Inhibition of grain growth by pearlite improves hydrogen embrittlement susceptibility of the ultra-low carbon ferritic steel: the influence of H-assisted crack initiation and propagation mechanisms

The influence of grain growth inhibition by pearlite on hydrogen embrittlement (HE) behavior of ultra-low carbon ferritic steels was studied. The Fe-0.02C alloy has a considerable sensitivity to HE, while the Fe and Fe-0.1C materials have lower HE sensitivity. The high fraction of high-angle grain boundaries (HAGB) is responsible for the shift of the peak to a higher temperature in the Fe-0.1C alloy. The trapping of hydrogen by HAGBs was observed by hydrogen mapping. Higher density of HAGBs contributes to higher trap density in Fe-0.1C alloy and lower H diffusion coefficient. The coexistence of hydrogen enhanced decohesion (HEDE) and hydrogen-enhanced localized plasticity (HELP) mechanisms was identified and discussed. The results indicate that HEDE is active in the initial stage of the tensile loading during crack initiation and HELP in the rest of tensile testing duration that controls the crack propagation. Due to the reduction in grain size caused by the addition of pearlite, the normalized hydrogen content per HAGB length is significantly lower than Fe-0.1C, which means that the critical amount of local hydrogen required for crack initiation is less likely to accumulate due to the weakening of cohesive interatomic strength. In Fe-0.02C alloy, the ciritcal local content was built up at HAGBs, and resulting in activation of HEDE mechanism for crack initiation. In Fe-0.1C, a lower possibility of accumulating the critical H concentration at HAGBs, resulting in limited activation of HEDE-based crack initiation at HAGB and less crack propagation events based on the HELP mechanism.


Introduction
The energy transition from CO 2 -emitting fuels such as natural gas and petroleum fuel to green energy sources is the focus of global interest following the Paris Agreement and was recently announced at a recent policy forum [1]. Despite advances in the various green energy production techniques, a critical challenge is still that production cannot meet the demand for energy. One solution to this problem is the conversion of surplus energy into fuel, specifically the production of hydrogen through electrolysis [2]. Hydrogen as a medium for energy storage can store energy until it is converted into electricity. Hydrogen energy will make an important contribution to the development of clean energy. Therefore, the development of safe and cost-effective technologies for hydrogen storage and transport is critical to the energy transition process. One of the obstacles to the development of these technologies is the sensitivity of structural materials to hydrogen when they react with hydrogen. Carbon steels, which are an important group of structural materials due to their weldability, are sensitive to hydrogen embrittlement (HE). HE is a common cause of failure of engineering alloys in a varying mixture of stress and corrosive media. Low-carbon steels are also considered susceptible in a hydrogen-containing environment to HE [3][4][5][6][7]. Hydrogen has a deleterious effect on mechanical properties at room temperature and leads to a loss of ductility [4]. We have already reported [4] that higher testing temperatures can lead to less HE susceptibility. Despite several studies on the hydrogen embrittlement behavior of carbon ferritic steels [8][9][10][11][12], there are few studies on the effects of hydrogen on the properties of ultra-low carbon steels to define the role of the determining parameters for the HE behavior of the alloys.
Hydrogen is absorbed from the hydrogen-containing environment, and the hydrogen absorption rate and the distribution of hydrogen among the trap sites play a crucial role in defining the HE behavior of the alloys [13,14]. Several studies have reported that the resistance to HE can be modified depending on the design of the trapping sites and the hydrogen trapping behavior in the microstructure [7,[15][16][17][18][19][20]. Hydrogen absorption and distribution depend on microstructural features resulting from chemical compositions. Thus, the microstructural features such as second phase particles, dislocations, grain boundaries, etc., can affect hydrogen absorption and distribution as well as the HE behavior of the alloy. Park et al. [9] reported that the grain boundary in carbon ferritic steels plays an important role in trapping and in defining the HE behavior of the alloy. Added to this, it was reported that [8] the prestrain can lead to a change in the nature of the dislocation and grain boundaries, resulting in a change in the H trapping behavior and the corresponding HE behavior. In the ultra-low carbon range in Fe-C systems, the role of trapping as a function of microstructural features and chemical compositions has not yet been elucidated. A systematic study is needed to clarify the role of hydrogen trapping behavior as a function of compositions and corresponding microstructure in the HE sensitivity of ultra-low carbon steels.
The objective of this study is to investigate the susceptibility to hydrogen embrittlement in ultra-low carbon steel with different carbon contents. The present work will be carried out by studying the hydrogen trap sites, the amount of trapped hydrogen and the activation energies are investigated by thermal desorption spectroscopy (TDS), electrochemical hydrogen permeation tests, hydrogen mapping by hydrogen microprint technique (HMPT) and hydrogen embrittlement susceptibility index by slow strain rate test (SSRT). In addition, the microstructure is studied using a scanning electron microscope (SEM) integrated with an electron backscatter diffraction (EBSD) detector. The obtained results contribute significantly to improve the current understanding of the effects of different carbon contents and the corresponding microstructures on the HE behavior of the alloys and the mechanisms of HE mitigation/susceptibility in the ultra-low carbon ferritic steels. Table 1 gives an overview of the chemical composition, microstructure, different phases and heat treatment of all investigated materials including pure Fe. The two alloys with 0.02 and 0.1 wt% carbon contents are hereafter referred to as Fe-0.02C and Fe-0.1C with higher carbon contents. Casting of the materials was performed in a vacuum induction furnace. The alloys with different carbon contents were prepared by the incremental casting method. The obtained materials were rolled at 950 °C. The materials were then soft annealed at 700 °C for 5 h and then cooled in the furnace with the aid of a table fan.

