Heat Treatment of Steel – Martensite Transformation

Steel is a remarkable material in that its microstructural characteristics can be altered by heat treatment, and consequently its properties and performance. Heat treatment would typically include i) heating of the steel (to an alloy dependent temperature), ii) soaking at that temperature, iii) cooling of the metal at some pre-determined cooling rate, and then iv) tempering the material at a specific temperature and for a given period of time to obtain the desired microstructure, and hence, properties. Besides altering the microstructure these heat treatments also affect the hardness, which is a function of the microstructure and hence carbon content.
The full understanding of the heat treatment of steels is a discipline in itself, and for full details you are referred to the textbooks and codes. It is appropriate here, however, to consider the behaviour, not only as the conventional steel phase diagram (of carbon composition – x axis, plotted against temperature – y axis), but also now to consider the time aspect. It is standard practice to plot transformations as a function of Isothermal Transformations, better known as TTT diagrams, (incorporating Time, Temperature and Transformation) for the particular composition. Such a plot for a 0.83%C eutectoid steel is shown in the figure, [Schlenker Fig 12.8 B] with temperature on the y axis, time (on a log scale) on the x axis and the transformation phases plotted as two large ‘C’ shaped curves, (open to the right).
The approach in using the TTT diagrams is to start in the austenitic region having heated the component being heat treated (assuming steel with a carbon content typically hypo-eutectic (in the range 0.05 to 0.8% carbon)) to above its austenite transformation temperature (which ranges 723ᵒC to 1000ᵒC dependent on its particular carbon content), and then cooling from this austenitic region at the cooling rate required to achieve the desired microstructure/properties. The transformation starts as the component cools through the first ‘C’ shaped curve and continues to the second ‘C’ shaped curve, at which the transformation ends. Thus in the figure, for this carbon content steel, Cooling Curve 1 represents a ‘full anneal’ from austenite to coarse pearlite, and takes about 45 minutes. Curve 2 represents a slightly faster cooling rate, typical of a ‘normalising’ processes, which in this case finishes at about 620ᵒC and takes about 9 minutes. If the rate of cooling is slightly faster, so the cooling curve passes through the ‘nose’ of the transformation line, Curve 3, some of the austenite will transform to form pearlite when the temperature passes the transformation curve. As the temperature drops below the martensite start temperature the remaining austenite will transform to martensite, a hard brittle microstructure that typically forms during rapid cooling. Such intermediate cooling would result in a final microstructure that consists of fine pearlite and martensite. For full hardening to occur, the rate of cooling should be such that the cooling curve passes to the left of the transformation curve, as illustrated by Curves 4 and 5. During these quenching processes the microstructure changes from austenite to martensite as the temperature drops below the martensite start line and ends as it passes the martensite finish line (which are around 260ᵒC and 110ᵒC in this case). This whole transformation only takes a fraction of a second. The martensite formed by these cooling processes would typically require tempering to reduce the hardness and improve toughness of the material.

Obviously the cooling rate of the material remote from the surface differs from that on the surface and the properties and microstructure of the component will change though its thickness. From a micro-structural point of view, there is a displacive transformation from face centred cubic (FCC) to body centred cubic (BCC) structures, as BCC lenses nucleate at FCC grain boundaries, almost instantaneously and grow across the grain. During the austenite to martensite phase change the BCC austenite changes to body centred tetragonal structure and there is a noticeable volume expansion. This volume expansion can introduce residual stresses of significant magnitude into the material, the magnitude and nature of which is dependent on where and when these changes occur relative to thermal contraction also taking place.
TTT curves change for different alloy compositions and tend to differ from relatively simple illustrative curve for the eutectoid composition shown in the figure. Depending on the chemical composition of the alloy, the martensite finish temperature can be below room temperature. In a particular case our client had changed the alloy used to fabricate gears and started experiencing cracking of the components soon after the change. What was surprising was that these gears did not show any evidence of cracking during initial post heat treatment inspection, but rather were found to have cracked after having been stored for a period of months and the cracking appeared to be more prevalent in winter. The investigation undertaken highlighted the martensite finish temperature of the new alloy was below ambient and the austenite to martensite phase transformation was not complete when the temperature of the heat treated gears reached ambient temperature. As the temperature of the gears dropped during the night and cold winter months some of the remaining austenite transformed to martensite introducing residual stresses which were of sufficient magnitude (and with associated low toughness) to lead to the cracking observed.
Understanding of the heat treatment processes should highlight that uncontrolled heating of alloy and heat treated components will obviously cause changes to the local microstructure achieved by heat treatment processes. These changes typically affect the mechanical properties and often introduce residual stresses which can cause premature failure. Failures related to the rapid self-quenching of material heated by friction in regions of mechanical damage on wire rope and in cases where high grade bolts are welded to stop self-loosening are encountered regularly.
The message it clear – heat treatment of steels is non-trivial and requires a good understanding of metallurgy, the chemical composition and required mechanical properties. Uncontrolled heating of heat treated and alloy steels should be approached with caution if premature failure or unexpected changes of mechanical properties are to be avoided. This includes small tack welds or arc strikes which are often perceived as being innoxious!!