Ferrite in 17-7PH

Delta Ferrite is very hard to see in this material unless you use a NaOH electrolytic etch. This etch will darken the Ferrite and leave the Martensite alone.

Without more information and a picture it is hard to make a definition as to the microstructure in you sample.

As this material is limited to use at 600F or below I have never heard of Sigma being a problem.

Thanks for rapid response, unclesid. How can I send you pics? We have tried numerous etches already with material in various conditions and cannot correlate microstructure, etch reactions and ferromagnetism readings to anything that makes consistent metallurgical sense.

I would appreciate a few words of wisdom on how the ferrite content grows (is it more time at carbide precipitation temp to produce more residual ferrite, or is ferromagnetism controlled (limited) by composition-and principally C content?

Look at Met-Tip #5 for the etching procedure to define the Ferrite.

It has been my experience that examining the microstructure of 17-7 can be a lesson in futility due to varying thermal and mechanical history.
You might contact AK Steel to see if you can get some of the old ARMCO Research Groups data. Also NASA and SPS all have done a lot of work on correlating the microstructure to physical properties. If you have access to Metadex there are several papers on this subject.
O

I have been able to look at two samples that had a slight difference in processing and correlate the differences in porperties to the micros. I don't feel that it is possible to look at single samples and determine the history.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

Ferrite forms in 17-7PH during solidification. It is delta ferrite. 17-7 is 100% ferrite at solidification and then most of the ferrite transforms to austenite. 5 to 20% remains as ferrite and you are stuck with it. You cannot transform it to austenite and therefore not to martensite.
Retained ferrite is the Achilles heel of all PH grades. It could transform to alpha prime or sigma, but it's bad enough as is. It's generally strung out and very detrimental to through thickness toughness.


Michael McGuire

Michael:
Thanks for the reminder that 17-7 starts life as a ferritic material. Your comments are well-taken and agreed.
The phenomenon I am struggling with is ferromagnetism in the alloy, and how to control it. Despite the residual delta ferrite after a 1950 solution anneal the material is not magnetic. But after 1450/1.5h +1050/3h it is. I initially overlooked the fact that there is 1% Al in the alloy. So now it is most probable that at 1450 we are forming Fe-Ni-Al intermetallics and so making the alloy more ferritic by removing Ni from SS. Then, because of this change the material transforms to martensite on cooling, and so becomes ferromagnetic. The 1050 treatment then tempers the martensite by precipitating carbides,-moving the alloy still further towards a ferritic character.
But what are the key factors for control of ferromagnetism? Is it Ni+Al content and the time at 1450-1750? Or is it the tempering step at 1050? Or is it carbon content-or a combination of all three of these? Where is the Curie temperature for this alloy?
Example: Tests with the Biddle meter (10=no magnetism; 1=High magnetism) give 3-5 for one heat and 1-2 for another with marginal chemistry differences and the same heat treatment.
Chris Bagnall

The main driver in ferromagnetism is having full transformation to austenite and then having full transformation to martensite.
Having enough carbon and a full solution anneal produces the absence of magnetism you observe after 1950 solution anneal.
Then the carbon is gain important because it de-stabilizes the austenite by precipitating as Cr23C6 during the 1450 treatment. This raises the Ms so you can have full transformation to martensite. If you have some retained austenite after the 1450 treatment, it can transform to ferrite during the aging, depending on chemistry.
17-7PH is in my opinion the most complex stainless steel because of the all the interplay among elements doing different things and the non-equilibrium phases.

Michael McGuire

In the annealed condition the carbon helps the alloy be less magnetic since it is an autenite stabalizer.

Now you have me curious. I'll start checking magnetism on 17-7.

