The Effect of Grain Boundary Character and Grain Boundary Engineered Microstructures on the Sensitization and Corrosion of 316L

There is strong evidence showing that the Brandon criterion for misorientation is insufficient determining whether grain boundaries in austenitic stainless steels will resist corrosion. Efforts have been made to use more restrictive criteria, however the prevailing theory is that the grain boundary plane plays a critical role in this phenomena. Understanding how they boundaries resist corrosion, and how they are integrated into the grain boundary engineered (GBE) microstructure, is critical to the successful optimization of these alloys for corrosion resistance. This thesis presents experimental evidence comparing the corrosion performance and microstructural characteristics of GBE microstructures arrived at by cold rolled and tensile strain routes. An iterative process was developed to propagate the Rolled GBE microstructure through the thickness of the specimen by straining by a 5% reduction in area followed by a 1020 degrees Celsius anneal for 2 hours, which was repeated three times. The Tensile GBE used the same iterated processing, but received a 5% tensile extension. The two GBE microstructures were examined using the stereological analysis method and found to have the same level of coherent [sigma]3 grain boundaries as the Annealed microstructure. Measurements of the length fraction of grain boundaries using different misorientation tolerance criteria showed that the Tensile GBE had significantly more highly deviated [sigma]3 grain boundaries than the Rolled GBE. The Rolled GBE possessed more 2-CSL triple junctions, however when only [sigma]3 grain boundaries are considered special, the Tensile GBE possessed more. The Rolled GBE also showed a consistently larger grain size than the Tensile GBE, when considered with [sigma]3, without [sigma] 3, and without [sigma]3, [sigma]9, and [sigma]27 grain boundaries. All of these measures, except the 2-[sigma]3 triple junction metric, suggest that the Rolled GBE should better resist sensitization and corrosion. However, the mass loss testing clearly shows that the Rolled GBE has higher mass loss than the Tensile GBE microstructure in 1, 10, and 100 hour sensitization conditions. The case is further complicated, as the 1000 hour condition shows that the Annealed and Rolled GBE conditions both had lower mass loss due to desensitization but the mass loss of the Tensile GBE continued to increase. Extended exposure of the 100 hour condition shows that the mass loss rate has not reached steady state yet, and may approach that of the Annealed microstructure when it is finished. U-bend testing shows that the high nickel alloy used in this work is immune to stress corrosion cracking under normal testing conditions, cracking only during the aggressive ASTM G36 test. A method was developed to use ACOM to calculate the local crystallographic faces in TEM specimens. This was used to evaluate grain boundaries which proved to be corrosion resistant. The [sigma]9 grain boundary that resisted corrosion was found to have a (111) grain boundary crystal plane on one side, and a continuous array of nanotwins on the other, while the [sigma]9 that did corrode did not have a low index crystallographic plane. Neither [sigma]3 grain boundaries which corroded had a (111) grain boundary, and one that appeared curved in SEM was found to have nanoscale facets. A pair of [sigma]7 grain boundaries are examined, one of which corroded and the other did not, and neither had a low index crystallographic plane at the grain boundaries. All examined grain boundaries except the coherent [sigma]3s are enriched in Cr and depleted in Fe, Ni or both. This shows that in the 1 hour sensitization condition, the grain boundaries are not sensitized in the classic sense, suggesting that the corrosion is related to an inherent feature of the grain boundary structures in this case.Ph.D., Materials Science and Engineering -- Drexel University, 201