[TheForge] Frozen anvils, round 2.5?

[Dennis Presser]

I am a beginning blacksmith, I live in an occasionally cold environment, and 
before I bought my anvil I asked a couple of friends about cold-fracturing 
an anvil during use.  They are chemists, not metallurgists, and tempered 
their answers (and I am keeping them anonymous); still, I think it is good 
information to share.  Here are my questions and their answers, paraphrased:

This is a set of questions for a metallurgist, but I don't know a 
metallurgist, and metallurgy is a sub-discipline of chemistry, right?  I 
have heard of engine blocks cracking in extreme cold.  I have also heard 
that an anvil will crack if it is hit when it is cold.  I have an anvil, and 
my shop is unheated except when I am in there with the forge going.  So, my 
series of questions:
·	How cold does the ambient temperature have to be, to have steel crack 
under stress?
·	The steel from which my anvil is made is a manganese-steel alloy; what 
effect does the alloy have for the answers to these questions?
·	Is it the cold itself that is the issue, or is it the difference between 
the ambient temperature of the anvil and the temperature of the hot metal 
placed on it?
·	One author recommends "preheating" an anvil by placing hot irons from the 
forge on the anvil to warm it up before working on; does this sound like a 
sensible thing to do?
Thanks for your help in clarifying these things for me.

B.F. [Chemist 1]:  At low enough temperatures, the ductility of metals 
decrease to the point that they begin to demonstrate brittle fracture.  This 
leaves a rough surface like broken cast iron (In the cast iron case there is 
enough silica in the metal that it is an inherently brittle material.) 
Microscopically, ductility involves planes of metal atoms slipping past 
adjacent planes within a single metal crystal. For really large-scale 
movement some slippage between individual crystals occurs as well.  This 
generally does not happen to a great degree because impurities and crystals 
of other alloys can 'pin' these dislocations to one spot and prevent them 
from allowing the movement. Conversely, hot metal allows the easy movement 
of planes of atoms  - thus the need for a hot forge.

As an aside:  The ductility decreases and hardness increases as the crystal 
size decreases.  This is why quenching and work-hardening work.  Both cause 
the decrease in crystal size.  Working a metal also forces more of the 
impurities out of metal and to the crystal edges where they pin the 
dislocations and further harden the material.

So, if your anvil is cold enough all of the force of your hammer blows will 
be sent through the body of the anvil, possibly focusing on a weak point and 
causing failure. If the surface is warmed sufficiently, at the point of 
impact there will be some amount of ductile flow, dissipating the energy to 
a more manageable level.

The other issue is thermal expansion.  Metals generally elongate as they 
warm.  If the temperature gradient is large, the strain between the small 
cold metal and the bigger hot metal will overcome the chemical bonds holding 
it all together and again cause failure.  You may have seen this with glass 
or when putting a cold cast iron frying pan directly into the fire.  Putting 
some hot irons on the anvil may warm it to a depth such that the temperature 
gradient is small enough that the anvil can handle it.

How cold does ambient temperature have to be, to have steel crack under 
stress?  Depends on the stress, of course, and the steel.  My experience is 
that zero is not a big deal, as far as common metals go.  -20, -40, I'd 
start being more cautious.  A lot of interesting things start happening when 
you get to -40.  Recognize too that the temperature differential is 100 
degrees more than at 60!  (absolute zero is only -400 - odd, so you've gone 
a serious fraction of the way!)

·	The steel from which my anvil is made is a manganese-steel alloy; what 
effect does the alloy have for the answers to these questions?

A lot.  The ductility, stress/strain relationships, thermal expansion, etc. 
are all properties of the alloy, grain structure, and impurities.  The 
following from a web site found with a quick Google search 
(http://maxpages.com/msper2000/A1_Alloy):

“Hadfield’s Manganese Steel is a very useful engineering material due to its 
high toughness, ductility, high work-hardening capacity and good resistance 
to wear. It is particularly useful for severe service that combines abrasion 
and heavy impact as in power shovel loader, bucket teeth, railway frogs, 
rock crusher, etc.

“The as-cast Hadfield’s manganese steel is quite brittle for normal use due 
to carbide precipitates [CITATIONS DELETED].  It does not have a sufficient 
physical properties to withstand impact due to carbides behavior which tends 
to brittleness.  Such an as-cast structure should be solution treated to the 
austenitizing temperature, at a certain holding time, then water quenched in 
agitated water to preserve the full austenite structure.  The mechanical 
properties are greatly affected by changes in its microstructure. The 
perfect dissolution of grain boundary carbides to austenite grains has a 
marked effect on its susceptibility to withstand impact under service 
condition. The toughness of these alloys is dependent on precipitates formed 
at grain boundaries. It is thus desirable to investigate these phenomena in 
these alloys. The present paper describes experimental results an 
investigation of as-cast and solution treated Hadfield’s manganese steel 
alloy. . .

“Austenitic manganese steel has a lower thermal conductivity and a higher 
thermal expansion ratio of about 1.5 times that of carbon steel.  Rapid 
heating has might cause cracking on grain boundaries due to the thermal 
stresses.”
---------------
So it looks as though you need to worry about thermal stress with 
austentitic manganeese steel while cast Mn-steel has grain-boundary deposits 
that make it more brittle at all temperatures.

D.F. [Chemist 2]:  My understanding is that cracking is not due to low 
temperatures per se, but rather to a fast change in temperature (creating a 
"temperature gradient" in nerd-speak).  Most materials expand while heated 
and contract when cooled. If a big hunk of cold metal is heated rapidly in 
one place, the hot area tries to expand, but the rest of the hunk stays cool 
and doesn't expand. This creates a stress between the hot and cold areas.  
Pounding on the anvil with a hammer makes cracking even more likely under 
those circumstances.

Based on that reasoning, the idea of preheating the anvil strikes me as 
eminently sensible. If it's heated up slowly (and without pounding) there is 
time for the anvil to reach a uniform temperature, and there will be a 
smaller temperature gradient (and less stress) when you start pounding the 
hot metal on it. I would recommend placing the hot irons on opposite sides 
of the anvil, to reduce the temperature difference experienced by different 
parts of the anvil.

I don't have specific answers about safe temperatures, alloys, etc. I guess 
what's important is the temperature difference (bigger difference, more 
likely), how rapidly the heat is transferred through the metal (depends on 
the alloy and its size), and how strong the metal is (again depends on the 
alloy).