Date of Graduation


Document Type


Degree Name

Master of Science in Mechanical Engineering (MSME)

Degree Level



Mechanical Engineering


Ashok Saxena

Committee Member

Rick J. Couvillion

Second Committee Member

Paul C. Millett


Fracture Mechanics, Oxidation, Oxidation Kinetics, P91 Steel, Predicting Crack Growth


There are only few methods available for predicting the age of cracks that are found in high

temperature structural components during service; among the promising ones is the oxide

thickness measurement technique. Oxide thickness profiles are taken from crack surfaces of

components and used for predicting the rates of crack propagation. This technique is particularly

suitable for high temperature components fabricated from ferritic steels commonly used in power

plants that run on fossil fuels. To implement this technique, it is necessary to fully understand the

kinetics of high temperature oxidation in these steels. In this study, the oxidation characteristics

of an American Society of Testing and Materials (ASTM) Grade P91 ferritic steel used in high

temperature piping is characterized.

The literature shows that there are four primary mechanisms that influence the oxide thickness

during high temperature exposure. Initially, the oxide thickness increases in a linear fashion with

time and then as steady-state conditions are established, the parabolic relationship takes over.

Multiple types of oxides with different rate characteristics can also form. Oxide degradation can

occur by spallation due to porosity and formation of cracks. Evaporation or volatility can also

occur and result in loss of oxide thickness. These factors must be considered in oxide thickness

analysis to determine crack growth history.

Two sets of laboratory experiments were conducted. The first consisted of measurement of oxide

thicknesses after exposure to high temperature for various periods to determine the oxidation

kinetics. The oxidized samples were subjected to SEM examination and measurements of

physical properties such as density and porosity levels. The second set of experiments consisted

of measuring the oxide layer thickness on the fracture surfaces of creep-fatigue crack growth

samples tested as part of a previous study where the crack growth rates were measured. These

reported measurements are used to compare with the predicted crack growth rates from the

analytical models that are developed as part of this study. The success of the technique is

measured by finding the correlation coefficient, which is within a factor of 2.58.