*3.2. Effect of Prior Austenite Grain Size on the Continuous Cooling Transformation Temperature and Transformation Rate*

Figure 4 shows the simulated heat-affected zone continuous cooling transformation diagram. Figure 4a is a SHCCT diagram for coarse-grained HAZ with the peak temperature of 1320 ◦C, while Figure 4b is for fine-grained HAZ with the peak temperature of 850 ◦C. Both diagrams indicate that martensite was the main phase. However, the martensite's start transformation temperature in FGHAZ was higher than that in CGHAZ by 17~31 ◦C at the same cooling rate, while the finish transformation temperature in FGHAZ was higher than that in CGHAZ by 31~58 ◦C, with an exception at the t8/5 of 7.5 s. Moreover, the transformation temperature range in FGHAZ was smaller than that in CGHAZ, as shown in Table 3.


**Table 3.** Continuous cooling transformation temperature at different t8/5.

**Figure 4.** Simulated heat-affected zone continuous cooling transformation diagram for peak temperature of (**a**) 1320 ◦C; (**b**) 850 ◦C.

Mn has a definite segregation tendency due to its differential diameter of 5 × <sup>10</sup>−<sup>12</sup> m with the Fe element. During the welding thermal cycle, Mn may segregate towards gain boundaries by the nonequilibrium segregation mechanism. The higher the peak temperature is, the higher the segregation concentration of Mn at grain boundaries is. For FGHAZ, the small grain size allows Mn and C to diffuse short distance to grain boundary. Furthermore, the great number of grain boundaries can enhance more Mn and C segregates at the grain boundaries. Therefore, the concentration of Mn and C should be lower in FGHAZ than in CGHAZ. Consequentially, the martensite start temperature was higher in FGHAZ than in CGHAZ.

The in situ observation of martensite transformation by SLCM (Figure 5) shows that lath martensite nucleated inside austenite grain and then grew quickly along the length direction. The austenite was divided by the first generated longer lathes. Then the lath widened. In the fine austenite grain, the martensite grew in the limited space, so short and thin martensite lath was found in FGHAZ. On the other hand, long and thick martensite lath was found in CGHAZ.

**Figure 5.** *Cont*.

**Figure 5.** Martensite transformation in the cooling stage in CGHAZ at different temperature of (**a**) 528.2 ◦C, (**b**) 504.8 ◦C, (**c**) 429.4 ◦C.

Figure 6 reveals the martensite transformation rate vs martensite transformed volume fraction. At the peak temperature of 1320 ◦C, the transformation rate from austenite to martensite was measured to be a maximum of 0.60, while at the peak temperature of 850 ◦C, the transformation rate from austenite to martensite was measured to be a maximum of 0.98. The transformation rate is closely related to the transformation start temperature and transformation temperature range. The higher start transformation temperature often leads to a relatively high transformation rate. Therefore, because the martensite transformation temperature was higher at the peak temperature of 850 ◦C than that at 1320 ◦C, as well as the grain size was smaller at the peak temperature of 850 ◦C than that at 1320 ◦C, FGHAZ expressed the relatively higher transformation rate.

**Figure 6.** Martensite transformation rate VS martensite transformed volume fraction with different cooling rate at peak temperatures of: (**a**) 1320 ◦C, (**b**) 850 ◦C.
