Cast Steel, Cast Iron, Forging vs. Casting, and Understanding Grain

[Patrick Nowak]

Let's talk just about grain size. To do that we first have to know what we mean by "grain" in this context. For purposes of this topic a grain is the same thing as a crystal. A crystal is just a group of atoms that are arranged in a repeating pattern. In the case of iron with no other elements, inclusions, impurities etc., iron can have two different atomic arrangements. One is body centered cubic and the other is face centered cubic. What this means is that atoms are arranged such that they form a network that can be broken down into cubes. Imagine a square frame completely filled with pool (billiard) balls. Each ball represents one atom. Now imagine a second frame of balls positioned on top of the first. The natural tendency is for the top layer to nest in the depressions between the balls of the first layer. Each ball in layer two is touching 4 balls in layer one. Adding additional layers in the same way gives you a stack of balls that is pretty stable. that is the body centered arrangement. If we balance the second layer on the first layer such that each ball in layer one touches only one ball in layer two and we stack additional layers in the same fashion we now have the face centered arrangement. There are metals other than iron that have these structures but for our discussion we are assuming iron for all the examples. In the iron system, the BCC or FCC arrangement is dependent on temperature and if we heat/cool pure iron under thermodynamically equilibrium conditions the change from BCC to FCC happens at about 1333 F. At very high temperatures, well in excess of 2000 F, iron will switch back to BCC, but that structure is not of significant importance to blacksmiths so I am ignoring it.

Our stack of pool ball represents a single crystal. If we have other stacks that are next to the first, but not in the exact same alignment, then we have a new crystal or group of crystals. They can all be the BCC or FCC, but they have different orientations and are therefore each a unique crystal. The size of that crystal can be measured using a microscope and we know that some performance properties are influenced by the size of those crystals so it is good to know how to control that property. the size of a crystal generally increases with increasing temperature, but the RATE of growth is temperature dependent. Grain size will change slowly, or maybe not at all for practical purposes, at low temperatures. In general practice, we assume that BCC iron really doesn't change size with increasing temperature, though it may be measurable with the right instruments. Even if it does, it is not significant to what a blacksmith is doing. So, for practical purposes, grain size does not change when the grains have the BCC arrangement. 

When we heat iron above 1333 F, the BCC arrangement is no longer the favored arrangement and the atoms begin to rearrange themselves into the FCC form. This happens by a process call Nucleation and Growth. First, a tiny FCC crystal nucleates in a pool of BCC crystals. This little FCC crystal starts to grow by consuming the BCC crystals via a diffusion of atoms from one position to a new position. Of course, lots of new FCC crystals are nucleate at the same time and they are all growing by consuming the BCC crystals so pretty soon you have no more BCC crystals. If the temperature is just hot enough for the FCC structure to be stable, let's assume 1350 F, then the nucleation and growth will be fairly slow and will be even slower once all the BCC crystals are gone. If however, we heat up to some very high temperature, say 2000 F, now there is enough thermal energy to cause the FCC crystals to continue to grow by consuming each other. The thermodynamic driving force for this is the reduction of crystal boundaries. The fewer there are, the more stable the system is from a thermodynamic perspective. Growth will stop once the crystal size vs. temperature has reached a balance or equilibrium condition. Raising the temperature to say 2300 would cause additional growth until a new equilibrium condition had been reached. 