Electrochemical hydrogen charging
Electrochemical H charging was performed in a setup with the material as a cathode and Pt as the anode. The samples were cathodically charged in a 3.5% NaCl solution with 1 g/l thiourea and 1 mA.cm −2 current density and a charging time of 24 h. The H-charged specimens were used for SSRT and TDS experiments.

In-situ slow-strain rate testing (SSRT)
Before starting the SSRT, the specimens were precharged in the same solution, current density, and charging time mentioned in "Electrochemical hydrogen charging section". The SSRT tests with in situ H charging were carried out at the strain rate of 3.4 × 10 -6 s −1 . The dimensions of the gauge part of the flat tensile specimen are 12.0 × 5.0 × 2.0 mm 3 . As a reference test, an SSRT with a similar strain rate was done in the air. For having reasonable repeatability, the tests were done three times per condition, and the representative results were reported.

Thermal desorption spectroscopy (TDS)
The H-charged specimens were immediately put into the liquid nitrogen since at the low temperature the possibility of hydrogen escape is lower. The samples were tested using a quadrupole mass spectroscope coupled with a furnace and a temperature logger. The test was done at heating rates of 300, 600 and 1200 k.h −1 . The changing of the heating rate enables us to use the Choo-Lee method to calculate the desorption energy (E d ) of H traps based on the evaluation of the peaks identified in thermal desorption (TD) spectra [21,22]: where Ф, R, T m are the heating rate, the gas constant, and the desorption peak center, respectively.

Hydrogen microprint technique
To visualize the local sites of hydrogen accumulation in the alloys, a hydrogen mapping technique, the hydrogen microprint technique (HMPT) was utilized. The Fe-0.1C specimen with polished surface was electrochemically charged and subject to HMPT mapping. The solution containing silver bromide (AgBr), formalin for hardening the gelatin layer, Na 2 S 2 O 3 water solution as a fixing solution were used. The other details of the hydrogen mapping procedure can be found elsewhere [23,24].

