PSI - Issue 13

4

T. Depover et al. / Procedia Structural Integrity 13 (2018) 1414–1420 Author name / Structural Integrity Procedia 00 (2018) 000–000

1417

Although H trapped at dislocations shows E a ’s in the same range (Pressouyre (1979)), this trapping site is confirmed to be undetectable due to the specific experimental requirements to perform the TDS analysis with the equipment used in this work. Moreover, the H which has been released from the sample before the TDS measurement started, has been confirmed to be linked to H trapped by dislocations (Depover et al. (2018 (c)) and Pérez Escobar et al. (2012)). Furthermore, a single peak was also exposed for the as-Q condition of Fe-C-Ti, indicating that the incoherent large TiC particles, shown in Fig. 1(a) were not able to trap H from electrochemical cathodic charging. This confirmed previous results where gaseous H charging at elevated temperature was required to charge these large particles (Wei et al. (2006)) and H from the gaseous charging being presumably trapped inside the carbide rather than at its interface. Considering the Q&T condition, a significantly different trapping behavior was observed due to the tempered induced carbides. The additional peaks in the TDS spectra were therefore attributed to the presence of these precipitates. The small TiC and V 4 C 3 precipitates were capable of trapping a lot of H, resulting in three peaks with activation energies in the range of 44-71 kJ/mol. The H trapping characteristics of the carbides will be further considered by increasing the tempering time. 3.3. Effect of carbide addition on H diffusivity evaluated by permeation transients H diffusion plays an important role to understand the interaction between H and a material. Therefore, H permeation tests were performed. The permeation curves for both materials are shown in Fig. 3. A remarkably slower hydrogen permeation was obtained after tempering which was due to the presence of abundant nano-sized TiC or V 4 C 3 particles. The resulting H diffusion coefficients for Fe-C-Ti were 1.14 x 10 -10 and 3.02 x 10 -12 m 2 /s for the as-Q and Q&T condition, respectively. This is consistent with previous reported results on the effect of nano-scale precipitates on the H diffusivity (Brass et al. (2006) and Liu et al. (2017)). For Fe-C-V, the corresponding diffusion coefficients were 8.53 x 10 -11 and 1.16 x 10 -12 m 2 /s for the as-Q and Q&T material, respectively. The slower H diffusivity properties for the as-Q condition of Fe-C-V compared Fe-C-Ti can be understood by the density of the martensitic matrix. Since a considerable amount of large TiC was not dissolved during austenitizing (cf. Fig. 1 (a)), less carbon was present in the lath martensitic matrix of Fe-C-Ti after quenching. Consequently, a harder and stronger material was obtained for Fe-C-V as-Q, as confirmed in our previous publications, as such slowing down the H diffusion in Fe-C-V as-Q.

As-Q Q&T

As-Q Q&T

1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

Normalized current

Normalized current

0.0

0.0

0.1

1

10

100

0.1

1

10

100

b) Figure 3: Hydrogen permeation curves of Fe-C-Ti (a) and Fe-C-V (b) in the as-Q and Q&T condition. Time (h) Time (h)

a)

3.4. Effect of carbide characteristics on TDS and permeation results in Fe-C-Ti Fe-C-Ti tempering was also done for 2 h to increase the carbide size and evaluate the H trapping ability. The corresponding TDS spectra and carbide size distributions are included in Fig. 4. The larger incoherent TiC were unable to trap electrochemically charged H. When the material got tempered for 2 h (Q&T 2h), peak 4 disappeared, whereas peak 2 and 3 diminished compared to Q&T 1h. Since the total interfacial area between carbides and matrix decreases with carbide growth during tempering, H was confirmed to be present at the interface of the carbides. Drexler et al. (2018) recently made a model based interpretation of these TDS data and revealed that peak 4 was linked to carbon vacancies inside TiC, whereas indeed the other carbide related peaks were linked to the carbide/matrix interface. An alternative estimation of the available trapping sites can be made by performing permeation tests, shown in Fig. 5.