Hydrogen Embrittlement 1

[9two5]

We are having problems with 4350 quenched and tempered parts failing an endurance test. The failures are linked to hydrogen damage at the site of large inclusions, app. one eight of an inch below the surface. The mill that produced the material has claimed that we induced the hydrogen into the material during the heat treatment. We heat the parts for two and one half hours at 1550F in an endothermic atmosphere, quench in oil and double temper at 375F for two hours. I have never seen this problem before. I was under the impression that only atomic hydrogen can do this type of damage, and that much higher temperatures are needed to produce atomic hydrogen. Can hydrogen be diffused into steel at 1550F. Thanks in advance for any help.

 

[unclesyd]

There is a lot information in your post that is troubling.

What are the inclusions and do you find them before heat treating?
What specification are you call out for the material?
Is the material hot or cold rolled.
Is the part thick enough to require 2 hrs at temperature?
What is the composition of the end gas?
Why such a low tempering temperature?
Are you using a “quench” oil?
Do you have an idea of the dew point of the endo gas?
Where did the 1550̊ F come from? This seems a little high.

I have never had a problem with H2 during heat treating even went using N2 plus 2% hydrogen.
You can produce H2 at these temperatures with water and iron but the rate is very slow.

Sorry for so many questions.

 

[9two5]

Sorry for the lack of information. The inclusions are aluminum oxide, and we found them when the part broke during the endurance test.The material spec is 4350 aircraft quality. The material is hot rolled, turned ground and polished. The part diameter is two and one half inches. The gas is endothermic, 40% Nitrogen, 40% Hydrogen, and 20% Carbon Monoxide. The low tempering is to meet a 54-57 Rc requirement.We Quench in oil, Beakon K9. We run the endo at 30 dew point. We use 1550 to get the part in the hardest condition possible before tempering, we have found by doing this the cycle life is improved.

 

[Metalguy]

More questions. If AQ steel, why the "large" inclusions? Did previous non-failing parts have the same inclusions?

How sure are you that H is involved? There are ways to measure the amount of H in metal, and if you have a reasonable temp. history you can even back-date the H amount-since it slowly leaving the steel even at room temp-assuming it's not being recharged by being wet!

 

[unclesyd]

If the inclusions are Aluminum Oxide the fault lies with the producer of the bars. There is no way this is H2 embrittlement. This is what you would see in an Al killed steel on the head and tail of a bar. These sections are normally trimmed.

There is a possibility that there wasn’t enough metal removed due to the hot rolling prior to the TGP process. Depending on the producers process up to 1/4" material (diameter) from raw blank may have to be removed.

Again I question the ?H2 damage point of view? Probably if you section a bar prior to heat treating you will probably find incipient cracking originating at an inclusion site. Do the same after heat treating. You will have to go up to high magnification (750x) if using optics. It will be quite evident on SEM.

The endo is a little high (about 10%)in H2 from what is Ive seen used. The temper at 375°F X2 is the older standard for removal of H2 from parts after plating, now its 24 hrs @ 375°F. 4 hrs would get 90-95% of the H2 if there was any.

Are you comparing your fracture surface against a good one?
I would MT the bar as received to look for grinding cracks.

 

[TVP]

I don't agree completely with unclesyd's last comments. First, aluminum-killed steels will have some amount of aluminum-based inclusions (aluminates, globular oxides, etc.) as part of the melting and refining process. It is the responsibility of the specifier to set limits on their presence in the final product. It is the responsibility of the producer to use a suitable process for achieving the specified levels.

Second, Aircraft Quality is not a real specification. I am not familiar with any SAE AMS (Aerospace Material Specification) specifications for 4350, so can you provide more details on exactly how the material is specified/ordered? Typically, Aircraft Quality grades have very low inclusion levels, due to the use of "clean steel" practices in the ladle and possibly other techniques (tundish slags that attract aluminum oxides, vacuum degassing, etc.).

The last point is the identification of hydrogen damage. It is true that inclusions can trap hydrogen. The endo gas environment with 40% hydrogen can be embrittling for high carbon martensitic steels. What was the indication for hydrogen damage? Intergranular fracture near an aluminum oxide inclusion? Are you sure that it wasn't just a fatigue failure, with the origin near the inclusion?

 

[mcguire]

I see no basis to exclude the possibility that the failures were due to HE. The heat treat certainly guaranteed its presence. Both fatigue and HE are step-wise zero ductility failures so fractography will be of little value.
If the part failure can be duplicated under static load, then you can conclude HE was causal. Or degas thoroughly and run the fatigue test and see if the normal endurance is then seen.

 

[TVP]

mcguire,

I am not sure I understand your statement about fractography having little value. If the failure mode is by fatigue, then striations, beach marks, etc. are possible/likely to be revealed by fractography. If hydrogen embrittlement is the failure mode, then zero ductility should be manifested as intergranular fracture (rock candy appearance). Are you operating under the assumption that the fracture area is likely to be obscured by post-fracture damage?

 

[mcguire]

TVP
No, my point is that HE procedes discontinously in steps as does fatigue, but of course for different reasons. This would make it difficult to distinguish one from the other, especially is a very hard material.
HE could, I suppose, produce a rock candy appearance, but classic HE is very like fatigue with fracture coming in steps paced by hydrogen diffusion to the zone of triaxial tension ahead of the crack tip.