I think to try and develop a correlation you would need to heat treat many samples from the same heat and compare them. After you had that trend identified you could try the same with other heats.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

Ed/Michael:

The whole situation with this alloy is a metallurgical maze. The fact that it is a very close cousin to 304SS is what really gets me going. We have alredy started a few lab tests. So far:
1. The delta ferrite is not magnetic at any stage, and takes no part in any of the reactions at 1450, 1750, 60, or 1050F. After 1950F and air cool, there is no martensite and there is no ferro-magnetism
2. More time at 1050, or time at 1000F has the same result: Magnetism decreases. So much for my thought that carbide formation would result in more magnetic ferrite in the structure
3. We do not believe carbides form at 1450-1750. We think this is a crucial step involving formation of intermetallics, of which Ni3Al is probably the most important. This step would then change the composition enough (removal of Ni from SS) to allow the martensitic transformation to occur on cooling. We are trying variations of time at 1450F to see what influence this has on the air-cooled microstructure, which we expect to be martensite and magnetic.

Your continued comments/ideas would be welcome! And if you have any data to help take the mystery out of this strange alloy, please let me know.

You will have to bear with me as I'm using my memory instead of my notes/notebook.

Early on in the development of the PH alloys my big boss was a fanatic about using these alloys for everything, I might say with varying results. The URL is a
lead to some work for NASA concerning the transformation to Martensite in 17/7. We did somewhat parallel work on several of the PH alloys. This was the result of loosing a large number of very expensive parts fabricated from a PH alloy due to SCC.



We also found that PH alloys tended to further age with extended time/temperature cycles. Our process cycle was preheat to 600F then an extend run at 550F then to RT a short excursion to 900F then RT and then to preheat again 600F. All the PH steel we used exhibited this trait with 17/4 showing the most pronounced effect.

We also found that most of the in service PH alloys didn't respond well to a solution anneal and precipitation heat treatment. It was our experience the most pronounced effect was seen in 17/7.

The reason we stay close to the problem is that the majority all our PH components were pressure containing.

As stated before correlation of physical properties with microstructure is problematic. We manufactured a small diameter pen in house for use in a process.

You will have to bear with me as I'm using my memory instead of my notes/notebook.

Early on in the development of the PH alloys my big boss was a fanatic about using these alloys for everything, I might say with varying results. The URL is a lead to some work for NASA concerning the transformation to Martensite in 17/7. We did somewhat parallel work on several of the PH alloys. This was the result of loosing a large number of very expensive parts fabricated from a PH alloy due to SCC.

Look at the section on heat treating.

We also found that PH alloys tended to further age with extended time/temperature cycles. Our process cycle was preheat to 600F then an extend run at 550F then to RT a short excursion to 900F then RT and then to preheat again 600F. All the PH steel we used exhibited this trait with 17/4 showing the most pronounced effect.

We also found that most of the in service PH alloys didn't respond well to a solution anneal and precipitation heat treatment. It was our experience the most pronounced effect was seen in 17/7.

The reason we stay close to the problem is that the majority all our PH components were pressure containing.

As stated before correlation between the microstructure and physical properties is very problematic. We manufactured a small pin from different PH materials that had to be somewhat precise, in 0.0001" range and we were never a able to sort or measure these pins electrically whereas we could sort or measure tool steel pins to millionths with no problem. Photomicrographs of two pins that measured appreciably different showed no discernible difference in the microstructure.
On some larger components we tried to measure crack depth with a improved version of the commercially available crack depth meter with the results being totally unreliable. Examination of the microstructure revealed no discernible difference.

The solidification ferrite that remains in the structure must be ferromagnetic. The martensite that forms on aging must also be ferromagnetic.
Exactly how magnetic though will depend on composition. Both the permiability and sturation magnetization will vary, so measurements at a fixed field level may not tell you anything.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

