The composition of nano-sized alloy carbides formed by interphase precipitation in V–Nb and V–Ti multiple microalloyed low-carbon steels is analyzed by using three-dimensional atom probe. Carbide-forming alloying elements including V, Nb, and Ti, are simultaneously precipitated from the early stage of isothermal treatment, whose atoms are uniformly distributed in the carbide particles, even after prolonged holding. Cluster analysis by the maximum separation method, with parameters optimized using different methods, is carried out to extract alloy carbides from matrix. The composition of alloy carbides evaluated by site fraction of substitutional carbide-forming alloying elements indicates that at the early stage of their formation, Nb and Ti are more strongly enriched than V.

Atom probe is a powerful technique for studying the composition of nano-precipitates, but their morphology within the reconstructed data is distorted due to the so-called local magnification effect. A new technique has been developed to mitigate this limitation by characterizing the distribution of the surrounding matrix atoms, rather than those contained within the nano-precipitates themselves. A comprehensive chemical analysis enables further information on size and chemistry to be obtained. The method enables new insight into the morphology and chemistry of niobium carbonitride nano-precipitates within ferrite for a series of Nb-microalloyed ultra-thin cast strip steels. The results are supported by complementary high-resolution transmission electron microscopy.

Local magnification effects and trajectory overlaps related to the presence of a second phase (clusters) are key problems and still open issues in the assessment of quantitative composition data in three-dimensional atom probe tomography (APT) particularly for tiny solute-enriched clusters. A model based on the distribution of distance of first nearest neighbor atoms has been developed to exhibit the variations in the apparent atomic density in reconstructed volumes and to correct compositions that are biased by local magnification effects. This model was applied to both simulated APT reconstructions and real experimental data and shows an excellent agreement with the expected composition of clusters.