Modeling Autotempering Effects on Microstructure and Properties in Large Section Sizes

Tuesday, September 13, 2022
Exhibit Halls F – HTS Poster Area (Ernest N. Morial Convention Center)
Mr. Michael Rupinen Jr , Colorado School of Mines, Golden, CO
Dr. E. Buddy Damm , TimkenSteel Corporation, Canton, OH
Mr. Michael Jacobs , Colorado School of Mines, Golden, CO
Dr. Amy J. Clarke , Colorado School of Mines, Golden, CO
Dr. John G. Speer , Colorado School of Mines, Golden, CO
Martensitic steels are often tempered at low temperature (below 200 °C) to achieve favorable impact energy and yield strength combinations. The recent development of AF9628 steel which achieves these properties while having significant hardenability and reduced alloy cost relative to previous steels has generated renewed interest in understanding the microstructural features controlling the strength/toughness combination in similar carbon content (~0.28 wt pct) alloys. One important consideration that is sometimes overlooked is that the ability to have a fully martensitic microstructure while having relatively low alloying additions typically leads to a high martensite start temperature. A high martensite start temperature along with large section sizes that experience a wide range of cooling rates during quenching means that autotempering (i.e. tempering of the martensite during cooling) may have a significant effect on the microstructure and properties. While the occurrence of autotempering is often mentioned in literature, the description of autotempering is usually limited to a characterization of carbides formed in the as-quenched condition. Little work has been done to attempt to predict the degree to which autotempering may affect the microstructure of a given steel. A model discussed here has been developed to quantify the tempering experienced by martensite formed over a range of temperatures during quenching (i.e. autotempering) and to relate that degree of autotempering to classical isothermal tempering stages. The results show significant differences in the predicted degree of autotempering experienced between different cooling rates for a given alloy and between alloys having different martensite start temperatures. Microhardness and SEM micrographs support the initial predictions from the model.