Agreed-although I was not sure about the delta ferrite for a while!
The 1950 delta ferrite is magnetic;-a polished section reacts to a magnetized needle hung above it. But the 10% volume fraction of this phase is not enough to have any effect on a Biddle meter, when tested on the OD of a tube.
Our problem in a nutshell is the response of two 0.035" wall tubes made from the same heat, drawn down from the same 1"dia x 0.045 stock. With the TH heat treatment (in which we found that the 60F soak for elimintating retained austenite is of major importance) we can produce a Biddle No. of 1 (very magnetic)after the 1450F +60F segment on 3/4" finish size.(After our earlier tests,we do not believe the 1050F stage does much at all for the magnetic properties). However, the 5/8" finish size only produces a Biddle No. of 4+.
Why? Cold work effects on solution HT? Solution HT time/temp? What else could be going on here?
Do you have any info on efects of longer times (in our case greater than 5 minutes) at 1950 and above? I recall that 304SS is solution heat teated at 2200F. Why not take 17/7 up to the same temp to be sure we really have everythng in solution.

What level of magnetic field does the Biddle use? It may be too low of a field to measure the ferrite effectivly.

We stay away from high temp for a couple of reasons. Not hte least of which is grain growth. The others are surface degradation and the formation/growth of primary MC carbides.

We do some 17-7 work with multiple draws. I'll try to watch the properties with further reductions.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

The Biddle meter is simply a pen-shaped device. The magnetic "point" is attached to a spring so that after contacting your test piece you pull against the spring and read off the spring extension against a scale from 1-10. Not sure what the feld strength of the magnet is. One (1) is high, 10 is no detectable magnetism)

We also us a Severn Low Mu Permeability Indicator for low-end ferite measurements; this is more sensitive but more complex to use. A weighted pendulum, with the bearing at its mid-point, has a magnetic hammer at the top. A range of Mu standards can be screwed into one side of the device opposite one end of the hammer. Your "unknown" is then placed on the opposite side, and the hammer swings if the field of the unknown is greater than the selected mu standard. A 1.0mu standard is approx 0.5% ferrite; 1.5mu = 5%. We cannot use this meter for the higher mag readings on 17/7, but to give you a rough idea, 8 on the Biddle gage is in the range 3-5mu on the Severn gage (10-15% ferrite equivalent).

I have inserts for my Severn that range from 1.01 to 6. but the problem is that these are both measurements of initial permiability taken at a low applied field strength.
I guess no magnetic measurements on this grade.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

Ed:
Confused by your last; can you explain?

We have the same inserts, but as I indicated the meter does not help with the 1450F and below HT material. It's only good for telling you about vry small amounts of magnetism (1900F condition. So Biddle picks up when Severn reaches its limit. And conversely, the Biddle is no good for the very low susceptibility samples.

I am very interested in your offer to measure effects of sequential reductions on 17/7 tubing (I assume). However, these data will be all in the solution annealed condition, I expect. Any chance of cutting a few inches off after each pass so we could look at the effect of a TH treatment on the different draw sizes?

Well, it goes like this. I know that there are phases in 17-7 that are ferromagnetic. But I don't know what the magnetization curves look like. That is the permiability as a function of applied field strangth. I do know that these curves will change based on both composition and thermomechanical history.
So, I don't see a straight forward way to use magnetic characteristics to tell me anything about the structure/condition of the alloy.
This may be doable with a Feritscope. I believe that you can alter the field strength on them. This might get over the variations in initial permiability.

Yes, we make welded and drawn tubing. I'll start collecting more samples in process and see what I can find.

= = = = = = = = = = = = = = = = = = = =
Rust never sleeps
Neither should your protection

DocCB,

Can you elaborate on the importance of the 60F soak you mentioned in one the threads.

"Our problem in a nutshell is the response of two 0.035" wall tubes made from the same heat, drawn down from the same 1"dia x 0.045 stock. With the TH heat treatment (in which we found that the 60F soak for elimintating retained austenite is of major importance)..."

I understand this is part of the HT for TH. AMS 2759/3D: "...cool to below 60F within 1 hour, hold below 60 F for not less than 30 minutes".

I'm interested in using magnetic measurement to find high hardness (45 or greater HRC)17-7PH TH1050 parts (I've got 50,000 parts and don't want to do hardness testing on all of them). I've had two system wrecks related to this hardness condition. Do you think this is possible?