PSI - Issue 68

Julie Papin et al. / Procedia Structural Integrity 68 (2025) 727–733

732

6

Julie Papin et al. / Structural Integrity Procedia 00 (2025) 000–000

The reference temperature T 0 rises for a tempering parameter over approximately 20.00 at the quenching rate of 2000°C/h (Fig. 6). This tempering parameter corresponds to either a tempering temperature larger than 640 °C or/and a tempering duration larger than 6h. This graph also highlights an equivalence in fracture toughness for different microstructures defined by a set of quenching rate – tempering conditions (800°C/h – 640°C for 6h results in the same T 0 than 8000°C/h – 640°C for 20h) (Fig. 6a). The same evolutions of T 0 with tempering parameter after quenching at the other rates are observed, except after quenching at 8000°C/h and tempering parameter of 21.00.

$

"*$(,-./ &$$(,-./ %($$$(,-./ &($$$(,-./

!%$

!"#$%C$' )#(

!%$

!#$

!#$

C'#$%

!'$

!"%$ !"#$%&'$()*+ , - ./01 !"$$ !&$ !'$

!"#$ !"#$%&'$()*+7, -7 ./01 !"%$ !"$$ !&$

!!#$%

+#,,#$%#

!"#$%C$' !(

( !)

!"#$%

!"#$%

( !*

!"#$

!"'$

"$$

"($$$

"$($$$

"&)$$

"&)*$

"+)$$

"+)*$

%$)$$

%$)*$

%")$$

%")*$

,$'#$%&(67#*%*'$)$%

23$(45&(6 %*)$7./08517

Fig. 5. Reference temperature T 0 vs quenching rate. Tempering at 640°C for 6h (blue) and for 20h (grey).

Fig. 6. Reference temperature T 0 vs tempering parameter. Four quenching rates.

4.4. Effect of heat treatment on initiation site types The distribution of the types of crack initiation sites presented in Fig. 2 was obtained by merging the distributions obtained for each heat treatment condition, but these individual distributions are significantly different from one tempering condition to another. Although the number of specimens per heat treatment was not sufficiently high to draw quantitative conclusions, the distributions were then analysed for each heat treatment for qualitative comparison. From Fig. 7, the frequency of initiation on inclusions decreased when T 0 increased, regardless of the increase in T 0 being due to a decrease in quenching rate or to a variation of the tempering parameter. Indeed, the size of carbides increases with increasing TP or decreasing quenching rate. As the carbides found at crack initiation sites were larger than the overall carbide population, a competition between crack initiation from micrometre-sized inclusions and from the carbides at grain boundaries might take place. This competition might result of the quenching and tempering conditions, with more frequent crack initiation at boundary carbides obtained for larger values of TP and lower quenching rate (Chekhonin et al., 2023; Druce et al., 1992; Li et al., 2016). One can thus consider that, in the absence of large boundary carbides (i.e., low TP), brittle crack initiation more readily occurred at inclusions or at grain boundary but without any visible particle, while when coarser carbides are present due to favourable tempering conditions (i.e., high TP), their higher density led to a much more probable crack initiation at these particles than at intragranular inclusions. At the lowest cooling rate, namely, 150 °C/h, grain boundary carbides could already form and start growing during the quenching step. The resulting size of boundary carbides, for a given value of TP, was thus expected to be higher for this slower quenching condition, resulting in a higher probability of crack initiation from carbides. The fastest quenching rate, namely, 8000 °C/h, could induce numerous but finer carbides, even after tempering, which leads to higher propensity to crack initiation from inclusions. Tempering at 610°C for 6h seems to be less favourable to the growth of grain boundary carbides after quenching at 2000 °C/h, because there was mainly initiations from inclusions and no initiation from boundary carbides. The case of grain boundary initiation with no visible particle is more difficult to understand. The corresponding physical crack initiation mechanism, as well as its dependence on tempering conditions remains to be further investigated.