1084 grain size trouble

[Tony Mont]

I recently had some unexpected grain size and would like some help because I cannot figure this out.

The test knife in question had this heat treatment protocol (it was a forged blade):

1) 1600F-870C for 15 minutes and air cool
2) 1450F-787C for 15 minutes and air cool
3) 1380F-748C for 30 minutes and let it cool overnight in the oven
4) Aus temp 1500F-815C for 10 minutes and quenched in parks 50 oil
5) 2x2hours temper at 400F-200C

This is the grain structure I ended up with.


This morning, I made 5 coupons and began testing different protocols. I will explain my thought process on each test. All the coupons were marked with the corresponding number to avoid any mix ups. All pictures were taken using the macro setting on my phone.



I usually put my knives in the oven while it heats up, that's why I decided the first two coupons would have the same treatment but one of them heated up along with the kiln and the second one was put inside after the kiln was up to temperature. I did this to check whether the kiln had a hot spot while it heats up or not.

On the left it's coupon #1 and on the right it's my test knife. The grain is similar in both which made me curious, I've never had steel to have such a bad grain structure from the "as received" status.



From left to right: coupon #1, knife from the test and coupon #2. This coupon was again heated to 1500F-815C from the as received condition but put in the kiln once it was up to temp. Even though it looks better in the picture, it's the same exact grain structure as the first. When the coupons were freshly broken, they had a lighter colour which makes it look like a finer grain. A few minutes later when all of them had the same darker colour, their true nature revealed.


Coupon #2 on the left and coupon #3 on the right. Coupon #3 was cycled 3 times from the as received condition:
1650F for 10 minutes and cool to black
1550F for 10 minutes and cool to black
1450F for 10 minutes and cool to black
The idea behind this protocol was to add another lower cycle compared to my test knife. Once again, nothing changed. I know it looks better in the picture but it was not.


Here are all the coupons so far, from left to right coupon #1, coupon #2, coupon #3.


Finally, coupon #4. On this one, I did the same three cycles as coupon #3 and then added another one at 1365F and let it cool in the oven (is this slow enough for a DET anneal?)
On this coupon there might have been the slightest improvement. However, I would not call it acceptable. The grain structure was still visible with the naked eye.


From my understanding, the last coupon should have given me a fine pearlite structure, right?
What would be your suggestions?

EDIT: forgot to attach the pictures.

Thank you in advance!

 

[Larrin]

1500F is too hot for 1084. How thick are the coupons?

 

[Tony Mont]

1500F is too hot for 1084. How thick are the coupons?

All coupons were 0.196"/5mm thick with a notch ground about half way into them to snap them easier.

Should I conduct the same tests but lower my aus temperature to 1475F?

 

[Larrin]

All coupons were 0.196"/5mm thick with a notch ground about half way into them to snap them easier.

Should I conduct the same tests but lower my aus temperature to 1475F?

It could be with the thick steel you are either undersoaking at austenitizing temperature or not quenching fast enough. When you get soft steel it also looks like coarse grain when it is broken.

You should grind one down to 1/8" and repeat with 1475F.

 

[Tony Mont]

It could be with the thick steel you are either undersoaking at austenitizing temperature or not quenching fast enough. When you get soft steel it also looks like coarse grain when it is broken.

You should grind one down to 1/8" and repeat with 1475F.

A couple of things I did not mention in my original post. All the coupons were soaked for at least 10 minutes and I'm using a parks50 equivalent oil.

With that being said, I'm going to repeat all the tests with a thinner cross section, 1475F aus temperature and make sure I do not undersoak.

 

[Joshua Fisher]

Where did you source the 1084 from and roughly when? There was a batch of 1084 from GFS in Sheffield that had issues with grain structure from the mill and no amount of cycling was able to fix it, I tried probably a dozen combinations or processes including DET and full annealing cycles and nothing seemed to have an effect on the grain structure on any of the test coupons to try and figure out the issue.

 

[Tony Mont]

Where did you source the 1084 from and roughly when? There was a batch of 1084 from GFS in Sheffield that had issues with grain structure from the mill and no amount of cycling was able to fix it, I tried probably a dozen combinations or processes including DET and full annealing cycles and nothing seemed to have an effect on the grain structure on any of the test coupons to try and figure out the issue.

The steel was recently bought from a local Greek supplier. However, the bar has a sticker from GFS and have no Idea when they bought it.

 

[Joshua Fisher]

The steel was recently bought from a local Greek supplier. However, the bar has a sticker from GFS and have no Idea when they bought it.

Could likely be from the same batch, I’d test what Larrin recommended first but if you still get the same results I’d contact either the local supplier or GFS directly and just get some new material. If you get a replacement bar from GFS definitely test a piece before making anything from it. I don’t know if they ever figured out what was wrong we were basically waiting to hear back from them or the mill on the issue and I never got a final update.

 

[Tony Mont]

I conducted more tests following Larrin advice. All coupons were quenched in parks50 equivalent oil.

Coupon #5- 0.196 thick piece. Aus temperature at 1475F from the as received condition, the kiln was already hot when I put it in. No change at all. The new coupon is on the far right and to the left are yesterday's tests.


Since I was wondering about my kiln's accuracy, I heat treated a 1095 coupon because I've been using this steel a lot. I used 1475F and 10 minutes soak. As expected, the grain was perfect. Not visible in the naked eye. The 1095 coupon is roughly 1/8 thick. Here it is compared to coupon #5.


