Ultrasonic spectral analysis for microstructural characterization of austenitic and ferritic steels

Abstract

The paper discusses the use of ultrasonic spectral analysis for microstructural characterization in AISI type 316 stainless steel and modified 9Cr-1Mo ferritic steel (T91/P91). Several specimens of AISI type 316 stainless steel have been heat treated at different temperatures varying from 1373 to 1623K. for different time durations (from 15 to 120min) to obtain specimens with different grain sizes. Spectral analysis has been carried out on the rf signal corresponding to the first back-wall echo obtained with a 25 MHz delay line longitudinal beam transducer. The shift in the peak frequency and the change in the full width at half-maximum of the autopower spectrum have been correlated with the average grain size and yield stress. Various specimens of modified 9Cr-1Mo ferritic steel have also been subjected to a series of heat treatments consisting of soaking for 5min at different temperatures in the range 1073-1623K followed by oil quenching. These treatments were given to obtain different microstructures and grain sizes in the specimens simulating the different regions in the heat-affected zones of the weldments. Spectral analysis has been carried out on the rf signal corresponding to the first back-wall echo obtained with the 20MHz longitudinal beam transducer. Two distinct peaks at around 7.0 and 17.5MHz have been found in the autopower spectrum of the first back-wall echoes of all the specimens. The ratio of these two peaks (the amplitude of the lower-frequency peak to the amplitude of the higher-frequency peak) has been correlated with the prior austenitic grain size. A linear correlation between the spectral peak ratio (SPR) and the grain size is found to be valid for a wide range of microstructures, that is ferrite, ferrite + martensite and martensite. The variation in the SPR with the soaking temperature shows that various metallurgical events affecting the grain size, namely the formation of martensite (Ac₁-Ac₃ temperature range), dissolution of NbC and V₄C₃ and formation of δferrite (Ac₄ temperature), can also be determined using the SPR but, for more accurate determination of these temperatures, on-line measurement is required. For the first time, we show that the various spectral parameters used in this study are independent of the coupling condition. The significance of the results for practical utility is also discussed.