H permeation experiments
The H permeation experiments were carried out using a Devanathan-Stachurski (D-S) method [25,26]. The dimension of the specimen was 40 × 40 × 1 mm 3 . The Pd coating with a thickness of 90 nm was done by physical vapour deposition. The Pd layer was applied to lead to a faster oxidation rate and hinder the H recombination to H 2 and avoid the formation of a passive layer. A double cell with a KCl saturated electrode (reference) and a Pt electrode (counter) was used for the measurements. Argon purging was done before starting the measurements. The charging cell started to work with 1 mA cm −2 current density. The scheme of the D-S double cell and the applied parameters for the electrochemical permeation test were shown in Fig. 1 and Table 2. More details about the H permeation experiment can be found in our previous work [26].
The D eff , effective diffusion coefficient, can be measured by the time lag method [16,27]. The time lag was defined as the intersection of the flux with the time axis. According to Devanathan and Stachurski [25], the following potentiostatic control method can be used for measuring the D eff via the time lag method: where L is the membrane thickness, t lag is the time lag for potentiostat charging, i ∞ the steady-state current density, t is time. Then, c app was calculated and used for the calculation of the permeation coefficient/permeability ( Φ): and the following equation can give us an estimation of total hydrogen trap sites (N T ): More details about the technique and calculations can be found in the other studies [25,26].

Microstructural characterizations
The texture, phase fraction, grain size and grain boundary types were analyzed by EBSD, using an EVO Zeiss SEM integrated with an EBSD detector. Figure 1 Schematic representation of the electrochemical permeation test facility. After mirror-finishing the surface, the ion slicing was applied via Hitachi IM4000 + . During ion slicing, a titanium mask is used with a part of the specimens outside of the mask for directing the specimen to the beam and flattening the surface. The EBSD measurements were done with a step size of 100 nm and at 20 kV energy of electron beam. For more details on sample preparation and EBSD measurement, please see refs. [4,22].

Results and discussion
Crystal orientation and phase analysis depending on the carbon content and Fe-0.1C specimens while te fraction of LAGBs in pure Fe is lower than the other specimens ( Fig. 2b, f, j). HAGBs and LAGBs were defined when adjacent pixel in the map exhibit a misorientation of > 15° and < 15°, respectively. During the heat treatment, in the two alloys containing carbon, recrystallization occurs. Due to the partially incomplete recrystallization, LAGBs form, which form sub-grain boundaries within the grains [28]. The phase map results ( Fig. 2c, g, k) show that the fraction of carbide increases with increasing the carbon content, indicating the higher content of pearlite structure in higher carbon content. This is in agreement with Fig. 2m (Fe-C phase diagram), showing the increasing content of pearlite with increasing carbon content. Figure 3 shows SEM micrographs with high magnifications of the Fe-0.02C and Fe-0.1C alloys, showing the pearlite microstructure formed at the grain boundaries and triple junctions. The grain size is decreased with increasing the carbon content due to grain growth inhibition in higher content of the pearlite that was located at grain boundaries and triple junctions (see the grain distributin histogram in Fig. 2d, h, l). The grain distribution histograms were obtained by analysis of the 3 EBSD dataset to plot a large area of the specimens. The average grain sizes of Fe, Fe-0.02C and Fe-0.1C are 84, 32, and 17 µm, respectively. Figure 4 shows the engineering stress-strain curve of Fe, Fe-0.02C, and Fe-0.1 materials with and without in situ hydrogen charging. Increasing the carbon content leads to an increase in the cementite content in the alloy (as seen in Fig. 2). Since the cementite is harder than the ferrite matrix, the yield strength and ultimate tensile strength of the alloy increase at higher carbon contents. This increase in strength is accompanied by an increase in elongation at fracture and thus higher toughness as a result of testing in air ( Table 3). The larger work hardening rate during SSRT of the Fe-0.02C speciem and then Fe-0.1C specimen results from a higher toughness of the sample due to the higher carbon content and higher fraction of pearlite particles [29]. In addition, the yield strength and ultimate tensile strength for each specimen remained almost unchanged when the test environment was changed, which is consistent with other reported results [30,31].

