ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Special Issue on "Toward Suppression of Hydrogen Absorption and Hydrogen Embrittlement for Steels"
Effects of Alloying Elements on Hydrogen Diffusion in Iron
Tomohiko Omura Hideaki SawadaKenji KobayashiYuji Arai
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML

2021 Volume 61 Issue 4 Pages 1287-1293

Details
Abstract

We investigated the effects of substitutional alloying elements on the microstructure, hydrogen diffusivity, and tensile properties of Fe–X binary ferritic alloys (X = Si, Al, Mn, Cu, Ni, Co, Cr, Mo, V, W, and Ti) in air and under hydrogen charging. We find using X-ray diffraction that these elements, except for Si and Co, cause ferrite lattice expansion. The hydrogen diffusion coefficient D (measured via hydrogen-permeation tests under cathodic charging at 24°C) reduces as a function of the added alloy concentration. The D-value reduction is enhanced more for Ti, Mn and Cr than other elements. This D variation cannot be simply explained based on the lattice expansion effect, which means that D depends on both hydrogen trapping at the expanded internal lattice spaces adjacent to substitutional solute atoms and hydrogen-solute-atoms chemical interactions. As regards the tensile properties obtained based on slow strain rate tests in air and under hydrogen charging, we find that the all elements, except for Al and Co, afford alloy strengthening in air. Under hydrogen charging, Ti, Mn, and Cr addition reduces the fracture elongation, thereby indicating that these elements increase alloy susceptibility to hydrogen embrittlement. The elongation loss due to hydrogen does not depend on the strengthening effects; however, it exhibits good correlation with the observed D-value reduction and the increment in surface hydrogen concentration C0, which is inversely proportional to D. This correlation indicates that substitutional alloying elements act as reversible hydrogen-trapping sites, which supply hydrogen to potential and developing cracks.

1. Introduction

It is well-known that hydrogen causes embrittlement in high-strength steel, Ni-based alloys, and other metals. In this regard, the microstructural or environmental factors affecting hydrogen embrittlement and the underlying mechanisms have been extensively investigated.1,2,3) In the context of alloying element addition to steel or its alloys to improve the strength, corrosion resistance, etc., researchers require a better understanding of the role of alloying elements and their effects on hydrogen-trapping in the resulting alloy for suppressing and preventing hydrogen embrittlement.

The hydrogen-trapping phenomenon can be evaluated via the measurement of hydrogen diffusivity based on the fact that such trapping hinders hydrogen movement.4) In this regard, Hagi has reported the effects of substitutional alloying elements (Al, Si, V, Cr, Mo, Mn, Ni, and Co) on diffusion coefficient D in Fe.5) He demonstrated that D decreases with the types and concentrations of alloying elements and further that D strongly depends on the degree of lattice expansion of the body-centered cubic (bcc) crystal structure of Fe upon alloying. These results imply that hydrogen is present at the interstitial sites around the substitutional atoms that attract hydrogen owing to elastic (mechanical) interactions. In this regard, Haruna et al. investigated the effects of Ni and Cr addition on D,6) and their study demonstrated D values similar to those reported by Hagi. In general, most substitutional alloying elements in the solid-solution state are known to act as hydrogen-trapping sites and form obstacles for hydrogen diffusion in Fe.

In the context of alloying elements and their effects on hydrogen trapping, Kushida and Kudo have investigated the effects of Ni, Cr, and Mo addition on the hydrogen diffusivity and hydrogen embrittlement susceptibility of martensitic steels.7,8,9) They demonstrated that these alloying elements cause a D reduction, probably via a mechanism similar to that observed in the Fe–X binary alloys in Hagi’s study. Kushida and Kudo also showed that adding alloying elements increases surface hydrogen concentration C0, which is inversely proportional to D, thereby resulting in the promotion of hydrogen embrittlement. These results indicate that substitutional alloying elements enhance hydrogen embrittlement because of the resulting increase in C0 under certain hydrogen charging conditions.

