Degree Name

Doctor of Philosophy


Department of Materials Engineering


The topic of this thesis covers a broad range of phenomena through three significant phase transformations in iron alloys, being solidification, delta-ferrite to austenite and austenite decomposition. An appreciation of the links between these diverse transformations is necessary to understand the purpose of the present study. One such link is the concept of morphological stability of interphase boundaries during transformations. Another link between the three phase transformation is the lack of studies conducted utilising in-situ, real-time observations of phenomena. With the advent of high temperature laser scanning confocal microscopy, it is now possible to overcome previous experimental difficulties associated with resolving images at high temperature. Therefore, experimental analysis can be conducted into not just the morphology of transformation interfaces, but into a range of additional phenomena such as the development of recovery structures and growth kinetics.

For the studies conducted utilising LSCM it was necessary, due to the novelty of the technique, to establish both a frame of reference for interpreting observations, and to characterise the influence of the free surface on the observations. In regard to the nexus between events on the free surface and in the bulk, endeavours were undertaken to establish phenomena that are affected by the free surface and those that are not. Serial sectioning analysis has led to the conclusion that solid-state phase transformations of delta-ferrite to austenite and austenite decomposition do show a correlation between events on the free surface and those in the bulk. Phenomena that do not exhibit a nexus between events on the free surface and in the bulk were identified. These were the pinning of grain boundaries by surface defects, anomalous massive phase transformations and precipitation of non-metallic particles.

The aim of the experimental studies into the relationship between crystallographic orientation and the morphological stability of a solidification front was two-fold. Apart from the aim of establishing such relationships, it was also used to assess LSCM as a tool for directional solidification experiments. It was demonstrated that the LSCM technique has potential in the study of the relationship between crystallography and solidification behaviour, however limitations associated with the present design precluded a quantitative assessment of this issue. The thermal gradient in the LSCM hot stage was determined to be dependent upon the temperature of the sample, and not to be linear. Qualitatively, a variety of interface morphologies were observed for an iron-chromium alloy, including a doublon structure, associated in the literature with crystallographic orientations that lead to isotropic surface tension.

Experiments conducted to probe the issue as to whether the delta-ferrite/austenite interface exhibits morphological instability failed to support this hypothesis. Novel observations of the development of a recovery structure following the austenite to deltaferrite phase transformation in low carbon steels were made. The sub-boundaries that develop through polygonisation were determined to have an interfacial energy in the range of 7.5 to 36% (0.035 to 0.17 J/m2) of delta-ferrite grain boundaries. Austenite was observed to grow along sub-boundaries in preference to the bulk, thus the interface developed an apparent unstable morphology. An alternative mechanism for the observed unstable grow morphology is therefore proposed, it is the result of the preferential of austenite along delta-ferrite recovery structures. This mechanism is believed to more adequately account for the appearance of so-called intragranular austenite cells that merge with grain boundary austenite cells, than does the model proposed by Yin et al [1999].

Further experiments were undertaken that provide evidence the recovery sub-structure can play a role in the delta-ferrite to austenite phase transformation, and subsequently influence austenite decomposition. Through judicious selection of thermal cycles, a 0.06mass% silicon killed steel's microstructure was modified from fully Widmanstatten to polygonal ferrite, with a statistically significant reduction in hardness from 124 to 114Hv(10). The modified structure was achieved without recourse to thermo-mechanical processing, and this finding may have implications for the control of microstructure in the strip casting of low-carbon steels.

Experiments into austenite decomposition have provided evidence that a grain boundary allotriomorph can grow with a cellular morphology, In-situ observations recorded such a morphology in real-time. The wavelength of the instabilities was measured to be 5.5μ width, thus indicating that the observed growth was not pearlite. Subsequent optical metallography and crystallographic orientation analysis support the conclusion that the observed growth was cellular.

The growth rate of individual Widmanstatten ferrite plates in low carbon steels has been studied. It was observed that the growth of plates exhibits a random variability in their growth velocity. This observation reflects previous in-situ studies of Eichen et al [1964] and Onink et al [1995], but not analyses based on quench-arrest techniques such as Townsend and Kirkaldy [1968]. T h e variability of growth was not found to be associated with pinning effects of surface defects. The growth behaviour is not predicted by current diffusion controlled or mixed mode (diffusion and interface reaction) models. Of these two models, the mixed mode has the greater potential for assessing variable growth behaviour as the pre-exponential factor Mo of the inherent interface mobility M, incorporates effects such as transformation stresses and coherency of the interface.

In-situ observations were also conducted on the formation of Widmanstatten ferrite plates. These observations indicated that the plates were not forming via an unstable interface mechanism, but were separate nucleation events. Subsequent EBSD analysis revealed low-angle misorientation between grain boundary allotriomorphs and the plates that formed on them, of between 5 and 10 degrees. This indicates that sympathetic nucleation is the mechanism of Widmanstatten ferrite formation in these instances.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.