Hydrogen-induced damage behavior
To quantify HE resistivity, the total elongation loss (hydrogen embrittlement sensitivity (HES) index) is calculated based on the following equation and the results tabulated in Table 4: where δ Air is the total elongation in air and δ H is the total elongation in the hydrogen environment. The  elongation losses of the pure iron and Fe-0.1C alloy were relatively low, compared with the high elongation loss in Fe-0.02C alloy. These results show that Fe-0.02C alloy has a considerable sensitivity to HE, however, Fe and Fe-0.1C materials showed a relatively lower HE sensitivity. Figure 5 shows SEM images of the tensile-tested Fe-0.02C alloy where most areas of the fracture surface of the H-charged Fe-0.02C specimen contain quasicleavage (QC) indications with the serrated patterns (Fig. 5a). There is some limited tear ridge morphology, however, no intergranular fracture surface can be observed for Fe-0.02C specimen. The fracture surfaces of the uncharged specimen were fully covered with dimples. This implies a fully ductile fracture of the tensile-tested Fe-0.02 specimen in the air (Fig. 5b). The fracture surface observation results agree well with the HES index of Fe-0.02C alloy, implying considerable sensitivity to hydrogen-induced fracture. The H-induced QC with serrated indications were observed for some low-carbon steels. Serrated indications are related to the area which consists of {011} facets that the longitudinal directions are parallel to < 110 > or < 112 > direction [4,32]. The slip is accelerated by hydrogen and hydrogen-enhanced plastic deformation leads to the coalescence of vacancies/ voids. These coalescences are responsible for the QC fracture. Figure 6 presents SEM images of the tensiletested Fe-0.1C specimen. The H-charged Fe-0.1C   sample contains a limited area of QC fracture, while the fracture surface is covered by a high number of dimples, implying a relatively ductile behavior in the H-charged Fe-0.1C specimen. This is accompanied by fully covered fracture surfaces with dimples in the uncharged Fe-0.1C specimen. This is in a good agreement with the HES index of Fe-0.1C specimen, showing mitigation of the hydrogen embrittlement in Fe-0.1C alloy. Since the level of embrittlement and the hydrogen absorption is low in the pure Fe and the fracture surface is fully ductile, it was not possible to analyse the fracture surface. Figure 7 shows the TD spectra (at the constant heating rates) of the hydrogen-charged Fe, Fe-0.02C and Fe-0.1C specimens with the condition mentioned in "Electrochemical hydrogen charging" section. The results indicate that the centers of the peaks in Fe and Fe-0.02C specimens are comparable. However, the TDS peak of Fe-0.1C alloy was shifted to the right side with a higher peak center compared to those of Fe and Fe-0.02C specimens. The total hydrogen contents were detected by TDS and reported in Table 3. The results showed that the total hydrogen content of the hydrogen-charged specimen enhanced with increasing the carbon content of the alloy. Pure Fe has a very low content of hydrogen, relating to a very low amount of trap site in the material. In Fe-0.02C alloy, there is a higher total hydrogen content than that of pure Fe, due to a higher amount of hydrogen trap sites. In Fe-0.1C alloy, the total hydrogen content is higher than those of pure Fe and Fe-0.02C materials. Thermally ramped desorption spectra were performed on Fe, Fe-0.02C, and Fe-0.1C specimens at various heating rates, and the results were shown in Fig. 8a, b, c. The obtained H desorption results were evaluated by the Gaussian function. The analysis indicated that there is one peak in the TDS results in all three specimens. Higher heating rates shift the peak to higher temperatures. The H desorption energies were calculated via this peak shifting phenomenon at different heating rates. By Eq. (1) [21] and the centers of peaks obtained from Fig. 8a, b, c, one can calculate the H desorption energies. The TD spectra, peak analysis, and energy calculations revealed that the trap activation energy corresponds to the H desorption energies of 29.5, 30.6, and 51.3 kJ mol −1 for the peaks in Fe, Fe-0.02C and Fe-0.1C specimens, respectively ( Fig. 8d and Table 4). The desorption energies of the trap sites in Fe and Fe-0.02C specimens are comparable while there is a very low total H amount in pure Fe. The desorption energy of the trap site in Fe-0.1C is higher than the energies of Fe and Fe-0.02C specimens. Since there are high overlaps of the peaks, it is difficult to deconvolute the main peak to the smaller peaks [21,33,34]. In the ferritic structure of iron and steel, the trap sites were reported as interstitial lattice sites, dislocations and grain boundaries ( high angle) [9,21,22,34]. It was reported that the HAGBs are related to the trap sites with the highest hydrogen desorption energy.