In this study, we examined the effects of the addition of a comprehensive series of substitutional alloying elements on the microstructure, hydrogen diffusivity, and tensile properties of the resulting Fe alloys in air and under hydrogen charging via microstructural evaluations, hydrogen permeation tests, and slow strain rate testing (SSRT). To eliminate the complexity induced by other metallurgical factors, we used simple Fe–X binary ferritic alloys. The mechanisms of the alloying effects are discussed from the viewpoints of lattice expansion and interactions between the absorbed internal hydrogen and solute atom, and their effects on the hydrogen embrittlement susceptibility.

2. Experimental Procedure

The chemical compositions of the binary alloys considered in the study are listed in Table 1, with the base material being pure Fe. The effects of Si, Al, Mn, Cu, Ni, Co, Cr, Mo, V, W, and Ti on the resulting Fe alloys were investigated. Here, we note that several alloys contain small amounts of Ti and Al, which eliminate free interstitial C, N and O atoms. All the alloys were melted in a vacuum induction furnace and cast into ingots that were subsequently soaked at 1280°C for 48 h for homogenization of the solute atoms. The ingots were forged, hot-rolled into 15-mm-thick plates, and solution-heat- treated at various temperatures for 8 h. Thermodynamic calculations were used to determine the appropriate solution heat treatment temperatures to obtain a single ferritic phase without any secondary phases including precipitation. For Mn- or Ni-bearing alloys, the heat treatment temperature was set to 600°C. For Mo-bearing alloys, the heat treatment temperature ranged from 850°C to 1000°C. For the base pure Fe and the other alloys, the heat treatment temperature was 800°C.

Table 1. Chemical compositions of alloys used in study (mass%).
MarkCTisolAlAlloying elements
Base0.001<0.0010.016
1.25Si0.0020.0370.019Si: 1.25
2.84Al0.0020.0472.84
0.35Mn0.0030.0380.030Mn: 0.35
2.62Mn0.0020.0300.014Mn: 2.62
1.15Cu<0.001<0.0010.033Cu: 1.15
2.63Ni0.0030.0380.027Ni: 2.63
2.84Co0.0020.0280.003Co: 2.84
5.7Co0.0020.0390.012Co: 5.70
0.92Cr0.0020.0390.014Cr: 0.92
2.37Cr0.0020.0390.031Cr: 2.37
4.56Cr0.0020.0420.018Cr: 4.56
18.75Cr0.0030.0390.029Cr: 18.75
1.69Mo0.0020.0260.001Mo: 1.69
4.23Mo0.0010.0370.028Mo: 4.23
8.4Mo0.0020.0410.019Mo: 8.40
0.93V0.0010.0010.022V: 0.93
2.3V0.0010.0010.013V: 2.30
3.24W<0.0010.0010.034W: 3.24
7.74W<0.001<0.0010.040W: 7.74
0.15Ti0.0010.15<0.001
0.84Ti0.0010.84<0.001
1.41Ti0.0011.41<0.001

P≦0.002, S≦0.001, N≦0.005, O≦0.003

Coupon specimens were cut from the plates for microstructural investigations. The coupons were mounted in epoxy resin, polished with emery paper, and finished to a mirror surface. The surface was etched using a Nital solution containing nitric acid and alcohol, and the etched surface was observed by means of an optical microscope. X-ray diffraction (XRD) analysis was conducted to confirm the presence of secondary phases and to estimate the lattice constants of the bcc crystal structure. Prior to the XRD analysis, the surface of coupon specimens was mechanically ground and chemically polished to remove the deformed layer on the surface. XRD analysis was conducted by using a diffractometer with a Cu Kα target. Based on the positions of (110), (200), (211), and (220) peaks of the bcc crystal structure, the lattice constants of the alloys were calculated. In addition, the presence of peaks from the secondary phases was also examined.

