2016 Volume 56 Issue 9 Pages 1656-1661
We investigated the suppression mechanism of strain-age-hardening in a ferritic carbon steel associated with hydrogen uptake. We considered hydrogen-related three factors suppressing the strain aging: 1) solution softening, for instance, arising from a reduction in Peierls potential of screw dislocations and a change in Young’s modulus, 2) suppression of dislocation-carbon/nitrogen interaction through hydrogen/carbon and nitrogen site competition, and 3) change in plastic strain evolution behavior by hydrogen-enhanced localized plasticity (HELP). According to the present experiments under in-situ hydrogen charging, it was concluded that the solution softening (factor1) and the site competition (factor2) by hydrogen did not significantly suppress the strain aging but the change in the pre-straining behavior (factor3) did.
Delayed fracture and fatigue fracture in steels cause unpredicted accidents under long-term use. These fracture behaviors have been known to depend on temperature,1,2) frequency,3,4) member shape,5,6) etc. Therefore, we need to understand the critical conditions causing these fractures. In particular, mechanical properties in hydrogen atmosphere have recently attracted attention and have materialized into the hydrogen energy society. Moreover, it has been reported that hydrogen strongly affects the above-mentioned fracture behaviors. For instance, the fatigue crack propagation rate on carbon steels is significantly accelerated by hydrogen,7) and this fact has been noted as a primary barrier to the utilization of various steels under hydrogen atmosphere.
More specifically, it has also been reported that the fatigue limit on carbon steel disappears when hydrogen is introduced.8) The main mechanisms of fatigue limit on carbon steels are plasticity-induced crack closure9,10) and stabilization of the plastic zone around the crack tip because of strain-age-hardening.11) In general, hydrogen embrittlement mechanisms are mostly suggested on the basis of the following underlying phenomena: lattice decohesion,12,13) hydrogen-enhanced strain-induced vacancy formation,14,15) and hydrogen-enhanced localized plasticity (HELP)16,17,18) which affect the dislocation motions. In addition to these basic hydrogen effects, hydrogen has also been reported to affect the behavior of strain aging of carbon,19,20) which has a significant effect on the fatigue limit of carbon steels. In general, the strain-age hardening occurs through solute carbon and nitrogen segregation to mobile dislocations.21) Therefore, we must discuss the effect of hydrogen on the fatigue behavior in terms of strain aging of the carbon and nitrogen in steels. However, we have not comprehensively understood how hydrogen affects the fatigue limit in terms of the influence of strain aging.
We can consider three main hydrogen-related factors suppressing the strain aging on body centered cubic (BCC) steels: 1) solution softening, for instance, arising from a reduction in the Peierls potential of screw dislocations and a change in Young’s modulus,22,23,24) 2) suppression of dislocation-carbon/nitrogen interaction through hydrogen/carbon and nitrogen site competition,25) and 3) change in plastic strain evolution behavior by HELP.16,17,18) Although a considerable amount of research has been conducted on dislocation mobility and stress-strain response of hydrogenated carbon steels,26,27,28) there are no detailed experimental reports dividing the three factors.
Therefore, in this study, we focus on the hydrogen diffusion coefficient that is larger than that of carbon and nitrogen.29,30) Specifically, the three possible hydrogen effects on dislocation-carbon/nitrogen interactions are separately examined by changing the timing of hydrogen charging. First, we investigate the effects of hydrogen charging after the dislocations are fully desorbed from the carbon/nitrogen atmosphere by deformation. Solution softening by the reduction in Peierls potential of screw dislocations and the change in Young’s modulus takes place when hydrogen is in the dislocation core or its vicinity. In other words, we examine the strain-age-hardening behavior under a condition where carbon and nitrogen are not present in or around the dislocation cores before strain aging, and thus, hydrogen easily diffuses into the dislocation cores or its vicinity during strain aging under hydrogen charging. Second, we perform hydrogen charging after strain aging for long time to examine the suppression of dislocation-carbon/nitrogen interaction through hydrogen/carbon and nitrogen site competition. Strain aging for long time provides a condition such that the carbon/nitrogen fully segregate in the dislocation. Namely, based on the comparison between the first and second experimental results, we discuss two hydrogen effects on strain-age-hardening: the effect of pre-existing hydrogen that prevents carbon/nitrogen segregation and the effect of hydrogen entry pushing out the carbon/nitrogen from the dislocation or its vicinity. At last, by pre-hydrogen charging before deformation in carbon steel, we can observe the effect of the change in plastic strain evolution behavior by HELP. The aim of this study is to clarify the hydrogen-suppression mechanism of strain-age-hardening in ferritic carbon steel by considering the carbon/nitrogen coefficient with a controlling hydrogen charge point.
