52100 Wootz Steel by Kevin Gray

"52100 Wootz Steel" by Kevin Gray

Introduction

The aim of the sword project was to reproduce the microstructure found in Ed Fowler's knives by following forging and heat treating procedures learned at Fowler's weeklong bladesmithing seminar. Ed Fowler uses 52100 steel for his knives, the same steel was used for this project. The forging techniques and heat treating practices produce knives with the ability to cut a half-inch thick hemp rope hundreds of times without sharpening and be bent to a 90 degree angle then back 180 degrees multiple times without cracking. The extraordinary edge retention and ductility properties produced can be explained by examining the microstructure and understanding the purpose of every heat treatment step. To track the microstructure, samples of the steel were taken throughout the forging process and after every step of heat treating. The microscopy samples were then examined to determine the impact on the microstructure that each step has. Using information gathered by the microscopy samples possible improvements will be proposed and later tested.

Theory

Figure 1. Iron Iron-Carbide phase diagram with adjusted lines for 52100.

As shown in Figure 1. the eutectoid composition for the phase diagram adjusted for the alloying elements in 52100 (dotted line) is around 0.45 Wt. % carbon. At the 1.0 Wt. % carbon composition Figure 1. shows that between 755⁰C and 910⁰C two phases, Austenite and Cementite, are present. 52100 austenitized to a temperature in the two phase region would contain Cementite and Austenite with dissolved carbon. The amount of carbon dissolved in the Austenite increases as austenitizing temperature and time increases, for example; on Figure 1. the horizontal dotted line at 843⁰C represents an austenitizing temperature and the white dot on the dotted line represents the composition of 52100. The grey dot on the dotted line represents the solubility limit of carbon in the Austenite at that temperature, for this example between 0.75 and 0.8 Wt. % carbon. Figure 1. was used to validate max forging temperature and in conjuncture with Figure 2. determine austenization times and temperatures.

Figure 2. carbon Wt. % in Austenite vs. austenitization time in minutes. (Austenitizing temperature 1550⁰F)

As seen in Figure 2. steels containing Chromium take much longer to reach the equilibrium amount of carbon dissolved in Austenite. Chromium carbides are more stable than iron carbides and therefore take longer to dissolve. Also, notice in Figure 2. that the stable amount of dissolved carbon is much lower for 52100 than for 10100 (a plain carbon steel with 1 Wt. % carbon). Examining Figure 1. explains the disparity. 10100 at 1550⁰F is just above A_cm according to the Iron Iron-Carbide phase diagram meaning that all of the carbon will be in solution with austenite. 52100 is well below the A_cm on the Iron Iron-Carbide phase diagram corrected for the Chromium in 52100 meaning that the stable amount of carbon in austenite for 52100 at 1550⁰F is close to 0.775 Wt. % this correlates directly to the steady state portion of the 52100 line in Figure 2.

Figure 3.. Iso Thermal diagram for 52100 steel with cementite present.

The cooling rate required to quench 52100 from a two phase austenitization to produce cementite in a martensite matrix was determined using an IT diagram for 52100 seen in Figure 3. It is important to note that the Austenite grain size greatly impacts hardenabiltiy. The Austenite grain size rating in Figure 3. is an ASTM grain size rating of nine. As the size of the grains decreases so does hardenabiltiy due to more grain boundary area or more nucleation sites for pearlite. More nucleation sites promotes pearlite growth decreasing the time to miss the nose of the curve, in Figure 3., and form all martensite. Figure 3. was used to validate the use of Martempering quenching oil.

Table 1. Martempering effect on impact toughness.

Martempering is the method for quenching that cools fast enough to miss the nose of the curve in Figure 3. but cools slowly through the martensite start temperature to room temperature. Cooling slowly during the Austenite to martensite transformation prevents adding additional thermal stresses to an already dramatic change in the microstructure. By minimizing the additional stresses to the microstructure during martensite formation an improvement in impact toughness is produced as seen in Table 1.

Figure 4. Comparison of martensite in sample A ASTM grain size 11 and sample B ASTM grain size 15.

Martensite forms at grain boundaries and other high energy areas within the steel and grows in needle like structures through the austenite until arresting by colliding with other martensite laths. A small grain size steel would have much more grain boundary area than a large grained steel and therefore would have more nucleation sites for martensite. Martensite laths formed in steel with more nucleation sites would collide with one another in a shorter distance creating smaller laths and a finer martensite structure as seen in Figure 4.

Figure 5. Martensite start and finish temperatures vs. carbon weight percent.

After quenching, retained Austenite remains in the steel. As shown in Figure 5. by cooling below room temperature more retained Austenite transforms to martensite. Figure 5. was used to confirm the effectiveness of the sub room temperature post quenching heat treating steps. Martensite despite being metastable is more stable than Austenite at room temp in 52100 steel. Cooling below room temperature provides the energy to transform the less stable Austenite to martensite.

