by Don Klesser

This article is a follow-up to the Beaver Creek event in August, where I provided an overview of the metallurgy of hardening steel through quenching.

Hardening requires two steps: heating steel to austenite, and then quenching to obtain martensite. Steel is a combination of iron and a small amount of carbon. At low temperatures (below 1300º F.), in the unhardened state, the carbon and iron atoms form two phases, ferrite and cementite (together they are called pearlite). As steel is heated above 1300º, the iron atoms begin to form austenite, which has the ability to dissolve carbon atoms – like sugar dissolving in iced tea. All steels must transform 100% to austenite before they can be hardened. It is widely, but falsely, believed that austenite has formed when steel loses its’ magnetism; that is, when a magnet will no longer stick. Steel begins to form austenite around 1333º, and loses its magnetism around 1430º. However, many steels are not 100% transformed to austenite until they reach 1500º or even 1600º.  A reliable, but low cost, method to ensure proper heating temperature is a Tempil® stick. The stick will melt above its ‘rated temperature’ – and for most steels, 1650º minimum is a good starting point for hardening.

 Once austenite is formed, it can be quenched to form martensite. Under normal cooling conditions, the iron and carbon atoms transform back to ferrite and cementite as the temperature falls below 1333º.  When steel is cooled rapidly however, carbon atoms don’t have time to move out of the austenite solution, and another steel phase is formed called martensite. Martensite is hard and brittle because the carbon atoms act to strengthen its structure. (By adding a heat to martensite, the carbon atoms gain energy, begin to move out of the structure, and form ferrite and cementite again.  This softening process is called tempering.)

The phases of steel are summarized below:

Steel condition

Low & medium carbon steels

High carbon steels (>0.60% C)
At 70º (unhardened)   ferrite & pearlite, magnetic   pearlite, magnetic 
Heated to 1300º   ferrite & pearlite, magnetic   pearlite, magnetic 
Heated to 1350º   ferrite & pearlite, austenite, magnetic

pearlite & austenite, magnetic

Heated to 1450º ferrite & pearlite, austenite, non-magnetic  austenite, non-magnetic
Heated to 1650º  austenite, non-magnetic  austenite, non-magnetic
Slow cooled to 70º  ferrite & pearlite, magnetic pearlite, magnetic
Quenched to 70º  martensite, magnetic  martensite, magnetic 
Semi-quenched to 70º ferrite, pearlite & martensite, magnetic  pearlite & martensite, magnetic 

 The maximum hardness of quenched steel is fully dependent on how much carbon is in the steel; higher carbon enables higher hardness.  Hardenability, on the other hand, depends on both carbon content and alloy content.  Two steels having the same carbon content, but different alloy content will quench to the same maximum hardness.  However, the steel with the higher alloy content (chromium, nickel, molybdenum, vanadium, etc.) has better hardenability, or ‘ability to be hardened’.  Alloy steels are more hardenable because the alloying atoms slow the movement of the carbon atoms out of austenite during quenching. If sufficient alloy is in the steel, the movement of carbon atoms is so slow that it can be quenched in oil, or even air – these are oil hardening or air hardening steels. 

Quenching involves rapid removal of heat. During quenching, only the steel surface is being rapidly cooled as it is in contact with the quenchant – water, or oil, or air. Inside the steel, heat has to move to the surface so that it can be removed in contact with the quenchant.  Because this process takes time, the interior of the steel cools more slowly than the surface.  The bigger the part, the slower the interior cools.  As a result, martensite may form at the surface of the steel, but the inside cools too slowly, and only ferrite and cementite result. Alloy steels are better able to harden into the center because of the greater hardenability.

Low alloy and carbon steels require a very fast quench to remove heat quickly, such as water.  As the water contacts the hot steel surface, it is heated and forms steam. The formation of steam draws a lot of heat from the steel. Unfortunately, once the steam is formed, it also forms a vapor barrier around the steel and slows the quench process. To obtain an effective quench, the blacksmith must stir the part during quenching to break the vapor barrier. Violent agitation, obtained by adding water pressure, is even better. If a water supply is available, a simple way to add pressure is to use a water hose or faucet with a pressure nozzle.

There is much interest in the effectiveness of ‘superquench’ (a solution of water, salt, soap and wetting agents) for improving the hardness of steel during quenching. It is well known that salt does somewhat improve the quenching ability of water (brine), and soap significantly decreases the quenching ability. In order to compare superquench with other available quenches, we heat treated some test pieces during the Beaver Creek demonstration. Later these test pieces were hardness tested with a Rockwell hardness tester.

Two sets of samples, made from a plain carbon (approximately 0.20% carbon) and a low alloy (grade 4140, with 0.40% carbon) steel, were heated in a coal forge.  Prior to quenching, the temperature of the bar samples were measured using a 1650º Tempil stick to ensure austenite was formed.  The parts were randomly quenched in one of three solutions: water with salt (brine), ice water & salt, or superquench – two sets in each quench.  The samples were cut apart, ground slightly with a water cooled knife sharpener (to remove any decarburization), and tested with a calibrated hardness tester in three locations – for a total of 72 hardness readings.

 The average hardness results are summarized below:

Steel Type


Outside Hardness* Inside Hardness*
Plain Carbon water & salt 37.5 Rc 37.2 Rc
Plain Carbon ice water & salt 42.0 Rc 36.8 Rc
Plain Carbon superquench

18.2 Rc

19.7 Rc

4140 alloy water & salt 56.2 Rc 52.7 Rc
4140 alloy  ice water & salt 53.2 Rc 50.3 Rc
4140 alloy  superquench 53.7 Rc 52.7 Rc

 *Outside is the surface touching quenching solution.  The inside hardness is about 1/4” deep on the plain carbon steel, and 3/8” deep on the 4140 steel.

 From metallurgy literature, a 0.20% carbon steel (such as hot rolled bar stock), can have  martensite hardness of about 44 to 46 Rc.  The ice water was able to get an average hardness of 42.0 Rc (note one of the 6 readings on the two samples was near the maximum of 46 Rc) on the outside surface – this was the best result for the plain carbon steels.  The inside surface was about 37 Rc – this is due to marginal hardenability of the plain carbon steel.  The superquench did not harden the plain carbon steel at all, having an average hardness of about 18 to 19 Rc (the hardness of ferrite and pearlite). 

The expected hardness of the 4140 steel is about 56 to 62 Rc (the variation is due to the allowable tolerance on carbon content).  All of the quenches were capable of quenching the outside surface of 4140 to near the maximum expected martensite hardness.  Inside hardness dropped only a little, indicating that 4140 steel has good hardenability for this size part. From some blacksmithing articles that I have read concerning superquench, it is not recommended for use on alloy steels because of its severity, however, in this test, the results were similar to that of plain water with salt.

Top of Page