Table 1 shows the chemical composition of low carbon steel (JIS S10C). Compared to the solid solubility limit of carbon (≈0.02 mass%), the amount of nitrogen is significant to consider effects of strain-age hardening. Therefore, the strain-age hardening in the present study stems from a combined effect of carbon and nitrogen. A bar with a diameter of 22 mm was annealed for 1 h at 900 °C and subsequently furnace cooled. The annealed bar was cut into the following dimensions: 15 mm in width, 16 mm in thickness, and 130 mm in length. Then, it was cut into tensile specimens with 0.5 mm thickness by spark erosion. This specimen was polished using a mixture of HF and H2O2 to remove the spark erosion layer. Figure 1 shows the specimen geometry. The initial microstructure of the steel is ferrite with a small amount of pearlite as shown in Fig. 2.
| C | Si | Mn | P | S | Cu | Al | N | Ni+Cr | Fe |
|---|---|---|---|---|---|---|---|---|---|
| 0.13 | 0.22 | 0.39 | 0.01 | 0.02 | 0.09 | 0.01 | 0.0044 | 0.01 | Bal. |
Shape and dimensions of the tensile specimen (unit: mm).
Undeformed microstructure of the steel used.
Nominal stress-strain curve obtained with an extensometer.
The strain aging tests were first conducted at different aging times without hydrogen charging to determine a time period for the accomplishment of strain-age-hardening. Then, with reference to the time-dependence of the strain aging behavior, cathodic hydrogen charging was conducted at various situations. The details of the selected situations for hydrogen charging are explained in section 3.
A pre-strain was introduced by tensile deformation. 10% pre-strain for the strain aging tests was obtained from the displacement data by referring to the relationship between displacement and elongation. The relationship between displacement and elongation was obtained by conducting preliminary experiments using an extensometer. The deformation temperature was 28 °C, which was controlled by a thermostatic chamber equipped with a tensile machine. The tensile tests were carried out at an initial strain rate of 10−3 s−1. We compared each stress increment Δσ in these experiments. Stress increment Δσ can be defined as the difference between the maximum stresses before the aging and subsequent yielding stress after the aging. A plot defining stress increment Δσ is shown in Fig. 4.
(a) Nominal stress-strain curve of the strain aging test. (b) and definition of stress increment Δσ.
Hydrogen charging was performed in 3% NaCl+3 g/l NH4SCN electrolyte with a fixed current density. The aging temperature was 28 °C, which was controlled by the thermostatic chamber equipped with a tensile machine. A platinum counter electrode was employed. In-situ cathodic hydrogen charging was carried out in the thermostatic chamber with a plastic cell containing the electrolyte.32) Figure 5 shows a schematic illustration of the cell. The current densities for the cathodic hydrogen charging were 50 and 100 A/m2. Because the present hydrogen charging condition is the same or stronger than the condition that can markedly induce hydrogen effects,33) it is assumed that a sufficient amount of hydrogen is introduced into the specimen by hydrogen charging.
A schematic illustration of the cell for in-situ hydrogen charging.
Figure 6 shows the summarized result of strain-age-hardening behavior without hydrogen charging. The stress increment is plotted against two-third power of the aging time, which is based on the stress increment-aging time relationship as reported by Cottrell and Bilby.21) The strain-age-hardening behavior of S10C carbon steel satisfies the theoretical relationship between the stress increment and aging time up to 27 h, and stress increment by strain aging was accomplished at 48 h. Thus, in this study, the target aging time was chosen to be 48 h to investigate the effect of hydrogen when dislocations are completely trapped by carbon/nitrogen.
Stress increment Δσ plotted against aging time [s2/3].