H permeation results
The H permeation measurements were carried out on Fe, Fe-0.02C and Fe-0.1C materials, and the results were shown in Fig. 9. The thicknesses of specimens were controlled to be 1 mm (± 0.02 mm). The H permeations were performed two times for each condition with new specimens, and the representative result was reported. One of the main considerations is stabilizing the surface reactions (occurring during the monitoring of the flux) that can affect the H recombination. The stationary in-fluxes on the entry side for the first and second cycles were comparable in the present study. An augmentation in flux would lead to an enhanced time lag for a given material state. For evaluation of the H permeation results, the time lag method was used considering 63% of the normalized flux. Table 5 presents the results obtained from the H permeation experiment. In Table 5, D eff,1 and D eff,2 are the effective diffusion coefficients for the first and second cycles. Φ and N T are the permeation coefficient/permeability and total trap density.
The results show that D eff,1 and D eff,2 for Fe-0.1C is lower than those for the Fe and Fe-0.02C materials. The permeability is decreased with increasing the carbon content in the alloy. The total trap density in the pure Fe is the lowest value compared with the other specimens. However, the total trap density of Fe-0.02C alloy is bigger than that of pure Fe, and the total trap density of Fe-0.1C alloy is the maximum value among the materials under study. While the increase in the pearlite content is approximately 1% (Fig. 2), the grain size is decreased with increasing the carbon content due to grain growth inhibition in higher content of the pearlite that was located at grain boundaries and triple junctions (Fig. 3). The decrease in grain size with increasing carbon content is in agreement with other findings [35,36]. Therefore, the higher density of the grain boundaries contributes to the higher total trap density in Fe-0.1C alloy, and lower hydrogen diffusion coefficient.