For measuring D, hydrogen permeation tests were performed by using membrane specimens with a 30 mm-diameter and 1.5 mm-thickness. Both sides of the membrane were mechanically and chemically polished, and one side was subsequently electroplated with Ni. The thickness of the Ni layer was 0.2 μm. Hydrogen permeation tests were conducted under hydrogen charging with the use of a Devanathan type double cell10) at 24°C. After the membrane was positioned, the cell contacting the Ni-plated surface (hydrogen detection cell) was filled with 1M NaOH solution. The Ni-plated side was polarized at 0 V versus an Ag/AgCl reference electrode. Saturated KCl solution was used as the internal solution in the reference electrode. Hydrogen absorption occurs in the other cell (hydrogen absorption cell) contacting the bare Fe surface. The current density at the Ni-plated surface (measured by means of a potentiostat) reached a stable low value due to the passivation of Ni. Next, the hydrogen absorption cell was filled with 3% NaCl solution containing 3 g/L NH4SCN and subsequently cathodically polarized. Hydrogen atoms permeating the membrane from the bare Fe surface were oxidized to hydrogen ions on the Ni-plated surface. Here, we note that the hydrogen oxidizing current density J (A/cm2) at the hydrogen detection side represents the instantaneous rate of hydrogen permeation through the membrane. Therefore, J can be used to calculate the hydrogen permeation coefficient JL (A/cm), where L denotes the membrane thickness. In the study, the hydrogen absorption side was initially polarized at −0.9 V (vs. Ag/AgCl) until the stabilization of J, and subsequently, the applied potential was switched to −1.2 V. This switching between −0.9 V and −1.2 V was repeated several times. The J value at −0.9 V was small, whereas that at −1.2 V was large. The permeation transient from −0.9 V to −1.2 V corresponds to a “build-up” process, whereas the transient from −1.2 V to −0.9 V corresponds to a “decay” process. In the study, the variation in J with the permeation time during the build-up or decay processes was fitted to the theoretical curve on the basis of the one-dimensional Fick’s second law. Parameter D was obtained from the build-up or decay transient of J according to Eq. (1), which is the first-term approximation of the Fourier or Laplace transform solutions for diffusion:11,12)   

D= L 2 /( 7.14 t 1/2 ) (1)
Here, t1/2 represents the time at which J reaches the half value of the steady-state current density, J, at the applied potential. As the first permeation transient is affected by the presence of strong irreversible hydrogen traps,13) the cycle of the build-up and decay was repeated two or three times to “fill” the irreversible traps with hydrogen before D measurement. As the decay curves of the investigated alloys exhibited a close agreement with each corresponding theoretical curve,11,12) D was mainly derived from the decay transition in this study.

Next, SSRT was performed to evaluate the role of the absorbed hydrogen on the tensile properties. Round bar tensile specimens or plate tensile specimens with their axis in the rolling direction were obtained from the plates. The round bar specimen had a 4-mm-gauge diameter with a 25-mm gauge length. The plate specimen had a cross section of 2 mm × 2 mm with a 20-mm gauge length. Prior to the test, the gauge section was polished with No. 600 emery paper. SSRT was performed at ambient temperature in a solution identical to that used for the hydrogen permeation tests at a constant applied potential of −1.2 V (vs. Ag/AgCl). The strain rate was 3 × 10−4 s−1 at the cross-head speed referred to the original gauge length. The tensile strength and fracture elongation under cathodic hydrogen charging were measured and compared with the corresponding values in air. The side portions and the fracture surfaces of the specimens were observed by means of a scanning electron microscope (SEM) after SSRT.

3. Results

3.1. Microstructure

Figure 1 shows an optical micrograph of the base pure Fe. The micrograph exhibits a single ferritic phase with no precipitation. In the study, the average grain size of pure Fe was 200 μm. As per our XRD investigations of pure Fe, no secondary phase including precipitation was observed. Other alloys also exhibited a single ferrite phase in the optical micrographs and XRD patterns. The grain size of all alloys ranged from 50 μm to 3 mm. Ni- or Mn-bearing alloys exhibited a smaller grain size (approximately 50 μm) than other alloys because of their lower solution heat treatment temperature (600°C).