First, we discuss the possibility of solution softening by hydrogen. It is possible that solution softening stems from a reduction in the Peierls potential of screw dislocations and a change in the Young’s modulus. Specifically, it is assumed that the reduction in the Peierls potential of screw dislocations is affected by the hydrogen existing in the dislocation core or its vicinity. In order to clarify the hydrogen effect on the Peierls potential requires carbon/nitrogen-free dislocations where the introduced hydrogen can segregate preferentially. These carbon/nitrogen-free dislocations can be created by dislocation depinning from carbon/nitrogen and dislocation multiplication. In other words, the required condition corresponds to the strain aging experiment with hydrogen charging that starts immediately after pre-straining. Figure 7 shows the results of three strain aging experiments: 1) strain aging for 9 h without hydrogen charging, 2) strain aging for 9 h with hydrogen charging at 50 A/m2 immediately after pre-straining, and 3) strain aging for 9 h with hydrogen charging at 100 A/m2 immediately after pre-straining. Our considerations to correlate the results with the effect of solution softening are as follows. Because strain-age-hardening is not accomplished at 9 h as shown in Fig. 6, the dislocation core or its vicinity have sufficient spaces for the entry of carbon, nitrogen, and hydrogen. Furthermore, because the hydrogen diffusion velocity is 2.9 × 104 times faster than that of carbon in ferrite/pearlite steels29,30) (nitrogen diffusion velocity is similar to that of carbon31)), hydrogen can segregate into dislocations without any significant effect on carbon/nitrogen. Subsequently, the hydrogen segregation at dislocations is considered to affect the strain aging behavior. The hydrogen diffusion coefficient of 12% pre-strained carbon steel is approximately 2.0 × 10−10 m2s−1.34) According to the equation l=
Results of strain aging tests. The aging time is 9 h. The hydrogen charging for 9 h was started immediately after unloading.
Next, we discuss the suppression of dislocation-carbon/nitrogen interaction through hydrogen/carbon and nitrogen site competition. The softening phenomena may occur through two types of mechanisms associated with the site competition: (A) the pre-existing hydrogen in dislocations inhibits subsequent carbon/nitrogen segregation, (B) the hydrogen that segregates into the dislocations pushes out the pre-existing carbon/nitrogen from the dislocations. In fact, the experiment result of section 3.2 has already denied the possibility of the former mechanism A). To investigate the possibility of the latter mechanism B), hydrogen was introduced after pre-straining and subsequent aging for 48 h was carried out. After aging, strain-age-hardening was accomplished as shown in Fig. 6, indicating that carbon/nitrogen fully diffuse and segregates into the dislocations during this aging time. In this experiment, the hydrogen charging time was configured to conform the hydrogen charging condition with the experiment of section 3.2. Figure 8 shows the stress increments after strain aging for 48 h with and without hydrogen charging for 9 h in the unloading condition. The stress increment after the strain aging for 48 h without the subsequent hydrogen charging shown in Fig. 8 is smaller as compared with that after strain aging for 48 h with the subsequent hydrogen charging for 9 h. Although the total aging time in the experiment with the hydrogen charging was 9 h longer than that without hydrogen charging, the higher stress increment was not due to the difference between the aging times. This is because the strain-age-hardening was accomplished at 48 h aging time as shown in Fig. 6. This means that the suppression of the strain-age-hardening by hydrogen was not observed in this condition. As mentioned for the experiment with hydrogen charging immediately after pre-straining in section 3.2, the solution hardening of hydrogen is considered to occur, leading to higher stress increment than that without hydrogen charging. Hence, the suppression of the dislocation-carbon/nitrogen interaction through hydrogen/carbon and nitrogen site competition as a factor for suppressing the strain-age-hardening as mentioned in the introduction is also ruled out from the underlying mechanisms for the suppression of strain-age hardening.
Results of strain aging tests. The aging time is 48 h. The hydrogen charging for 9 h was started after aging for 48 h.
In this last section, we discuss the remaining suppression mechanism of strain-age-hardening which was mentioned in the introduction, i.e., the change in plastic strain evolution behavior by HELP. To investigate the HELP effect on the pre-straining behavior, hydrogen must be fully introduced into the specimen before the pre-straining. Subsequently, the hydrogen-charged specimen must be deformed under hydrogen charging to avoid hydrogen desorption from the specimen. Consequently, the deformation under hydrogen charging enables the fabrication of a pre-strained specimen with HELP-altered dislocation distribution, dislocation density, and stress field around the dislocations.14,15,16) From the estimation of hydrogen diffusion distance as mentioned in section 3.2, the hydrogen charging time before pre-straining was determined to be 1 h to provide a macroscopically homogeneous HELP effect in the specimen. In detail, after the hydrogen charging for 1 h, pre-straining and subsequent aging for 9 h were performed under hydrogen charging. Figure 9 shows the result of strain aging for 9 h with and without hydrogen charging. These results indicate that the stress increment by strain aging with pre-straining under hydrogen charging is approximately 5.4 MPa smaller than that without hydrogen charging. We conducted this experiment with hydrogen charging three times, and all of those experiments showed reproducible stress increment values with an error range of ±0.3 MPa. Hence, the reduction in the stress increment by pre-hydrogen charging is concluded to be significant. This fact indicates that the last mechanism out of the three possible suppression mechanisms of strain-age-hardening is suggested as the major factor.