During forging, we generally start at a temperature of 2300 give or take a bit but then we start to deform the metal. When we do this, we are causing the atoms within each crystal to get knocked out of alignment. this misalignment is called a DISLOCATION. We can have an enormous number of dislocations within a single crystal since these crystals have thousands or maybe even millions of layers of atoms within each one. Each disrupted layer is a single dislocation. Dislocations cannot pass from one crystal to another. They are stopped by the boarder or boundary (we call it a grain boundary) between each crystal. They can also be stopped by other things within the crystal such as alloying element. Once a dislocation hits an obstacle within the crystal or at the crystal boundary it stops moving. New dislocations are formed and will move, but they will get caught the dislocations already stopped by obstacles so pretty soon you have this tangle of dislocated atoms and at some point you will not be able to put in any more dislocations. (This is why a paper clip breaks when you bend it back and forth while fidgeting in meetings).  So, as dislocations build up within a crystal, that crystal is no longer in equilibrium and at some point, the number of dislocations will be so high that a new crystal will form to get rid of those dislocations. When new crystals form for this reason that is called DYNAMIC RECRYSTALIZATION. This is very common during all kinds of hot working operations. The actual temperature at which it happens and how fast it happens will be a function of how much deformation has been done (strain), how fast the metal was deformed (strain rate) and the temperature at which the forging was done. Dynamic recrystallization is different from the other type of recrystallization event because it does NOT depend on a change from BCC to FCC. Dynamic recrystallization happens when the iron is already in the FCC arrangement. How big the crystals will get after a dynamic recrystallization event depends on how many dislocations there were to start with since these really are a form of stored energy and how hot the iron gets. The more stored energy and/or the higher the thermal energy (temperature) the larger the resulting grains can be.

So that is all super interesting, but where is the practical application? 

The practical application of this is in grades which do not undergo a change from BCC to FCC like the 300 series stainless grades. Controlling the process variables lets you control how big the grains or crystals will be in those materials. In the case of CARBON AND ALLOY STEEL (not pure iron) we can forge at low temperatures near the end of the job to add dislocations, trigger recrystallization and, by keeping the forging temperature very low, say less than 1600 F, we can keep the as-forged grain size small. This has very limited practical benefit with moderns steels, but if you are working with historical steels (bloomery steel, blister/sheer steel etc.) in which there are almost no alloying elements other than carbon such a practice can be helpful in ensuring the final grain size after heat treatment is small, which is beneficial for toughness.

In modern steels, most contain very small amounts of aluminum or other elements that can be used to keep grain sizes small during heat treatment so that we don't have to limit ourselves to finishing forgings cold. The process works like this:

Aluminum will react with nitrogen, both of which are frequently found in modern steels to make an Aluminum Nitride. Aluminum nitride will dissolve in austenite at temperatures over about 1800 F, so it is not a discreet particle during the forging process. Grain size during forging is allowed to change via the dynamic recrystallization process discussed above. Once the forging is complete, it is cooled to room temperature and then subjected to the heat treatment cycles. A normalizing cycle may or may not be done depending on a variety of factors but if it done, the piece is not heated above 1750 F. In the range from 1350-1750 BCC crystals transform to FCC crystals and will grow, but that growth will be arrested by the presence of the aluminum nitride particles, which act as fence posts at the crystal boundaries, pinning them in place. If the Aluminum nitride was not present, the crystals would continue to grow. After cooling from the normalize, the crystals revert back to BCC. The piece is again heated to a temperature above 1350, this time to 1650 F. New FCC crystals are formed, but this time they will not get as big as they did when heated to 1750 F because there is less thermal energy to push them to get bigger. At this point the piece is quenched and after that it is tempered. There is no change in crystal size during tempering because that is done at a temperature less than 1333 F. If an extra small crystal size is desired, multiple normalizing cycles can be done, but for this to have the desired effect, each one must be at a lower temperature than the previous. So that would look like this: Normalize #1 at 1750, Normalize #2 at 1700, Austenite for quench at 1650.

In industrial forge practice, multiple normalizing cycles are the preferred method for controlling grain size in carbon and low alloy steels since there is more control over that than than the finishing temperature of the forgings. Note that low temperature forging is sometimes advocated in hand forgings of blades and in that field it is described as "edge packing". That is really a misleading term as there is no increase in density along the edge, but if it is done correctly AND the heat treatment cycles are handled properly it can help to give a final grain or crystal size along the edge of the blade. A more reliable method of achieving the same result is to use a steel of know composition with elements that will form grain boundary pinners. I already mentioned aluminum nitrides, but some others are vanadium and tungsten carbides and columbium (niobium) carbo nitrides. You really don't need much of these elements to get this effect, 0.01% usually is sufficient. If you are working with a high alloy knife steel you probably have much more of these elements than you need since they have extra high concentrations of these alloys for improved cutting performance.