Coupon #6 was roughly 1/8 thick and cycled 3 times:
1650F-10 minutes soak
1550F-10 minutes soak
1450F-10 minutes soak
Aus temperature 1475F with a 10 minutes soak. Again, no improvement. Coupon #5 on the left, coupon #6 in the middle and the 1095 on the right.


Currently, I have coupon #7 soaking at 1365F and will let it cool in the kiln (it was cycled 3 times, same as coupon #6). It's roughly 1/8 thick and will be heat treated tomorrow.

 

[Ken H>]

That 1095 coupon shows you know what you're doing. Looking forward to the 1/8" thick 1084 coupon tomorrow.

 

[Tony Mont]

Just heat treated coupon #7.
Roughly 1/8" thick, did the following cycles
1650F-10 minutes soak
1550F-10 minutes soak
1450F-10 minutes soak
1365F-30 minutes soak and will let it cool in the kiln.
Aus temperature 1475F-10 minutes soak.
Again, no improvement compared to the other 1084 coupons.

Side by side with the 1095 coupon from yesterday.

And another compared to the previous 1084 coupons.

 

[Ken H>]

I think you've got some bad 1084 - it's way beyond my knowledge how this could be, but something ain't right, and I don't think it's anything you're doing {g}

 

[Stuart Davenport Knives]

This may indeed be a bad batch of steel, it is pretty odd. I don't know if this will help at all, but there was a batch of W2 floating around some years back that simply would not harden properly. And I do not have first hand experience with this, but was told this story by another maker who sometimes will pop in and out of Shop Talk. They tried normalizing at the usual 1650F and nothing. 1700F and nothing. They normalized at 1900F and then the steel hardened correctly, getting 67HRC+ post quench. I am assuming cycling and possibly annealing was done after the very high normalizing heat.

The problem with that batch was very very heavy spheroidizing (annealing) done from the mill as received. It took that high of a heat to break up the coarse carbide structure, freeing up the carbon to get proper hardness. It is possible, although admitted slightly possible, that this is the issue you're having. The factory anneal was just way too coarse. I really don't know, but I don't doubt was I was told, especially who told me. I never used much W2, or 1084, so I can't directly speak to that issue personally.

Other than just a really bad batch of steel, this does all seem odd. I assume you're not taking hardness readings off of the coupons, just checking grain size? I used quite a bit of 52100 with normalizing/cycling and every time I snapped a coupon it just looked like smooth gray to the naked eye. But 52100 has the carbides to pin grain boundaries where 1084 is about as simple of a steel as it gets.

I don't know Tony. After a puzzle like this, I would be inclined to try it myself with the 1084, but won't be holding my breath.
Normalize at 1900°F for 15 minutes. Air cool. This will for sure break up any possible heavy spheroidizing and reset the grain structure.
Cycle at 1650°F for 10 minutes. Air cool. This is the industry standard normalizing heat for 1084, but it will refine the grain after 1900F
Cycle at 1475°F for 10 minutes. Air cool. Further grain refine
Skip the anneal. Let's quench from pearlite structure and shorten the austenitizing soak time
Austenitize at 1475°F for only 5 minutes and quench in P50.

See what, if any, that does. Good luck, mate!

 

[Stacy E. Apelt - Bladesmith]

I agree with Stuart. I think it needs a very high normalization and then some grain refinement, then austenitize and quench.
We have seen this scenario before and 1900°F did the trick.

 

[Joshua Fisher]

If it’s from the same batch of GFS 1084 I tested I tried a high heat cycle between 1900-2000 and then cycled from there and it still didn’t correct the grain size. I tried both cycling and annealing after high heat and neither had any effect.

 

[Tony Mont]

First of all, I'd like to thank eveyone for helping solve this conundrum.

Stuart Davenport Knives Other than checking with a chainsaw file, I do not take any hardness readings off the coupons. The knife that set this chase into motion was a forged blade, that's why I haven't tried any high heat normalization.
However, I'm going to follow your suggestion since it will be in a more controlled matter and see what results it yields.

I will report back in a few hours.

 

[Tony Mont]

Just snapped the latest coupon.
1900F-15 minutes soak and air cool
1650F-10 minutes soak and air cool
1475F-10 minutes soak and air cool
Aus temperature 1475-5 minutes soak into p50.

The latest coupon is the one of the far right. No improvement.


And here it is compared to the blade that set these tests into motion. Compared to the blade there is the slightest improvement but still nowhere near how it should be.


After all the cycles, the coupon had this spiderweb pattern on it. Not sure whether it matters or not but it looks cool nonetheless.

 

[Ken H>]

Yep, that spider pattern does look neat - I hope somebody can chime in what it's all about. What about those "dimples" that look like where you might have used a drill to "dimple" a few places? What are those?

After your last trial, I'd be tempted to give up on that batch of steel. I hope you don't have lots of it?

 

[Keanen]

I had that same problem awhile back when I first got my kiln and I thought it was my thermalcupler. But when I tried some 1084 from NJSB the grain was perfect. I tried everything but didn't get good results. It was Sheffield steel from GFS knife supply that I purchased through Pop's. I talked to Pop's and the refunded me for the steel. I asked a metallurgist and we didn't find a heat treatment that worked. I did a ton of testing. 1400 for 15 minutes got me the best of the poor results, although that seems really low.

 

[Stuart Davenport Knives]

I'm sorry. I didn't realize the knife was forged, otherwise I would never have suggested the 1900°F normalizing heat. Your forging for sure broke up any possible heavy spheroidizing.

I just looked at your OP and saw that you mentioned it was forged. Doh! My bad!