Steel Phase Basics

What follows is based on what I’ve learned from books and talks by Wayne Goddard, the ABS instructors in Old Washington, and others, plus the postings of metallurgists like Kevin Cashen, and  John Verhoeven’s public PDF document and his book Steel Metallurgy for the Non-Metallurgist.

The following is generalized for simple carbon steels. When a steel has other alloying elements this will change the temp and speed at which that steel will change phase. Every steel is a little different and has it’s own set of charts and specifications.

You probably know that steel is iron (Fe) with just a bit of carbon (C). Simple “high carbon” steels have less than 1.5% C and trace amounts of impurities. If you get more than 2% C you are getting into the realm of cast iron. The Fe forms a 3 dimensional lattice structure. A grain of steel is a tiny area where this Fe lattice is all oriented in a single direction. Each grain’s lattice is oriented or offset differently than its neighbors. Grain boundaries tend to be the weakest point in a piece of steel, with smaller grain size making for a tougher steel.

The eutectoid point for how much C you have in the steel is 0.77% C. This is the amount of C that fully occupies the lattice in the austenite phase of steel (more on austenite below). More than 0.77% C in the steel is referred to as hypereutectoid, less is referred to as hypoeutectoid. In numbered steels the last digits refer to the 100th of a percent of C. The reference is not exact – so 5160 the “60” refers to 0.60% C (the accepted specs range from 0.56-0.64% C) and is just slightly hypoeutectoid… while 1095 (0.90-1.03% C) is hypereutectoid.

Ferrite is the phase steel wants to be in at room temperature. It is like a cubic crystal lattice with Fe atoms at the cube corners. This Fe lattice can hold one C atom in the center of each cube – this structure is called Body Centered Cubic (BCC) – or “alpha iron”. Only 0.02% C can be held in the ferrite lattice.

Austenite is the phase of steel (created by heating to somewhere over 1340°F) that also has a cubic lattice structure, but in this phase C atoms can settle into the Fe lattice at each face of the cube – called Face Centered Cubic (FCC) – or “gamma iron.” This allows the Fe lattice to suck in much more C. As mentioned above, 0.77% C can be held in the austenite lattice, and when you include other metallurgical magic, I’ve read that austenite can hold 80 times the C that ferrite can hold.

Cementite and other carbides are where that extra C wants to live when it can’t get into the Fe lattice. Cementite (Iron Carbide – with the chemical formula Fe3C meaning 3 iron and 1 carbon atom are bound together) is harder than the regular steel lattice but softer than other carbides we find in alloyed steels.

The temperature at which steel becomes non-magnetic (Curie temp) is 1414°F regardless of the %C in the steel and has something arcane to do with Fe atom’s electron’s angular momentum and spin. Whatever the heck that means!

A phase diagram plots % C in the steel across the bottom of the graph, and temperature vertically. Areas of the graph show the phase of steel at that temperature for that % C. Here is such a diagram from Verhoeven’s public domain PDF:

The theoretical temp for transforming ferrite into austenite is marked on the phase diagram by the A3 line for hypoeutectoid steels and the Acm line for hypereutectoid steels. This theoretical austenitizing temp is lowest at the eutectoid point (1340°F at 0.77% C). The theoretical austenitizing temp varies according to %C. As you can see, both above and below 0.77% C the austenitizing temp increases.

According to Verhoeven, each time steel changes phase into or out of austenite the new phase starts growing grains at the boundaries of the old grains of the old steel phase. The grains of the new phase eat away the grains of the old phase until the steel is more-or-less all transformed into the new phase. This is the mechanism for refining grain size through thermal cycling. Unless of course you go way over the austenitizing temp – in which case the new austenite grains start to merge with each other and form bigger grains than you started with.

The transformation into austenite is called the austenitizing temp, and often referred to as the “critical” temp. As you can see from the chart above, even in a simple steel this varies depending on the %C of the steel.

The recommended real-life austenitizing temps are higher than the A3 and Acm lines. As I understand it, this is to fully transform ferrite into austenite and to allow for existing carbides to dissolve and free up C to be absorbed into the Fe lattice and/or to reform as new carbides. For instance, a recommended temp for austenitizing 5160 is 1525°F even though the diagram’s A3 line at 0.60%C is about 1380°F.

A curious thing about simple steels is that – for some obscure reason at the atomic level – they go non-magnetic at 1414°F – regardless of %C. So when folks say “bring the steel up a little above non-magnetic” to get to critical temp for heat treat – that’s just a rule of thumb. If you have a controlled forge or oven with a temperature readout it pays to look up the recommended austenitizing temps for various heat treating steps like normalizing, annealing, and hardening… as well as the recommended tempering temps.

Normalizing (where the steel is soaked at above the austenitizing temp and then cooled in still air) is used to relieve stresses in the steel created from the heat and hammering of the forging process. It also dissolves large carbides so that the C becomes redistributed more evenly in the steel. Grain size is also evened out by the thermal cycling. Thermal cycling just means bringing steel above the critical temp into austenite, then back down through critical temp again – regardless how fast you do it. Normalizing evens out the steel’s internal stress, grain sizes, carbide distribution, etc.