Mechanisms of hydrogen embrittlement susceptibility and mitigation
The hydrogen desorption activation energy of the peak of Fe-0.1C alloy is higher than the two other materials. As explained in "Hydrogen trap site studies" section, the hydrogen trap sites in the alloy comprise  interstitial lattices, dislocations and high-angle grain boundaries. In Fig. 3, it was shown that there is pearlite at the grain boundaries and triple junctions in Fe-0.02C and Fe-0.1C alloys, with a higher amount of them in Fe-0.1C alloy ( Fig. 2 and 3). However, ferrite-carbide interface and carbide itself do not have a significant contribution to the trap sites due to small volume fraction in the alloys. Instead, it significantly affect grain size by inhibiting grain growth during soft annealing (see "Materials" section). Among the mentioned trap sites, grain boundaries showed higher desorption energies according to several reports [9,21,22,34]. Figure 10a shows the correlation between the HES index and the normalized H content in unit length of HAGB. For measuring the normalized H content in unit length of HAGB, we first estimated the total length of the grain boundaries (from the grain size distribution in Fig. 2). Then we divided the total hydrogen content (Table 4) by the total length of the grain boundary. The H content normalized by HAGBs length for the Fe-0.02C specimen is the maximum value among the specimens. This result indicates that in the Fe-0.02C specimen, a high hydrogen accumulation at HAGBs leads to providing a critical H content to crack to nucleate [37] easier than the two other materials. In pure Fe, since the total hydrogen content is lower than those for the other materials, normalized desorbed H content per HAGB length is considerably lower than the Fe-0.02C specimen. The fracture surface of the specimens tested in a hydrogen environment contains quasi-cleavage fracture features with serrated markings. As described in "Hydrogen trap site studies" section, there is a higher contribution of grain boundaries to the hydrogen trapping in the Fe-0.1C alloy (due to the peak shifting to the higher temperatures and higher desorption energy of the peak), compared to the two other specimens, leading to the high energy of the desorption peak in Fe-0.1C alloy. Figure 10b and c shows the SEM images of the surfaces after microprinting, showing reduced silver particles in the tensile-tested Fe-0.02C and Fe-0.1C alloys with in-situ hydrogen charging, indicating the locations of the hydrogen trapping in the alloy. This observation confirms the TDS and hydrogen desorption energy calculation findings, implying that grain boundaries act as a relatively strong H trap site. These results are in good agreement with the other results reported in the studies [9,21,34,38]. While it is confirmed that hydrogen is irreversibly trapped at HAGBs (relatively), a critical concentration to be built up at a potential crack initiation site for crack initation [37,39] plays an important role in having different hydrogen embrittlement behavior in Fe, Fe-0.02C and Fe-0.1C materials. In pure Fe, while the total hydrogen content is very low, the H content normalized by HAGBs length is very low accompanied by a low HES index for pure Fe (as can be seen in Fig. 10a). In Fe-0.02C alloy, the total hydrogen content is higher and the normalized H concentration per HAGB length is higher than that for pure Fe (accompanied by a higher HES index for Fe-0.02C alloy). In Fe-0.1C alloy, due to a decrease in grain size by the addition of an effective amount of pearlite, the normalized H concentration per HAGB length is considerably lower than that in the Fe-0.02C alloy. The microprint results in Fig. 10b and c confirm the lower local hydrogen content at HAGBs in the Fe-0.1C than that of the Fe-0.02C alloy. This means a lower possibility to build up the critical amount of local hydrogen at potential flaws (here is HAGB), for a crack to initiate. Figure 10d shows the crack initiation sites in tensile-tested Fe-0.02C specimen stopped at 0.12 strain. The microcracks, HAGBs and LAGBs were indicated in the IPF map. The results indicate that the micro-cracks initiated at HAGBs of ferritic structure.
In Fe-0.02C alloy, the higher H concentration at HAGBs promotes the weakening of the cohesive interatomic bonds, and then, atomic separation during SSRT and leading to crack initiation at HAGBs. This is concluded based on the hydrogen-enhanced decohesion mechanism (HEDE) [40][41][42]. The activation of this mechanism for crack initiation can be confirmed by the observed microcrack initations in Fig. 10d at 0.12 strain. This clarifies that HAGBs is the preferential initiation sites of H-induced fracture in the ultra-low carbon steels.
The fracture surface observations in Fig. 5a are a QC fracture that means the transgranular crack propagation was observed for the Fe-0.02C specimen, and there is not enough indication of intergranular fracture. These results can indicate that the HEDE mechanism was active only in the early stage of the tensile loading in hydrogen containing solution. There is a good correlation between the HES index and normalized H content in unit length of HAGBs in Fig. 10a, indicating meaningful changes in the HES values and the normalized H contents as a function of the alloys. Thus, one can conclude that the obtained results supports the concept: a critical local hydrogen content at HAGBs required for crack initiation [23,37,43]. In Fe-0.02C alloy, the ciritcal local content was built up at HAGBs, and resulting in activation of HEDE mechanism for crack initiation.
As was observed in Fig. 11a-c, the crack is propagated transgranularly. Figure 11a shows that the crack is propagated transgranularly in two grains, whereas Fig. 11c shows the crack propagation planes. {011} slip plane has high contribution to H-enhanced planar slip, and leads to crack propagation dominantly along these planes (Fig. 11b and c) and resulting in a QC flat fracture accompanied by serrated indications (Fig. 5a). It was reported that the cracks propagated along the {011} planes are related to a transgranular crack propagation in hydrogen charged specimen [44,45]. The QC fracture occurred along the {011} planes even in the current study in Fe-C model alloys and other commercially alloyed bcc and bct steels indicates that fracture at the {011} plane is an intrinsic feature of hydrogen-induced QC fracture in bcc/bct steels. The transgranular cracks are also observed in the IPF map in Fig. 11a. The reduction in cross-slip is consistent with the hydrogen-enhanced localized plasticity (HELP) mechanism associated with the shielding model of hydrogen embrittlement [44,46]. The fracture surface of H-charged Fe-0.02C alloy (Fig. 5a) showed dominance of the QC fracture accompanied by serrated markings, indicating the dominance of HELP mechanism in the alloy at the later stage of the tensile test in the presence of hydrogen (leading to failure). The coexistence of HEDE and HELP mechanisms were reported by another study [47].
The fracture surface of the Fe-0.1C specimen contains a limited area of the flat surface showing a limited quasi-cleavage fracture, and the fracture surfaces were covered with a considerable number of dimples. This is mainly because of a lower local hydrogen concentration at HAGBs caused by the smaller grain size and the inhibition of grain growth by the addition of an effective amount of pearlite to the microstructure. This means a lower possibility of accumulating the critical H concentration at HAGBs, resulting in limited activation of HEDE-based crack initiation at HAGB. Consequently, less crack propagation events based on the HELP mechanism in the form of QC fracture occurs (Fig. 6a).

