Quantification of Creep Damage by Atomic Force Microscopy

Abstract

The progression of creep damage in heat resistant steels for boiler application has to be monitored regularly as part of an asset integrity management. The damage is normally evaluated using qualitative technique which involves in-situ replication and microscopic examination. The results are rather subjective and depend on the personnel who carry out the interpretation and evaluation of the microstructures. The assessment is based on the distribution of creep cavities present and rarely considers their quantity. This qualitative approach has been used extensively to evaluate creep damage in conventional heat resistant steel such as grade P22 due to the reason that the microstructures are relatively simple with creep cavities readily observed. However, with the wider applications of supercritical technology in power industry, more complex alloys such as grades P91 and P92 steels are gradually replacing the P22 steels in the hotter part of the boilers to accommodate higher operating stresses and temperatures. This type of steels is of tempered martensite with numerous strengthening precipitates of carbides and carbo-nitrites. For this reason, the creep cavities in these alloys tend to be masked by the presence of precipitates. The creep cavities sometimes become less readily detectable under optical microscope. This paper reports a study on a quantitative method in which the depths and areas of cavities were measured using Atomic Force Microscopy (AFM). The cavities were formed by exposing P91 samples in a creep machine at a temperature of 625°C and a constant uniaxial load corresponding to 100 MPa. The results show that the method can be used to quantify creep cavities and can potentially be used to provide information on the creep life consumption of components operating in creep regime. For samples exposed to the above test conditions for 1200 and 3100 hours, the quantification of the cavities number at threshold -250 nm yielded 12 and 131 respectively for 100 μm2 areas of the samples examined.