with the eutectic silicon.

Another good source of information on aluminum featuring details on physical metallurgy and microstructure is Aluminum: Properties and Physical Metallurgy by J. E. Hatch. It was published in 1984 by ASM, and can be obtained from ASM at

http://www.asminternational.org/

.

 

[trytrue]

tvp,

Thanks for your reply.

We have installed several controls in the heat treat process but we still have problems and no answers.

About the Mg2Si, are these micros (ASM HANDBOOK Volume 9Metallography and Microstructures)
in as cast condition or after heat treat?
In the AFS book Solidification Characteristics of Aluminum Alloys, they show Mg2Si in the eutectic as black particles at 500X in as cast samples. I tried to find out the same size and color in samples after heat treat but without success that is why we sent samples to labs with SEM and TEM.

 

[TVP]

trytrue,

There are micrographs showing Mg₂Si both in as-cast samples, and after heat treating. They appear as black particles in these photos because they are black and white images (no color).

By the way, if you didn't see my question in the other thread on 356 castings can you please take a look at it and get back to me? Thanks. A link to the thread is included below:

thread330-35850