[jlpservicesinc]

Great write up..    So if this is the case of "Grain" and "Crystal".  What are we actually supposed to call the composite units of say a hardened item when it is fractured and you see the jagged material inside. 

Grain often talked about as course, medium, fine or pearl like. 

From what I know of this was always referred to as grain structure. 

[anvil]

Interesting. Especially about packing. Back in the day via experience, Smiths discovered that if they basically followed the process you described, they got a better blade. They called this process "edge packing"

Lol, Jen, I'd call it an improper heat treat, figure out where i messed up and learn from my mistakes and not get bogged down in semantics. 

[Steve Sells]

In my apprenticeship I learned to edge pack as well, but the problem is that edge packing doesn't work that way because anything done to the grain during the 'edge packing' phase is undone during the thermal cycling of the heat treatment process, what it does do is give a smoother finished product as a result of the blows given as the temperatures get lower

[Frosty]

Thanks Patrick!

Patrick will correct me if I'm wrong, please. Jennifer: I believe what's called "Grain" originates from a crystal structure large enough the glint from crystal faces looks like grains of salt. It's a term from a time when nobody KNEW atoms and molecules from ant colonies and mollusks. Many knew crystals and how to grow them though, they just never made the connection.

As experience has shown us joints are the weak points of objects, mortared brick walls tend to break along the mortar, welds break beside the weld bead, etc. The larger the joint the weaker it becomes. Crystals are no different, crystals can NOT move past each other easily the larger they are the more difficult. For a grossly exaggerated example visualize a bag of crushed pea gravel and an equal sized bag of bricks. Which changes shape more easily?

I've been saying this for a long time but everybody gets stuck thinking steel has a "GRAIN" structure, it does NOT it has a "CRYSTAL" structure. It's much easier to understand how crystal structures form and behave than a mythological concept like "steel grain."

By breaking up crystals the boundaries become smaller and less resistant to movement and so less brittle. 

In this sense, think of crystal boundaries as stress risers or cold shuts. Instead of force conducting smoothly around them it is interdicted by large sheets of crystal face that is NOT bonded equally strongly to the crystal face or face in contact so if it is forced to flex the weaker bond breaks. 

Patrick? Close for a lay handle on the subject or way off base?

Frosty The Lucky.

[Patrick Nowak]

Crystal boundaries are indeed a weak point in the crystal structure.  I personally would not use the analogy of cold shuts or stress concentrators since I would put those other features into a different category, but  in the sense that they help you understand that the boundary between crystals is a weak point, they are correct comparisons. The size of the crystals can be gaged by looking at fracture surfaces. When the steel being evaluated has a high hardness, as in the case of tool steels that are quenched and tempered, the fracture appearance can be pretty closely correlated to the ASTM grain size #. At one time there was a test call the Shepard Fracture test. It was possible to obtain specimens showing different fractures and these could be compared to the item of interest and to get a pretty good idea of the grain or crystal size. Today this evaluation is usually done with a microscope. 

A further note about edge packing. Steve makes an excellent point about the improved surface finish achieved by the process. One situation in which the technique could provide some benefit is if the blade being made was NOT going to be heat treated. This would be the case for some historical blades. there are quite a few examples old blades that show no evidence of heat treatment after forging. the reason is because the dislocations created by the process could remain (since there is no phase change and the work is done at fairly low temperatures). These dislocations will result in an increase in hardness relative to an item that has not been subjected to low temperature forging and this increase in hardness could give improved performance. When properly heat treated modern steels are used there is really no metallurgical benefit.

However, there is a whole subset of industrial forge work called "cold forging" in which the forging temperature is kept below the temperature needed to have dynamic recrystallization. In these products, the dislocations created during forging do contribute to the final performance properties of the part.

[George N. M.]

Just to throw in a bit of mineralogy into the discussion:  Iron crystals act the way they do because they are roughly the same dimension in all directions (cubic, like salt) and lock together well.  Other minerals have a crystal structure which is asymmetric where the crystals are longer or shorter in different directions.  This results in more pronounced zones of weakness, often called cleavage.  One of the minerals which shows this in a pronounced way is mica (various types, including our friend vermiculite).  The layers in mica are crystal boundaries.