Fig. 1.

Optical micrograph of base pure Fe.

Based on the XRD patterns, we next measured the lattice constants of the investigated alloys. The lattice constant of pure Fe was 2.867 × 10−10 m. Figure 2 shows the variation in the lattice constants relative to that of pure Fe upon alloying. We note that every substitutional element except for Si and Co “expands” the bcc lattice. The lattice constant value is most enhanced by the addition of Mo, W, and Ti.

Fig. 2.

Deviation in lattice constant as function of alloy concentration. (Online version in color.)

3.2. Hydrogen Diffusivity

Figure 3 presents the effects of the alloy concentration on D. Parameter D decreases with the addition of all the elements as a function of their concentration. Among the investigated elements, Ti, Mn, and Cr enhance the decrement in D by a greater degree than the other alloying elements.

Fig. 3.

Hydrogen diffusion coefficient D as function of alloy concentration. (Online version in color.)

3.3. Tensile Properties in Air and under Hydrogen Charging

Figure 4 shows the effects of the alloy concentration on the tensile strength (TS) in air. All the alloying elements, except for Al and Co, afford a strengthening effect. In particular, Si, Mn, Ni, Cu, Mo, W, and Ti exhibit the most significant strengthening effect.

Fig. 4.

Tensile strengths vs. alloy concentration after slow strain rate testing (SSRT) in air. (Online version in color.)

Under hydrogen charging, the TS values of all alloys were found to be nearly identical to the corresponding values in air. These results indicate that the presence of hydrogen negligibly affects the alloy strength. On the contrary, the alloy ductility was strongly degraded by hydrogen; Figure 5 shows the appearances of the fractured pure-Fe samples (a)(b) in air and (c)(d) under hydrogen charging. Here, we remark that all the fractured samples of the investigated alloys exhibited transgranular cracking.

Fig. 5.

Fractured specimens obtained after slow strain rate testing (SSRT) (a) (b) in air and (c) (d) under hydrogen cathodic charging.

To assess the degree of deterioration by hydrogen, we next compared the fracture elongation under hydrogen charging was compared with that in air. Figure 6 shows the relative fracture elongation as a function of the alloy concentration. Here, “relative” refers to the ratio of fracture elongation under hydrogen charging to that in air. We note that the addition of alloying elements, except for Al, reduces the relative fracture elongation. These results indicate that the addition of alloying elements increases the alloy susceptibility to hydrogen embrittlement. The detrimental effects of Ti, Mn, and Cr are more significant than the others.

Fig. 6.

Relative fracture elongation vs. alloy concentration after slow strain rate testing (SSRT). (Online version in color.)

4. Discussion

4.1. Effects of Alloying Elements on Lattice Constants

From Fig. 2, it can be observed that all the alloying elements except for Si and Co afford an increase in the bcc lattice constants, which indicates that many alloying elements expand the bcc lattice of Fe. This dilation of the lattice by various substitutional alloying elements has been extensively investigated, and the lattice constant is generally identified to vary linearly with the alloy atomic concentration.5,14,15) In this regard, King has widely reported on the lattice-constant change of many binary systems,14) and proposed the following relationship between the alloy concentration and the deviation in the lattice constant of Fe:   