Conclusions
1. The elongation losses of Fe and Fe-0.1C materials tensile tested in hydrogen-containing solution were relatively low, compared with the high elongation loss in Fe-0.02C alloy. These results show that Fe-0.02C alloy has a considerable sensitivity to HE, however, Fe and Fe-0.1C materials showed a relatively lower HE sensitivity.
2. The TDS peak analysis, and the calculation of the H desorption energy of the corresponding peaks revealed that the trap activation energies of the peaks for Fe, Fe-0.02C and Fe-0.1C specimens are 29.5, 30.6, and 51.3 kJ mol −1 , respectively. 3. The desorption energy of the trap site in Fe-0.1C is considerably higher than the energies of Fe and Fe-0.02C specimens. The grain boundaries are responsible for shifting the peak to a higher temperature in Fe-0.1C alloy. The hydrogen trapping by grain boundaries was observed by HMPT mapping. 4. The permeability is decreased with increasing the carbon content in the alloy. D eff,1 and D eff,2 for Fe-0.1C is lower than those for Fe and Fe-0.02C materials. The total trap density in pure Fe is the lowest value among the specimens. However, the total trap density of Fe-0.02C alloy is higher than that of pure Fe, and the total trap density of Fe-0.1C alloy is the maximum value among the materials under study. 5. While the pearlite content is increased in Fe-0.1C alloy, the grain size is decreased with increasing the carbon content due to grain growth inhibition in higher content of the pearlite that was located at grain boundaries and triple junctions. Therefore, the higher density of the grain boundaries con-

Figure 11
The crack propagation path studies for Fe-0.02C specimen in a the EBSD result of the crack showing a transgranular crack propagation in the fractured surface, b EBSD orientation map near the transgranulary fractured surface of the concurrently hydrogen charged specimen after tensile fracture and c the IPF map of the crack propagation path, and d the corresponding cumulative IPF set for the grains close to the transgranulary fractured surface. The dashed yellow lines are the fracture traces that are parallel to {011}.
tributes to the higher total trap density in Fe-0.1C alloy, and lower hydrogen diffusion coefficient. 6. In Fe-0.02C alloy, the higher H concentration at GBs promotes the weakening of the cohesive interatomic bonds, and then, leads to atomic separation by the increment of load during SSRT and finally H-induced crack initiation (HEDE mechanism). The hydrogen-enhanced planar slip leads to crack propagation along the planes and leads to a quasicleavage fracture with serrated markings (HELP mechanisms). 7. In Fe-0.1C alloy, there is less opportunity for the critical H concentration to accumulate at the HAGBs, resulting in limited activation of HEDE based crack initiation at the HAGBs. Consequently, there are fewer crack propagation events based on the HELP mechanism in the form of quasi-cleavage fractures.