Tempering, I'm confused....

[Gerhard Gerber]

I've run into what I feel is conflicting advice regarding tempering, even here.

This is not about one alloy in particular, so if the mods feel this belongs in another section please feel free to move it......

My understanding from the charts that I've looked at basically comes down to resulting hardness being a function of time and temperature, with more heat, longer time or both leading to lower hardness and increased toughness as result etc etc.....

Generally the temperatures quoted are the same, but I've seen 2 x 2 hour temper cycles being recommended.  

I can fully understand wanting to draw back hardness on some types of blades e.g. swords and machetes, but 

 

I've watched the Real Engineering HT video several times and I'm getting closer to understanding what happens to steel during the HT process, but tempering is still a mystery to me in many ways.

First and foremost I struggle to understand how a relatively low temperature can have so much effect, but also the role that time plays.

Am I missing something or perhaps interpreting the charts wrong?

[Joël Mercier]

Ok, i'll try to answer this one 

Tempered martensite gives more "room" to retained austenite to change into martensite. So, the relatively long tempering is to allow that change of phase and temper the newly made martensite. 

[Alan Longmire]

The "longer time" thing is as Joel said.  It does not, however, mean that if you leave a blade cooking overnight that it will be softer than one cooked for two one-hour cycles.  I haven't had enough coffee yet to explain further, hopefully someone else will chime in...

[Jerrod Miller]

6 hours ago, Joël Mercier said:

Tempered martensite gives more "room" to retained austenite to change into martensite. So, the relatively long tempering is to allow that change of phase and temper the newly made martensite. 

I'd like to clarify this a little.  A hardened blade may or may not be 100% martensite.  This is dependent on alloy and heat treat process, which we'll ignore for now.  If at the end of a quench there is any retained austenite (that which did not convert to martensite) then heating the steel up a bit (the temper cycle) can give enough energy to convert the austenite to martensite.  Double tempering gives 2 chances for this, and ensures that the vast majority of all the martensite, including the sections that were retained austenite that converted to martensite during the temper, gets tempered.  

As for the time:  A phase change happens at the speed of sound for the material.  So when a given section of steel crosses a transformation temperature (heating up or cooling down), the arrangement of atoms shifts really fast.  In this case we are talking about BCC --> FCC --> BCT (body center cubic = ferrite, face center cubic = austenite, body center tetragonal = martensite).  Now, the whole blade won't change that fast, because the temperature in the blade doesn't change that fast.  It is just that when the temperature DOES change, it is a really quick shift of the atomic lattice.  I put this here just to say: this isn't how tempering works.  Tempering is better described as a diffusion process.  When we heat up BCC and it shifts to FCC, life is good and there good movement of the carbon was the gap in the lattice structure for it moves.  When we go from FCC to BCT, the carbon atom gets pinned not quite where it would like to be, thus BCC doesn't form, as the lattice has a carbon wedged where there isn't supposed to be a gap.  This stretches the lattice, thus putting it in a stress state.  When we temper steel we are giving it the energy needed for the carbon to diffuse out of the most stressed position, but only enough to get slightly less stress.  The more energy (heat), the more the carbon can get "unwedged" in the lattice.  Diffusion times energy (heat) and time.  Thus the soak time during a temper.  Energy is much more important, as that really is what allows the carbon to move.  The time length the steel is given this energy just ensures the carbon has time to use that energy.  

Think loose rubber band versus stretched rubber band.  The stretched rubber band is very difficult to stretch further, and more stretching will cause it to break.  Tempering is kind of like "un-stretching" the rubber band just a tiny bit.  Just enough to get off the breaking point, but not so much that it is loose.  

[Gerald Boggs]

That was great, thank you.  Now let me hijack a bit: Is there a point/ range in tempering time that will result in the BCT converting to BCC?  IE, if I left it in the oven for a day at 400F

[Jerrod Miller]

Technically maybe, but I don't think so.  Martensite is metastable.  Meaning energy has to go in to change it, it will not decay into another state over time all on its own.  See the image below (stolen from a quick Google search).  Energy put in will move the orange ball up the slope to the right, assuming enough time is given for it to move there.  But really think of it as energy gives potential to move up, time lets it use that potential.  400 will let it get so far up, but 600 lets it get even further up.  I know this is essentially true for practical time periods, like up to days.  If you held it at 400 F for a couple centuries it may change.  