Figure 6. HRC hardness of fresh Martensite vs. Wt. % carbon.

Figure 6. was used to confirm that the amount of carbon in the martensite matched the amount that was intended and predicted by the austenization times and temperatures.

Figure 7. Impact toughness vs. martensite tempering temperature.

As can be seen in Figure 7. TME or tempered martensite embrittlement is the name given to the tempering temperature range where a loss of impact toughness occurs. The loss in impact toughness is due the segregation of phosphorus to the austenite grain boundaries during austenization. The tempering temperature used on the sword was verified by use of Figure 7.

Experimental

The starting material was a four inch diameter cylinder six inches long of vacuum arc remelted 52100 steel. A sample was taken from the billet in the as received condition. 52100 is shipped in the Spheroidized condition with all of the carbon tied up in large carbide particles surrounded by a ferrite matrix. A pre-forging heat treatment was used to produce pearlite grains with proeutectoid cementite at the grain boundaries. For the pre-forging heat treatment the billet was wrapped in stainless steel foil and packed with "Carbrite" coke, to prevent decarburization, and placed in an induction furnace set to 1700⁰F and allowed to soak for two hours. As shown in Figure 1. 1700⁰F is just above the A_cm line for 52100. Soaking at 1700⁰F ensures all cementite particles (Spheroidite) are dissolved in the Austenite and upon cooling the microstructure will be pearlite grains with proeutectoid cementite. A sample was taken after the pre-forging heat treatment to verify that the desired microstructure was obtained. All micrographs of samples can be seen in the appendix Micrograph Timeline.

Figure 8. Propane forge

Figure 9. Trip hammer

A propane forge seen in Figure 8. was used to heat the billet for forging and a trip hammer Figure 9. was used for drawing out the steel. A max forging temperature of 1620⁰F was maintained by the use of a radiation pyrometer. The max forging temperature was kept under the Acm for 52100 seen in Figure 1. to prevent the proeutectoid cementite from dissolving into the austenite. Throughout the forging process the cementite formed at the austenite grain boundaries in the initial heat treatment was transformed into homogenized Spheroidite particles. Samples were taken at four different stages of forging at various billet cross sections detailed in the appendix Micrograph Timeline.

Figure 10. Sword and heat treatment sample.

When forging was completed a machining mill was used to shape the handle with the help of Jay Fisher. The blade was shaped using a belt sander. Prior to heat treating the edge of the sword was only brought down to slightly less than an eighth of an inch. The edge was left thick to prevent too sever of a quench and so any decarberization that took place during heat treating could be removed during final shaping. The sword was then ready for heat treatment. A separate piece of the billet was forged into a thin strip (bottom Figure 10.) to follow the sword through the same heat treating steps. After each step of the heat treatment a one inch sample was cut from the end of the thin strip and was examined using optical microscopy the results of which are contained in the appendix Micrograph Timeline.

Figure 11. Liquid salt tank

A liquid salt tank, Figure 11., was constructed to normalize and austenitizing the sword. The liquid salt tank is a propane forge with a stainless steel tube in the center to contain the salt mixture of 50% potassium chloride and 50% soda ash.

Ed Fowler uses an oxy-accetalyen torch to heat his knives before quenching. Heating a sword with a torch would have been impractical and would have lead to uneven heating and too long of austenitization temperatures for some parts of the sword and too short of time for other areas. A salt tank was chosen to heat the sword for the rapid rate of heating, evenness of heating and control of temperature.

Figure 12. Sword prior to first normalizing step in liquid salt tank.

The sword was normalized three times by heating to 1600⁰F and holding for four minuets then air cooling. Normalizing both homogenizes carbon and refines the grain structure. On heating to Austenite new Austenite grains form at Ferrite/Pearlite grain boundaries and then grow until all of Pearlite and Ferrite is consumed. During air cooling, Ferrite grains form at the Austenite grain boundaries and Pearlite forms in the remaining Austenite. Given time at elevated temperatures the larger Austenite grains will grow at the expense of small grains. If austenite grains are allowed only to form then the sword is cooled small grains result. If the normalizing process is repeated more grain boundary area translates to more nucleation sites for new grains to form. Each successive step of normalizing forms smaller grains. Performing the normalizing in a liquid salt bath produces finer grains than normalizing in an air forge would because; one, the sword is brought up to temperature faster and two, less time is needed to equalize the temperature throughout the sword. Heating faster only allows for small grains to nucleate and staying at the austenitizing temperature for a shorter period of time reduces grain growth. Liquid salt also prevents uneven heating due to the salt within the bath naturally creating a convection current that continuously stirs the salt eliminating any hot or cold areas.

Figure 13. Preheating quenching oil in quench tank

A quench tank was constructed with an adjustable height table to allow the sword to only be quenched to a certain depth to only harden the edge and leave the spine more ductile. A Martempering oil was used to quench the sword and was preheated to 160⁰F using an oxy-acetylene torch as seen in Figure 13.