Δd= d Fe ×lsf×c (2)
Here, Δd denotes the deviation in the lattice constant upon alloying, dFe the lattice constant of pure Fe, lsf the linear size factor (the coefficient of expansion) for each alloying element and c the alloy concentration. In Fe–X systems, the lsf values are +0.0408 for Al, +0.0050 for Co, +0.0143 for Cr, +0.0552 for Cu, +0.00160 for Mn, +0.00843 for Mo, +0.0152 for Ni, −0.0270 for Si, +0.0439 for Ti, +0.0338 for V, and +0.0997 for W.14) The upper limits of parameter c in Eq. (2) are 25 at% for Al, 20% for Co, 9% for Cr, 0.7% for Cu, 10% for Mn, 2% for Mo, 5% for Ni, 10% for Si, 1.8% for Ti, 30% for V, and 7% for W. Figure 7 compares the measured lattice deviations (vertical axis) and the calculated ones based on Eq. (2)14) (horizontal axis). The measured and calculated values exhibit a close agreement, except for the high Mo alloy (8.4 mass% Mo). The 8.4 mass% of Mo exceeds the upper limit of parameter c of Mo concentration (2 at%) in Eq. (2). In this regard, Leslie has suggested another dFe × lsf calculation, yielding +0.035 for Si, +0.031 for Ti, +0.055 for Cr, +0.067 for Mn, +0.008 for Co, and +0.07 for Ni.15) The upper limits of the application of these values are 6 at% for Si, 0.15% for Ti, 6% for Cr, 3% for Mn, 6% for Co, and 3% for Ni. We also confirmed that the theoretical values of Leslie’s equation closely agree with the experimental values in this study, except for the high Cr alloy (20 mass% Cr). The 20 mass% of Cr also exceeds the application limit of Leslie’s equation. This close consistency between the lattice constant measurements of this study and those of conventional researches confirms the validity of our measured values.
Fig. 7.

Relationship between calculated and measured lattice constants. (Online version in color.)

4.2. Parameters Affecting Hydrogen Diffusivity

From Fig. 3, we note that adding alloying elements reduces the D value. These results indicate that substitutional solute atoms interact with internal hydrogen atoms and hinder hydrogen movement. The detrimental effect of the elements on D can be ordered as Co, Ni, V, Cu, Mo, Al, W, Si, Cr, Mn, and Ti. In this regard, Hagi has reported a similar order: Co, Si, Cr, Al, Ni, Mo, and V.5) Moreover, the study reports similar D values ranging from 10−5 to 10−4 cm2/s for binary alloys containing up to 7 at% of alloying elements.5) However, the D values of high-Cr, high-Mn, and Ti bearing alloys in this study exceed the limit of Hagi’s data.

To confirm the lattice-dilation effect on D, we plot the correlation between the lattice constant deviation and D in Fig. 8. The D values of the alloys containing Si, Al, Cu, Ni, Co, Mo, V, and W are located on a single line, which corresponds to Hagi’s data. This result implies that lattice expansion is necessary for elastic (mechanical) hydrogen trapping by substitutional atoms. On the contrary, Ti-, Mn-, and Cr-bearing alloys exhibit a different tendency, thus indicating that only the lattice-dilation mechanism cannot explain all the data. That is, another type of interaction should be considered between hydrogen and the substitutional alloying elements.

Fig. 8.

Influence of deviation in lattice constant on hydrogen diffusivity. (Online version in color.)

In this regard, Pressouyre et al. investigated the trapping parameters of pure Fe containing different amounts of Ti and C from their analysis of hydrogen-permeation transients.13) They reported that several traps with different degrees of reversibility coexisted in the investigated alloys. They demonstrated that Ti substitutional atoms act as reversible hydrogen traps with low interaction energies for hydrogen, whereas TiC particles act as irreversible traps with high interaction energies. The reversible-trapping parameters were calculated from the second permeation transient, whereas the irreversible-trapping ones were calculated from the first transient. The authors presumed that the driving force for the hydrogen attraction of Ti can be of chemical or mechanical origin. They also showed that Fe–Ti binary alloys exhibited lower D values than Fe–Cr or Fe–Ni binary alloys.13) In the present study, a similar tendency was observed for the alloys containing Ti, Cr, and Ni (Fig. 3).