Different crystal shapes and properties are why pure metals behave differently than other elements and other minerals.

"Edge packing" can also be useful when working with copper based metals such as bronze.  Work hardening an edge in bronze can make it more durable.

Speaking of vermiculite, the reason it makes a good, fireproof insulator to cool things slowly is that it has some unusual mineralogical characteristics.  It is a mica that has water molecules incorporated into its crystal structure.  So, when it is heated the water expands and the mineral "pops" like popcorn and creates a very low density, expanded material which is fire proof because it is still a rock and pretty chemically inert.

"By hammer and hand all arts do stand."

[JHCC]

Not dissimilar to the water of crystallization trapped within the hydrates of borax, which is what makes it foam up when first heated.  

[dian]

here some thoughts.

grain flow: it depends if you look at some 100 ton slab a 50 ton part will be forged from, that might include 1/4" sized grains or a small low quality mild steel part that will be formed with a few blows (like the hook below) or a high quality part out of "good" alloy steel where forging is done to improve the microstucture (with quenching in between) or even some vim-var/esr melted, ultra fine grained steel. the first might really have elongated grains, the second elongated "inclusions", like fayalite stringers (or oxides of bi-films), the third will often have banding (segregations of constituens and elements) while the last will have a "clean", uniform and anisotropic structure. (thats when you would machine a crank from "billet".)

typically you see the expression "grain flow" in older literature, where they used "dirty" steel, while lately "metal/material flow is used".  so its really more about the orientations of segregated areas having better propertier in the longitudinal direction.

grain size: grain refinement up to 1000x is possible with hot working (rolling, forging). usually grains get reduced considerably (2-30x) provided its done right. and its easy to do it wrong because success depends on so many variables. i have included the example of a simple hook, where grain reduction in the neck takes place on the last pass. so if you screw that up the part will probably brake. (notice that the grain is groving in the body with subsequent passes.)

grain size is one of the rare ways to increase strenght and toughness simultaniously. But: that is (like always) with everything else being equal. example: you figure you want smaller grain size and reduce the forging temperature. good idea, however if this results in primary ferrite on grain boundaries (especially if its blocky) you have shot yourself in the foot.

forging mild steel hook 

btw, if we talk about grains, we mean prior austenite grains. e.g in perlite colony size and interlamellar spacing are im portant. there will be a number of colonies in such a grain. sometimes a special heat treat is neccessary to make the prior austenite grains visible.

hook.pdf

https://core.ac.uk/reader/195109013

(i was talking about fig. 8.)

[Patrick Nowak]

Dian- Thank you for sharing the paper on forging of that hook. It does a nice job of explaining how grain size is affected when no heat treatment other than quenching from the finished forging temperature is involved. I've already given a pretty detailed explanation of how grain size is controlled by heat treatment. And you make a good point that the grain size being evaluated is indeed the prior austenite grain size. Generally this can be revealed by either an oxidation or carburization heat treatment which involves heating a small specimen to the austentizing temperature so that the austenite grains are outlined either by oxides or carbides. This allows you to see them under the microscope after the specimen has cooled, at which time you will typically have a ferrite/pearlite structure. It is important to note that both of these methods only demonstrate the capability of a steel to achieve a particular grain size when processed in exactly the same manner as the test specimen. When you want to know the actual prior austenite grain size you must etch for that without re-heating the specimen. That is usually done with a saturated picric acid solution.

Grain size in large forgings: The company I work for can produce forgings from ingots approaching 200 tons and there certainly are other companies in the world that can go bigger. The biggest ingots I know about were between 600 and 700 tons and were cast and forged in Asia. In my industry, I cannot recall ever encountering a carbon or low alloy steel forging that was going to be used without some form of heat treatment. Therefore, the prior austentie grain size going into service would never be 1/4 inch, though in the as- forged condition prior to heat treatment that could be a possibility.