Results of strain aging tests. The 10% pre-strain during hydrogen charging was provided after pre-hydrogen charging for 1 h. The top and bottom results are shown as reference data which are same as Fig. 7.
Moreover, in this section, we discuss how the HELP effect suppresses the strain-age-hardening. We consider the following three HELP-related effects in this section: change in stress field around the dislocations, change of dislocation density, and change in dislocation distribution. It has been reported that the stress field around the dislocations shrinks with hydrogen entry into the specimen.18) This stress field shrinkage may contribute to the suppression of strain-age-hardening. However, note that the stress increments from strain aging did not decrease by hydrogen charging in the experimental conditions shown in sections 3.2 and 3.3. Because the stress field around the dislocations probably shrinks under those conditions, the suppression of strain-age-hardening cannot be explained by the hydrogen effect on the stress field around the dislocations. Further, the effect of increasing dislocation density by HELP is taken into consideration. The dislocation density has been reported to notably increase when a plastic strain is provided in environments with hydrogen17,35) as compared with that without hydrogen. Additionally, it has also been reported that the stress increment by strain aging is not significantly dependent on the pre-strain when the pre-strain provides sufficient number of dislocations for strain-age-hardening, e.g. higher than 10% pre-strain.36) These facts indicate that the increase in the dislocation density by hydrogen-enhanced plasticity does not directly affect the stress increment by strain-age-hardening. Therefore, it is considered that the reduction in the stress increment observed in the pre-hydrogen charged specimen is attributed to the change of dislocation distribution rather than a simple increase of dislocation density. In other words, an inhomogeneous dislocation distribution is important for the suppression mechanism of strain-age-hardening. A possible factor suppressing the strain-age-hardening based on the change of dislocation distribution by HELP is illustrated as follows. The dislocation distribution is known to localize when a pre-strain is provided under the hydrogen effect, i.e. hydrogen-enhanced shear localization.16) As a result, it is assumed that a slip plane having high dislocation density and another slip plane having relatively low dislocation density coexist as shown in Fig. 10. As evident from the result of Figs. 7 and 8, a typical strain aging phenomenon, mobile dislocations on each slip plane are pinned by carbon/nitrogen during the strain aging for 9 h. Similar to this fact, the dislocation pinning by strain aging must also occur in the case of hydrogen charging prior to pre-straining. Therefore, the locally distributed mobile dislocations in the pre-hydrogen-charged specimen are trapped. In this case, a considerable number of trapped dislocations hinder the motion of the remaining or newly formed dislocations. In contrast, when the number of trapped dislocations is less, the untrapped dislocations move easily on the same slip plane compared to the region having a high density of trapped dislocations. When the initial plastic deformation is activated by the dislocation motion on the slip plane having less number of trapped dislocations, the plastic deformation can initiate easily than in the case of the absence of the hydrogen effect where the trapped dislocation density is relatively homogeneous. As for one of the above model, it is considered that the stress increment affected by HELP exhibits low values as compared with that without HELP. Therefore, we consider that the main hydrogen effect suppressing the strain-age-hardening in carbon steel is the change of dislocation distribution associated with HELP.
A schematic image of localized dislocation distribution associated with HELP.
In this study, we concluded the suppression mechanism of strain-age-hardening by hydrogen uptake in carbon steel. The details are as below.
(1) The strain aging tests with the hydrogen-charged carbon steel exhibited solution hardening rather than solution softening.
(2) The strain-age-hardening in the present experiments was not suppressed by the two types of site competitions assumed: A) the pre-existing hydrogen in the dislocations inhibits carbon and nitrogen segregation, B) hydrogen entry into the dislocations pushes out the pre-existing carbon and nitrogen. Therefore, the site competition of hydrogen and carbon/nitrogen does not suppress strain-age-hardening.
(3) A change in the dislocation distribution by HELP during pre-straining is the major cause of strain-age-hardening suppression. Therefore, the main factor deteriorating the fatigue limit in hydrogen environments is considered to be the hydrogen-altered plastic deformation behavior and its associated formation of dislocation structures.