Figure 14. Oil quenching the sword

The sword was austenitized at 1600⁰F for four minutes and edge quenched in 160F oil then allowed to come to room temp before being placed in a freezer and left overnight. The process was then repeated with an austenitizing temperature of 1550⁰F and again at 1500⁰F. The sub cooling step after the first and second quench are necessary to produce a small and uniform grain structure. Sub cooling transforms the retained austenite to martensite if retained austenite remained in the sword the areas of retained austenite would act as seed grains. Areas of retained austenite would allow for early formation of austenite grains at these sites. The austenite grains that formed first would grow at the expense of smaller grains resulting in an irregular grain size.

The benefit of multiple quenching comes from two things; sufficient carbon dissolved in the Austenite and reduced grain size of Austenite prior to the final quench.

To better understand the benefits of multiple quenching examine Figure 2. As can be seen in Figure 2., to dissolve 0.6 wt% carbon into the Austenite the sword would have to be held at 1550⁰F for about 12 minutes. However, if the sword were allowed to soak at 1550⁰F for 12 minutes the very fine grain structure achieved by the multiple normalizing steps would be detrimentally affected by grain growth, the end result being a coarser martensite structure (see Figure 4). To accomplish both the fine grain structure and achieve a sufficient amount of carbon required in the Austenite to produce the desired hardness, the iron and chrome carbides must be dissolved incrementally. The first austenitization at 1600⁰F for four minutes would dissolve approximately 0.5 (more than Figure 2. would indicate because of the higher temperature). Quenching prevents any iron or chrome carbides from precipitating making the 0.5 wt% carbon in the Martensite much easier to dissolve into the Austenite upon heating for the second Austenitizing step. In this way multiple austenitizing has a cumulative effect on the carbon dissolved in the Austenite prior to the final quenching.

Grain growth is avoided because each austenitizing step only lasts four minutes and new grains are forced to grow for each of the three austenitizing steps. The reason for the descending austenitization temperatures is to dissolve most of the carbon needed in the first step then in the second and third steps dissolve the remaining amount while limiting grain growth. Quenching from a lower austenitizing temperature for the final quench results in lower thermal stresses in the steel producing a tougher martensite.

Results

The sword microstructure matched that of the knives I made during a seminar with Ed Fowler. The only differences were; no banding in the sword, slightly larger Spheroidite particles in the sword and more segregation of the small carbide particles in the sword. All micrograph comparisons are in the Micrograph Timeline Appendix. Rockwell C hardness of 63 was taken from the final temper sample. Referencing Figure 6 it can be seen that the martensite contains 0.55 to 0.6 Wt. % carbon.

Discussion

The uneven carbon distribution in the sword steel was caused by Ostwald ripening depleting the carbon percent of the matrix surrounding the Spheroidite the carbon percent based off of micrographs dropped below eutectoid levels producing ferrite grains and pearlite grains with large Spheroidite particles. The pearlite lamellar structure broke down into very small cementite particles only in the prior pearlite grain regions the normalizing cycling helped to even out the carbon distribution however some unevenness was visible on the final normalize micrograph. A possible solution for the Ostwald ripening problem would be a bigger trip hammer so the deformation can take place faster and not allow time for as much cementite coarsening. A normalizing process for future testing to achive smaller grains and evenness of normalizing temperatures throughout the cross section of the blade consists of the following steps. Heat in liquid salt to normalizing temperature then quench in liquid salt at 1100⁰F until equalized then allow to air cool to room temperature. this would increase the nucleation rate of pearlite grains and reduce time for graingrowth resulting in a finer grain and finer lamellar structure this would be good for two reasons one smaller grains leads to a lower stress martensite and finer lamellar structure means more ferrite to cementite surface area so carbon would dissolve into Austenite faster and homogenize more rapidly during austenization. Other experiments to take place in the future are the influence that banding has on the performance of the knives banding was not produced in the sword because Vacuum arc remelted steel was used "Because of the high arc temperature and the small pool of liquid metal, the process can produce very sound ingots of dense crystal structure, low hydrogen and oxygen contents, and with minimal chemical and non metallic segregation" The Making Shaping and Treating of Steel. An accurate comparison between banded and non-banded microstructures will be difficult to make because of what causes the microstructure to be banded has other affects as well. The theory of why banding would promote longer edge life makes sense with hard and soft layers preferential wearing would create a micro-serrated edge that would continue to cut and if this is the case band thickness could possibly be optimized to yield more cuts.

Conclusions

The microstructure found in knives forged during a bladesmithing seminar was replicated in the sword project. The properties of hardness, edge retention and ductility were explained with the use of the micrographs and well known metallurgical figures. Micrographs were taken at four points during the forging and after every step of the heat treating process. All micrographs are cataloged in the Micrograph Timeline Appendix. Possible improvements to test at a later time were proposed.