Karin Frisk

An austenitic alumina forming stainless steel was exposed in a nitriding atmosphere comprising 5 vol% H₂/N₂ for 100 and 1000 h. After 100 h, the sample displayed aluminium nitrides down to 450 μm and chromium nitrides down to 200 μm. Thermodynamic equilibrium calculations and kinetic modelling of the system to simulate 100 h of exposure time were performed using the software ThermoCalc 4.1 including DICTRA. The order of appearance of nitrides in the sample starting from the surface can be understood from the thermodynamic equilibria as well as from results of kinetic modelling. However, the nitridation depth predicted by the kinetic modelling, was larger than experimentally observed, and the cubic chromium nitride predicted to be present on the surface was not detected in the exposed sample. One difference between model and experiment was a thin oxide layer present on the surface of the sample. As this oxide layer can decrease the inwards flux of nitrogen, attempts were made to describe the effects of this layer in the model. The model was also used for qualitative comparison of varying alloy compositions, temperatures and initial nitrogen contents of the matrix.

The addition of alloying elements in low alloyed PM steels in the form of a master alloy gives the advantage of introducing oxidation sensitive but less expensive elements and also allows manipulation in composition adjustment to achieve desired properties. In this work, interrupted sintering trials of the Fe–2MA–0.5C (%) (MA = Cu based master alloy) are performed. The behaviour of the liquid forming master alloy, for instance in terms of liquid phase formation, alloying element redistribution and effect on the dimensional changes, is investigated. The results show that master alloy particles melt over a range of temperature, which is also supported by the thermodynamic calculations. The low swelling in the master alloy system, compared to a reference system of Fe–2Cu–0.5C, is attributed to the progressive melting of the master alloy. The mean diffusion distance of Cu in Fe at the interparticle boundaries is 5.8 μm after 34 min of isothermal holding.

The effective use of alloying elements in powder metallurgical steels requires a deep understanding of their redistribution kinetics during sintering. In this work, interrupted sintering trials of Fe–2Cu and Fe–2Cu–0·5C compacts were performed. Moreover, diffusion simulations of Cu in γ-Fe using Dictra were performed. It is found that transient liquid phase penetrates the Fe interparticle and grain boundaries in less than 3 min of holding time. However, C addition limits the penetration of liquid Cu, particularly into grain boundaries of large Fe particles. The results also show that the mean diffusion distance of Cu in γ-Fe in the C added system is slightly lower than that in the C-free system at 3 min of holding time; however, after 33 min, the mean diffusion distance is similar in both systems. The diffusion distances of Cu in γ-Fe, predicted by Dictra, are in good agreement with the measured values.

A comparison of the coarsening of nitrogen-rich vanadium precipitates and the coarsening of carbon-rich vanadium precipitates is presented. The precipitate phases are studied experimentally, via fabrication of model alloys, and theoretically, via simulations utilizing the DICTRA software. The experimental investigations indicate that the nitrogen-rich precipitates exhibit a slower coarsening behaviour than the carbon-rich precipitates. Analysis using thermodynamic and kinetic modelling shows that this can be explained by the higher thermodynamic stability of the nitrogen-rich precipitate compared to the carbon-rich precipitate. The calculated coarsening rates are compared with the measured rates, and found to be in satisfactory agreement using reasonable values for the interfacial energies. The investigations are motivated by the fine precipitate size distribution of nitrides and carbonitrides characteristic for high nitrogen alloyed tool steels produced by means of powder metallurgy.