 

[Gerald Boggs]

Thank you very much.

[Alan Longmire]

Dang, it's good to have an actual metallurgist on the board!  Thanks for the in-depth explanation, Jerrod!

[Jerrod Miller]

Let's just hope I'm right...

I'm pretty sure that is all sound.  A bit simplified, but why get even deeper?  

[Gerhard Gerber]

Thank you Jerrod, I think I get what you're saying, but I will need to break down and study your answer....along with watching that video a few more times and bed down proper understanding of the process.

From what I THINK I understand, temperature rather than time determines the eventual hardness, and a longer temper is beneficial.....?

By necessity my 2nd temper cycle is usually done between 10pm and 11pm, so I quench between tempering cycles........this now seems like a bad idea?

[Jerrod Miller]

12 hours ago, Gerhard Gerber said:

From what I THINK I understand, temperature rather than time determines the eventual hardness, and a longer temper is beneficial.....?

For practical purposes, yes.  As far as I know, things can change over really long times.  But if we are talking a few hours long per temper, you won't see much difference in blade sized parts tempering for 1 hour versus 5 hours at the same temp.  100 hours may be another story.  A higher temp for a 1 hour soak will definitely have a big impact (assuming it is not just a 5 degree higher temp).  

12 hours ago, Gerhard Gerber said:

By necessity my 2nd temper cycle is usually done between 10pm and 11pm, so I quench between tempering cycles........this now seems like a bad idea?

Not a problem at all.  

[Brian Dougherty]

I won't pretend to understand this as well as Jerrod, but I will offer another way to look at this for those struggling to understand why 2 hours is better than an hour, but 5 hours won't make a difference.  Think about it this way: 

Just because a reaction takes 2 hours to complete doesn't mean it will continue on indefinitely.  At some point the structure will have relaxed as much as it is going to for a given temperature. (For all practical purposes)

[Jerrod Miller]

14 minutes ago, Brian Dougherty said:

I won't pretend to understand this as well as Jerrod, but I will offer another way to look at this for those struggling to understand why 2 hours is better than an hour, but 5 hours won't make a difference.  Think about it this way: 

Just because a reaction takes 2 hours to complete doesn't mean it will continue on indefinitely.  At some point the structure will have relaxed as much as it is going to for a given temperature. (For all practical purposes)

Yes.  This.  At the very least the process slows down significantly, to the point where practical time limits give no measurable difference.  

[Gerald Boggs]

Since Jerrod appears to be willing to let us bend his ear :-)  (thank you for your forbearance)   Could you explain work hardening and stress?

[Doug Lester]

Work hardening is caused by the atomic bonds between the iron atoms being physically stretched by the hammering or bending and increasing the stress on the bonds.  This is not just something that works on iron and steel though.  Other metals are subject to this.  For instance, the only way bronze and copper tools were able to be hardened, as during the Bronze Age, was to work harden by hammering.

Doug

[Jeroen Zuiderwijk]

I've never understood what longer duration does either. As far as I've experienced, tempering is pretty much instant. If you only for a split of a second get your temperature too hot, your hardness is ruined. Might it be that the time does not really reduce hardness, but due to allowing more time to let the atoms find the least stressed state, it does keep on increasing the toughness? 

[Doug Lester]

Just one thought on work hardening.  There are some out there that advertise their "secret" method of edge packing (just a pseudonym for work hardening by hammering) that will give a superior edge to your blade.  They "pack" the edge and then put the blade back into the forge ignorant (or maybe just in denial) of the fact that when the steel changes phase which it will to on the way up in temperature and the way back down those atomic bonds that were stressed will reform.

Doug

[Jerrod Miller]

Doug beat me to it on work hardening, and I have nothing to add to his description/explanation.  

Jeroen,

Time is most important for austenite conversion and secondarily to actually tempering the martensite.  The best practice is to temper a little on the low side and let it soak to be fully up to that temperature, then if needed you can re-temper a little hotter and sneak up on the final temper desired.  If you just hit it for a short time with temperature you may or may not get the exact temper you want.  It is all about stable, repeatable results.