Grain flow vs. Metal Flow- In the US forging industry "grain flow" is still the common expression, but that may not be true in other parts of the world.

Grain Flow behavior will indeed be influenced by inclusions of all sorts as well as chemical segregation. Most of the inclusions present in the steels I deal with are manganese sulfides. Occasionally we do see an exogenous non-metallics but the grain flow behavior is mostly influenced by sulfides in the steels we are working with. Certainly if you have severely banded structures, say of pearlite and ferrite or bainite/pearlite that will contribute to the grain flow behavior. It is rare for our forgings to exhibit isotropic behavior no matter what the source of the raw material. In our work, the use of VAR/ESR materials is not done to achieve fine grains or get a part with uniform properties in all orientations but to ensure that non-metallic inclusion particles are as small and as few as possible. The act of remelting does not by itself ensure small grains. Solidification takes place quite slowly in those processes so in the as-cast condition you still have large grains. This is particularly evident on as-cast titanium which has beautifully colored grains that are often 1/4 inch or larger.

[LeeJustice]

9 hours ago, Patrick Nowak said:

Therefore, the prior austentie grain size going into service would never be 1/4 inch, though in the as- forged condition prior to heat treatment that could be a possibility.

Good morning,  would you please clarify: as small as 1/4" or as large as?

[Steve Sells]

1/4 inch is huge

[Frosty]

Another definitive example of cut and paste pretend knowledge. 

Patrick: dian is just looking for attention, everything he's said so far is easily found on the internet. He's shown no actual knowledge of the subjects he pretends to. 

You have shown yourself to be knowledgeable and very good at presenting the information and addressing questions. Thank you I've learned a lot from every one of your posts. Unfortunately the quality of your knowledge is diminished trying to respond to dian's foolishness. 

Please, don't feed the trolls. 

I'm hoping you'll do your outstanding good job of explaining how monocrystalline and amorphous castings are made.

Frosty The Lucky.

[ThomasPowers]

Funny you should mention that: the swordmaker I worked under, his Father was a research Metallurgist at Batelle and had a number of patents on metal glasses made by touching a copper disk cooled with liquid helium and spun at a high rpm touched to the surface of a melt and throwing a ribbon of metal glass off due to the crazy fast cooling rate it created.  The swordmaker held a patent for using it to create pattern welded billets as placing the metal glass between CLEAN strips of other alloys, clamping and heating would result in a diffusion weld at temps substantially lower than forge welding temps generally are.

[Frosty]

By metal glasses we're talking single molecule thick 2D crystal sheets, right? 

That is a SERIOUSLY DEEP rabbit hole you're dangling in front of the gang Thomas. You have an evil sense of humor, IF educational. It's one of the things I like about you.

Frosty The Lucky.

[LeeJustice]

What I think I know about glass is that despite appearing to us to be solid, it is actually classified as liquid, as it has not a crystalline structure.  So it would seem that a metal glass would be solid metal which was without a crystalline structure.

[ThomasPowers]

What Lee said, they are generally thin as the cooling rates to get them can't be maintained in thick sections.  They have some "interesting", (academia for "crazy"), properties as the atoms are not bound in a crystalline lattice like slower cooled materials.

I wonder what other metals and alloys have been explored since I met them in the early 1980's...

[Frosty]

I believe that's silica glass and it's sort of a debate about it being a super cooled liquid rather than an amorphous solid that can flow. The easy experiment is to cantilever a sheet of window glass over the edge of a table and place weight on the end. In time the glass will bend.

Really old window panes ARE thicker at the bottom than the top though I don't know of any that are flowing over the edges of the frames. Most notable are stained glass in Cathedrals.

If you really want to do some reading about metal glasses and other monomolecule thick 2D sheet structures look into Quantum Computers. Quantum computer chips are some mind bending crazy reading. Just don't expect to understand what you're reading, it's quantum theory after all. Just don't forget your mental spelunking gear!

Perhaps, Master Slag will post a few links, for the interested. 

Frosty The Lucky.

