This project was ordered by Sandvik Materials Technology and was performed by a group of students at Uppsala university. The purpose of the project was to study precipitation behavior and structure stability in six advanced heat resistant alloys. Each sample were subjected to a creep rupture test in 600 or 700°C depending on the alloy type. Two parts of each alloy where examined; one part which had been affected by creep and another part which was unaffected by creep. A literature study was performed first to gain knowledge of the scientific theory utilized in this project, namely creep, precipitation hardening, and about the different materials which were analyzed. Preliminary results for the phase composition of the materials were obtained from a Thermo-Calc (TC) simulation. The SEM-images showed nothing noteworthy for any sample due to the roughness of the sample surfaces. The EDS-analysis showed chromium depletion in the centers of the aged samples of HT9 and Sanicro® 75X. Other minority phases such as Cr23C6 in Sanicro®70, P-phase and a titanium nitride phase in sanicro® 60X, VB in Esshete 1250 and Sigma-phase in 4C54 were identified using EDSmapping. It was found that when using XRD to analyze the phase compositions of small samples it is impractical to have the samples cast in bakelite beforehand. The XRD-results obtained in this project showed that more than 90% of the XRD diffractogram for every sample was graphite, which made the identification of minority phases impossible. The quality of the LOM-images varied greatly between samples, for 4C54 grain sizes were measured in all images, for Esshete 1250 grain sizes were measured for the crept sample, and for Sanicro® 60X measurements could only be taken from one image. Most of the sample preparation was insufficient to achieve the test results necessary for complete microstructural analysis and phase analysis of the samples. The mistakes in the practical steps of the project were noted and improvements for these mistakes are presented in the conclusion.

The formation of nano–sized precipitates is considered to be the origin of hardening and embrittlement of ferritic steel used as structural material for pressure vessels of nuclear reactors, since these nanoclusters hinder the motion of dislocations within the grains of the polycrystalline bcc–Fe matrix. Previous investigations showed that these small precipitates are coherent and may consist of Cu, Ni, other foreign atoms, and vacancies. In this work a combination of on–lattice simulated annealing based on Metropolis Monte Carlo simulations and off–lattice relaxation by Molecular Dynamics is applied in order to determine the structure, energetics and thermodynamics of coherent clusters in bcc–Fe. The most recent interatomic potentials for Fe–Cu–Ni alloys are used. The atomic structure and the formation energy of the most stable configurations as well as their total and monomer binding energy are calculated. Atomistic simulation results show that pure (vacancy and copper) as well as mixed (vacancy-copper, copper-nickel and vacancy-copper-nickel) clusters show facets which correspond to the main crystallographic planes. Besides facets, mixed clusters exhibit a core-shell structure. In the case of v_lCu_m, a core of vacancy cluster coated with copper atoms is found. In binary Cum_Ni_n, Ni atoms cover the outer surface of copper cluster. Ternary v_lCu_mNi_n clusters show a core–shell structure with vacancies in the core coated by a shell of Cu atoms, followed by a shell of Ni atoms. It has been shown qualitatively that these core–shell structures are formed in order to minimize the interface energy between the cluster and the bcc-Fe matrix. Pure nickel consist of an agglomeration of Ni atoms at second nearest neighbor distance, whereas vacancy-nickel are formed by a vacancy cluster surrounded by a nickel agglomeration. Both types of clusters are called quasi-cluster because of their non-compact structure. The atomic configurations of quasiclusters can be understood by the peculiarities of the binding between Ni atoms and vacancies. In all clusters investigated Ni atoms may be nearest neighbors of Cu atoms but never nearest neighbors of vacancies or other Ni atoms. The structure of the clusters found in the present work is consistent with experimental observations and with results of pairwise calculations. In agreement with experimental observations and with recent results of atomic kinetic Monte Carlo simulation it is shown that the presence of Ni atoms promotes the nucleation of clusters containing vacancies and Cu. For pure vacancy and pure copper clusters an atomistic nucleation model is established, and for typical irradiation conditions the nucleation free energy and the critical size for cluster formation have been estimated. For further application in rate theory and object kinetic Monte Carlo simulations compact and physically–based fit formulae are derived from the atomistic data for the total and the monomer binding energy. The fit is based on the structure of the clusters (core-shell and quasi-cluster) and on the classical capillary model.