PSI - Issue 43

Marie Ohanková et al. / Procedia Structural Integrity 43 (2023) 300–305 Author name / Structural Integrity Procedia 00 (2022) 000 – 000

304

5

in gradual transitions through both heat-affected areas and gradual tempering of individual weld beads in the weld metal. The hardness peaks thus remain in the root part of the weld and close to the melting lines.

a

b Fig. 2. (a) SEM of exposed microstructure; (b) TEM analysis of a cementite particle at a grain boundary.

3.3. Microstructure

Microstructure of non-exposed and exposed base materials is very similar formed by ferritic matrix with mostly plane bainite, see Fig 2a. At the grain boundaries, we can observe an occasional occurrence of particles, such as Cr 23 C 6 type or cementite Fe 3 C according to TEM, see Fig 2b. Base metals BM1. BM2 contain a larger amount of inclusions such as globular iron oxides (predominant in BM2) and elongated manganese sulphides (predominant in BM1, where they make up 0.11 % of the area). Otherwise, the microstructure of all basic materials is homogeneous throughout the thickness in all sections and, as expected, contains a small amount of bainitic grains with a size of G7-G8. The weld metal is full of iron oxides, otherwise its microstructure is formed by upper bainite with ferritic mesh. Therefore, apart from the low content of bainite in both exposed base metals, it was not obvious in the structure at first sight why this welded joint should be so significantly brittle. TEM diffraction of particles at the boundaries of the ferritic-bainitic structure of exposed sample showed that if nitrides, as a sign of operational aging, are present at the boundaries at all (it is not easy to reliably distinguish them from hexagonal cementite), they are so small that their contribution to the embrittlement was only small. On the contrary, the identified particles of primary cementite at the grain boundaries and the preserved plate-like morphology of bainitic laths rather indicate, consisting to Briant and Banerji (1978), a state caused by an insufficient heat treatment (either new or already operated material). The only obvious feature demonstrating the difference between the brittle (exposed) and non-brittle (non exposed) material is the elongated appearance of the ductile dimples and the transcrystalline ductile fracture matrix in the non-exposed segment. This elongation is an indication of the crack resistance of the matrix, which lacks the exposed segment material, as shown by the SEM images in Fig. 3. However, since no intercrystalline fracture was found in the fractographic analyzes of tensile, impact, or creep bodies of the exposed sample, the embrittlement of the material is therefore not related to weakened boundaries (cementitic particles or others) but to the matrix, which must have been caused by improper heat treatment. This is reflected in the presence of cementitic particles at the grain boundaries and a high ferrite content with a low degree of spheroidization of the bainitic plates. 3.4. Root cause of embrittlement