 

 

[Chris Williams]

The aspect that most appeals to me about metallic glasses is how they seems to be the result of the atoms trying to arrange in five fold symmetry. Many of you will recognize that there are no natural  crystals with five fold symmetry. I don't know if any crystals have been created in labs with other than local five fold symmetry. I won't pretend to have researched it well enough to explain myself, but I will seed the conversation with a link. 

https://www.nature.com/articles/ncomms9310

 

[Patrick Nowak]

The professor I learned the most from in college, Dr. Carroll Mobley, worked with Bob Maringer, the inventor of the spinning copper disk technique at Battelle before he came to the university. He had lots of stories about that process and he had us use a variation of it to make rapidly solidified solder strips just to show us how it changes the solidification behavior. My favorite story of his was when he used that technique to rapidly put tinsel on the family Christmas tree.

A metallic glass is a glass because it lacks a crystaline structure. Most of the ones I know of are complex alloys and it is the high number of alloying elements and the atomic sizes of those elements that help to prevent the formation of the normal crystal structure. They do have fairly unique properties compared to normal metals. To the best of my knowledge the metallic glasses are really pretty thin due to the rapid solidification times required. As for single crystal castings, it has been a long time since I've given those any thought. I recall that control of solidification variables is critical. Also, you can get single crystals that are not the complext alloys needed for the metallic glasses. They are two very different situations.

The reference to 1/4 inch grains was intended to indicate that they are very large by normal standards. Usually grain size is measured with a microscope at 100x magnification and the units are in microns.

I do recall learning about glasses in general and the silica based glasses, though that too was a long time ago and is not my area of professional work. I have heard the statement about glasses being liquids and if given enough time they'll flow but I've also heard that the evidence for that (old windows) is based on a misunderstanding of the manufacturing techniques used and that old windows were in fact made with varying thickness. That makes sense to me if the panes were made by hand blowing since, in that case, the glass worker would blow a hollow ball, open the end and form it into a flat disk, all while the glass was still attached to either the blow pipe or the punty rod. In either case, the center of the disk was thicker than the edges so any panes of glass cut from that disk would tend to taper from the edge closest to center of the disk towards the outer edge. Once glass manufacture became more mechanized and they could make "float glass" then the thickness became uniform.

[Frosty]

I didn't mean to bring up mono crystalline nor amorphous castings to exercise your expertise, they're just as examples of how widely the manipulation of crystal structure extends. For example, jet engine compressor and turbine fans are mono crystalline titanium castings. The only thing I recall from reading about amorphous iron alloy castings is their superior strength do to having no crystal boundaries to act as fracture initiators. 

I misspoke, silica glasses is more commonly thought of as a super viscous plastic material rather than a rigid solid but that's not in all cases, especially in old days when every batch was a recipe of it's own. 

I'm sorry if I've added to the confusion in this thread it wasn't my intent. I'd like to thank you again for sharing your knowledge with us, your ability to put it in terms lay readers can grasp and research in more depth is very of high value.

Frosty The Lucky.

[ThomasPowers]

Want to see some good sized crystals in Ferrous material?  Look at the Widmanstätten patterns seen in some meteorites.

[Frosty]

Talk about long cooling times!

Frosty The Lucky.

[patrick]

The easiest way to see big crystals id to find a galvanized item: trash can, highway gaurd rail,  ductwork for a central heating system etc. I've seen grains much larger than 1/4 inch on all these things. Not every galvanized item will show these big grains but many do. The other place you can often see grains without the need for a microscope is old brass door hardware.  The are often castings with fairly large grains.  Over time the action of people grabbing the door handles will have an etching effect making the grains visible. 

When i was in college one of the examples Dr. MOBLEY used was those turbine blades Frosty mentioned.  He had a set of 3 that were all the same part. One was made with small equiaxed grains.  The next best had larger directionally solidified grains and the 3rd was a single crystal.

[ThomasPowers]

Well; large crystals are easy; just look at pegmatites!  wiki-"a large crystal of spodumene at the Etta Mine in South Dakota was 42 feet long, 5 feet in diameter and yielded 90 tons of spodumene"