 

[unclesyd]

Fractography is highly important in analyzing any failure as far as I’m concerned. Like “TVP” there are a lot of indicators for each of “9two5's’ possible failure modes. Hydrogen induced cracking leading to fatigue failure has some very distinguishing features from the aforementioned rock candy progressing to cleavage fracture and microvoid coalescence. Granted this sometimes takes a little SEM work. As stated there should be evidence of striations near the origin of he failure toward the center of the bar unless they are multiple origins.
"9two5" do you have access to an SEM?

Which of the impurities in the endo-gas inhibit the diffusion of hydrogen? Nearly all furnaces that use a gas generator have appreciable amounts of hydrogen as component of the gas.

I’ll agree with “TVP” that “Aircraft Quality” isn’t a very good specification for material. But all steel centers normally sell such a material by advertising it’s quality.
As to the Aluminum oxide inclusions, I’ve never seen any amount and or sizeable aluminum oxide inclusions which would be considered a detrimental defect in a properly made bar no matter the process. Every thing else, yes, including radioactive cobalt.

There are a lot of avenues to pursue in this failure.
Are the all the inclusions near the surface or they scattered through the bar?
Have you been able to determine the origin of the crack?
Are the bars out of the same lot?
At what time in the test are bars starting to fail?

 

[CoryPad]

The debate started by TVP and mcguire regarding fracture surface appearance stems from the two different types of hydrogen damage: 1) high fugacity hydrogen which produces transgranular cracking (possibly striated similar to cyclic mechanical stressing); and 2) low fugacity hydrogen, which produces intergranular cracking and the distinctive "rock candy" appearance.

unclesyd,

What do you mean when you wrote "4 hrs would get 90-95% of the H2 if there was any."? If you mean that the [atomic] hydrogen would be removed, then I disagree, as there is much literature that shows a variety of temporary and permanent hydrogen traps in steel. The de-embrittlement baking process can remove some hydrogen, but the main benefit is to promote diffusion throughout the entire piece so that there is no local hydrogen content that exceeds the critical concentration that initiates cracking.

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[unclesyd]

CoryPad,

I probably used the wrong terminology as we always said bake out in referring to treating the parts. I did see some papers on work where they did measure H evolution from metals. I have no idea of their experimental apparatus.

The old procedure for H removal for a part either after plating or acid cleaning was a 4 hr bake at 375°F to get the H to an acceptable level. Speaking mainly of bolting. It was known all along that there was not complete removal of H but there was an acceptable level and this time (4 Hrs) accomplished it. At the time we considered this adequate for the time as it was unusual to stress a fastener above 45,000# , but as higher strength fasteners and higher preloads started being use some problems started being noted. We able to avail our selves of some work by the Navy where they were contemplating extending the bake out time to 24 Hrs at 375°F. We had several applications were we used the additional time and it apparently helped, no failures. We had one area where the only solution was to a lower strength bolt only more of them, in other words a grade five or grade eight wouldn’t work.

Our group was acutely aware of the problems with H as we had a process that had H at 6750 psig and over 400°F. Any failures were examined for H damage if a failure occurred. All our components for another process were tool steels with H-11 fasteners that had to be chemically cleaned. The cleaning procedures were carefully monitored and if there were any suspected problems the parts were baked out.

In the early 70's we did a lot of basic research work on an electrochemical process which involved both electroplating and operation. We used the 4 Hrs at 375°F as a standard for our plated products to eliminate H. We stayed with the 4 Hr bake out for several years until the electro- chemists said that we weren’t gaining anything.
It was at this time we finally got time on the SEM to do a little follow up on some previous fasteners failures and failures from the high pressure system.

FYI, I was also involved in a another type of hydrogen event that is seldom seen. We had a failure of PSA vessel in H purification by fatigue. After an extended investigation it was determined that his failure was “Hydrogen Assisted Fatigue” up until that time had only seen in the laboratory in England.

 

[9two5]

Just wanted to say thanks to everyone who responded. Our goal was to prove to the mill that our process did not induce the hydrogen into the part. We hardened some parts in nitrogen, never exposing them to endo, and they failed at the site of the inclusions, in the same fasion as the parts we hardened with endo. To answer a few questions, the failures were fatique failures which originated at the site of the inclusions. There were typical beach marks and chevron pattern pointing to the defect. The material had a (fish eye) appearance the site of the inclusion. There were both intergranular, and transgranular fractures surrounding the inclusion. Thanks again.

 

[unclesyd]

Next time you use N2 please add a little Methanol to the mix.

 

[TVP]

Thanks for the update 9two5! It's always good to receive some feedback on these types of things. Obviously fatigue failures can be caused by excessively high levels of nonmetallic inclusions. Aluminum oxide is especially detrimental due to its angualr/blocky shape and extreme hardness. There are a number of items to address with the steel producer:

1. First, you need to specify the maximum inclusion levels allowed. There are a number of standards that cover this including ASTM E 45, ASTM E 1122, ASTM E 1245, and DIN 50602.

2. Better attention to clean steel practices while using aluminum deoxidation. Vacuum degassing may be needed. Consumable electrode melting is often required in SAE AMS standards in order to meet the highest quality requirements.

3. Discuss with the steel producer using a vanadium addition instead of aluminum. This eliminates the aluminum oxide inclusions. Silicates and sulfides can still be present, so make sure to adequately specify them as well.