Annealing: If austenite is cooled very slowly, the C atoms have time to migrate out of the lattice and bind into carbides. In plain steels the C will form pure cementite plates between layers of pure ferrite. This combination is called pearlite, and (other than another form that has spheroidal cementite) pearlite is the easiest phase of steel to grind, bend, or machine. When we anneal steel we are creating pearlite.

Hardening: If austenite is cooled very rapidly (quenched) the extra C atoms which austenite can hold do not have time to get out of the lattice. This causes a stressed lattice structure called martensite. The lattice is literally stretched into a Body Centered Tetragonal (BCT) form to accommodate the C that could not diffuse out in time. When quenching for martensite, each steel has a “Martensite Start” (Ms) temp and a lower “Martensite Finish” (Mf) temp.

Martensite starts forming at Ms, but you need to reach Mf within the time given in the steel’s Time Temperature Transformation or TTT diagram (see below) in order to get full martensite with minimal retained austenite grains. Martensite is both hard and brittle due to the stressed BCT lattice. It should also be noted that martensite has a slightly lower density than other room temperature steel phases – causing a slight expansion of the blade where it is created – an extra 4-5% volume at room temperature. This is what gives the Japanese Katana it’s iconic curved shape – it has the expanded martensite at the cutting edge and tougher phases of steel at the core and spine.

With some quenching methods – particularly for “shallow hardening” steels – the thin edge of the blade will be fully transformed to martensite while the thicker portions will become pearlite. You can see these phases of steel in the knife as a “hamon” or “temper line“. “Temper line” is something of a misnomer since this is caused by hardening the steel, not tempering it.

A critical temperature is where a phase transformation takes place. On heating this is where austenite forms. On cooling things get more “interesting.” The Time Temperature Transformation or TTT diagram (aka Isothermal Transformation or IT diagram) for a particular steel shows graphically the critical temperature lines where transition from austenite into other phases of steel (pearlite, bainite, or martensite) takes place. The nested curves indicate the start and finish of the phase transformation from austenite. Both time and temp are critical in these phase changes.

If steel is cooled so that time & temp trace a line through the Ps/Pf (pearlite start and finish) lines above the “nose” of the graph the result is pearlite. If the steel is cooled fast enough to miss the nose and goes through the Ms (martensite start) line at the bottom of the graph you transform austenite into martensite. If the steel is cooled fast enough to miss the nose, but then held at temp to go through the Bs/Bf lines you get bainite.

Here is Verhoeven’s TTT chart for 1080 – the Mf line (where the transformation into martinsite is complete) is not shown, but for 1080 it would be a horizontal line at 212°F. Note that the time scale is logarithmic.

Bainite is formed by initially cooling austenite very fast – to miss the nose of the steel’s TTT diagram, but stopping and holding the temp well above Ms for long enough to go through the TTT diagram’s Bs/Bf lines at a constant temp before finishing the quench. This is usually accomplished by the use of a molten salt bath to achieve the fast initial quench and then hold at that temp for the time required to reach the Bf line. Bainite is similar to pearlite in that it is a mix of ferrite and cementite – but in bainite, the cementite forms in smaller filaments and loose particles.

Actually there are two forms of bainite – upper and lower – with lower bainite having finer cementite structures. Lower bainite can be almost as hard as martensite and can be tougher than tempered martensite with the same Rockwell (Rc) hardness.

Retained austenite refers to austenite grains that do not transform (to martensite or pearlite or bainite) on cooling. If you quenched to a temp between Ms and Mf you would only transform a portion of the austenite. The rest is retained austenite. Even full quenching through Mf can leave some austenite grains in the steel. Retained austenite is unstable in the long run at room temp and is living on borrowed time. When it does transform it forms untempered martensite. The gradual creation of untempered martensite adds stresses to the steel over time. This might explain stories of an untempered blade “just cracking for no reason” if left on the workbench.

Retained austenite can be forced to transform into martensite by a low-temperature quench and to untempered martensite or ferrite/pearlite during tempering at temperatures above 375°F. The creation of untempered martensite explains why you want to temper multiple times (to temper the new martensite).

Tempering is primarily used to transform untempered martensite into tempered martensite. When martensite is heated to a few hundred degrees (exact temp varies depending on the steel and the desired effect), some of the C atoms trapped in the martensite will migrate out of the lattice, forming carbides and leaving the martensite in a less stressed state. This adds toughness to the blade, making it less brittle. Tempering to a recommended temperature causes very little loss of hardness.

It is general practice to temper three times to get the steel’s transformations settled out.

When steel changes from one phase to another (ferrite to austenite, austenite to pearlite, etc.) seed grains for the new phase generally start along existing grain boundaries and grow from there. This is why thermal cycling between steel phases tends to reduce grain size. This occurs during phase change both upon heating and on cooling. If the steel is not overheated then these new grains will be smaller than the old grains. Overheating causes some grains to consume their neighbors, forming larger grain sizes. The larger the grain size, the easier it is for a local stress in the steel to propagate into a crack in the blade.

And there you have the 10 cent tour of my current understanding of heat treatment and steel phases.