Abstract
The Quenching and Partitioning (Q&P) process starts with an austenitization process followed by rapid cooling to a temperature named quenching temperature ranged between the martensite start (M_s) and martensite finish (M_f) temperatures to establish initial fractions of martensite and austenite. Afterwards, the material is reheated to a particular partitioning temperature and for a given time to allow the carbon diffusion from martensite to austenite. Finally, the steel is quenched to room temperature and a fraction of fresh untempered martensite may form from the least stable austenite. The whole process determines the phase fraction of the constituent phases, namely: tempered martensite (M₁), retained austenite, and un-tempered martensite (M₂). This combination of phases has the potential to simultaneously improve strength and elongation of steels due to their small grain sizes and the transformation-induced plasticity resulting of the transformation of the retained austenite into martensite during a deformation. 
It is known that the fraction of retained austenite of Q&P steels depends on the quenching temperature, showing a bell-like relation. The volume fraction of retained austenite shows a maximum at a particular quenching temperature somewhat intermediate between the M_s and M_f temperatures, known as the optimum quenching temperature, below and upper which the volume fraction of retained austenite decreases in either way. 
Recently, it is reported [1] that retained austenite fraction after reheating up to 700^oC followed by quench losses its dependency of quenching temperature and reaches a constant value. This can be a manner to further control phase fractions and consequently the mechanical properties of Q&P steels. The mechanisms by which the volume fraction of austenite changes to reach that constant value are not well understood and are the focus of the present study. 
In the present work, Q&P samples with chemical composition 0.31C-4.58Mn-1.52Si (wt.%) were reheated up to 700^oC and quenched directly to compare the retained austenite fractions of reheated and un-reheated (Q&P) samples. It was found that the retained austenite fractions tend to have a constant value after reheating to 700^oC followed by quenching, whereas the retained austenite fraction of the Q&P samples showed an optimum value and usual dependency with quenching temperature. It was also observed that the highest and lowest drop of retained austenite occurred in the reheated Q&P samples at 160^oC and 80^oC of quenching temperature. These samples were reheated to 900^oC to have a complete observation of microstructural events by means of dilatometry, optical and electron scanning microscopy, hardness, and X-Ray diffraction measurement. 
Interrupted reheating was applied to understand the microstructural changes during reheating. The two samples were interruptedly reheated and directly quenched at 450^oC, 530^oC, 610^oC, 720^oC, and 740^oC to measure the retained austenite fraction and to capture the corresponding microstructure around the observed changes in slope. The retained austenite fractions evolution of both samples were measured and compared. It was found that the decreasing of retained austenite fraction started around the critical temperature of 600^oC and remained up to 700^oC of reheating temperature. Further reheating process increased the retained austenite fraction. 
In the specimen showing the lowest drop of retained austenite fraction, retained austenite decomposed by two mechanisms. The first one was the formation of discrete particles of cementite, which originally came from film type austenite. The second was the formation of globular carbides within tempered martensite. On the other hand, river-like patterns were observed surrounding the martensite region. The river-like patterns mainly consist of martensite from new austenite that formed between 720^oC and 750^oC.
In the highest drop of retained austenite fraction sample, retained austenite decomposed by three mechanisms. The first two mechanisms were similar to the previous sample. The last mechanism was the formation of substructure within Martensite-Austenite islands. River-like patterns were also observed surrounding the martensite region, consisting of martensite which originally came from new austenite as a product of (martensite and) carbide transformation that formed between 720^oC and 745C and lath martensite which originally came from the unstable austenite forming upon reheating to 740^oC.