Defect Density Imaging of Thick Steel Tensile Specimens

At the physics department of Idaho State University we have been developing accelerator-based in-situ pair-production techniques to produce positrons in thick samples (∼ 4-40 g/cm² , or ∼ 0.5-5 cm in steel). These techniques are called ”Accelerator-based Gamma-induced Positron Annihilation Spectroscopy” (AG-PAS). AG-PAS employs an electron accelerator to create a bremsstrahlung cone which, after collimation, strikes a sample and generates positrons in the sample volume via pair production. The annihilation radiation is counted by a high resolution HPGe, which allows us to analyze the Doppler broadened 511 keV peak in terms of S and W parameters. This approach creates a possibility to add defect density imaging capabilities of large scale engineering materials to the conventional PAS techniques, because it combines high penetration of MeV gamma rays with the high sensitivity of positrons to microscopic defects. The work presented here serves as the feasibility study to investigate the potential of this new technique. The main limitation of the AG-PAS experiments performed in the past was the low repetition rate of the linear accelerator (LINAC), which was around 200-300 Hz. The experiments in this project were performed by using a 1 kHz High Repetition Rate LINAC (HRRL). Our imaging sample was four 304L steel tensile specimens inserted in a plastic holder. One of the specimen was an undamaged, control sample, while the other three had preinduced mechanical strain with different loads - 2500 lbs-force, 3200 lbs-force and 4000 lbs-force. We generated 2D contour plots of the S parameter (which is proportional to defect density) after scanning our sample in x and y directions. The results confirmed the imaging potential of the AG-PAS technique and also revealed it's limitations. In order to produce high resolution defect density plots of large scale objects and have feasible data acquisition times, a higher repetition rate accelerator is needed.

Acknowledgments

This work was funded by a grant from DOE # DE-FC07-Q6ID14780