Recent studies have used density functional theory (DFT) calculations to quantify the binding energy of hydrogen with substitutional alloying elements in bcc Fe.16,17) Counts et al. calculated the binding energy of hydrogen for a series of substitutional solutes including 3sp elements, 3d and 4d transition metals,16) and they showed that solute Ti atoms attract hydrogen with a binding energy is 0.08 eV, which compares reasonably with the experimental value estimated by Pressouyre (0.14 eV).13) Counts et al. also suggest that larger binding energies correspond with lower electronegativities (chemical effect) and larger atom sizes (mechanical effect). Sawada et al. also performed similar calculations and distinguished the chemical and the mechanical effects.17) They showed that the main factor determining hydrogen trapping is the chemical interaction between the substitutional solute and interstitial hydrogen.

4.3. Parameters Affecting Tensile Properties in Air and under Hydrogen Charging

From Fig. 4, we note that the addition of alloying elements increase alloy strength, except in the cases of Al and Co. In this regard, the solid-solution strengthening effects of substitutional alloying elements have been widely investigated.18,19) Seono et al. reported the following orders of elements from the lowest to the highest effect in the strengthening of Fe–X binary alloys: Co, Cr, W, V, Ni, Mn Cu, Si, and Mo,18) similar to that reported by Takeuchi.19) We note here that the results of Fig. 4 closely agree with these reports.18,19) Therefore, we posit that the main strengthening mechanism in the alloys is solid-solution strengthening.

From Fig. 6, we note that the addition of alloying elements reduces the relative fracture elongation as per SSRT. These results indicate that adding alloying elements increases susceptibility to hydrogen embrittlement. In general, high-strength steels have a high susceptibility to hydrogen embrittlement.1,2,3,4) However, the ductility loss due to hydrogen (Fig. 6) does not correlate with the strengths of the alloys in air (Fig. 4). For example, Mo- and Ni-bearing alloys exhibit a higher relative fracture elongation, while Ti- and Mn-bearing alloys exhibit a lower ductility (Fig. 6), even in the same strength level in air (Fig. 4). Moreover, the addition of alloying elements leads to D reduction, as shown in Fig. 3. Therefore, we posit a correlation between D and the hydrogen embrittlement susceptibility, as the presence of internal traps interacting with hydrogen can affect the hydrogen embrittlement behavior. Figure 9 presents the relative fracture elongation as a function of D. It can be observed that alloys with lower D values exhibit lower values of the relative fracture elongation. From the hydrogen permeation tests, we next obtained surface hydrogen concentration C0 (mass ppm) using J and D on the basis of Fick’s first law under the steady-state hydrogen permeation condition. Parameter C0 was calculated according to the following Eq. (3):11,12)   

C 0 =1.318    J L/D (3)
Here, J denotes the steady-state hydrogen permeation current density (A/cm2) and L the thickness of the specimen (cm). In this study, the product JL was found to be nearly identical for all the investigated alloys: this is because JL does not depend on the trapping properties of the materials, although it strongly reflects the environmental severity (hydrogen charging condition).4,20) Therefore, according to Eq. (3), C0 is inversely proportional to D. Figure 10 shows the correlation between C0 and relative fracture elongation. In the case of SSRT, C0 corresponds to the absorbed hydrogen concentration in a test specimen under cathodic hydrogen charging. The results indicate that alloys absorbing larger amounts of hydrogen under cathodic charging exhibit lower values of relative fracture elongation. Therefore, in the alloys considered, the main metallurgical factor determining the susceptibility to hydrogen embrittlement is D, i.e., ability to absorb hydrogen. Here, we hypothesize that substitutional solute atoms weakly trap hydrogen and act as hydrogen sources to supply hydrogen to potential and developing cracks during SSRT.
Fig. 9.

Effects of hydrogen diffusion coefficient D on hydrogen embrittlement properties. (Online version in color.)

Fig. 10.

Effects of surface hydrogen concentration C0 on hydrogen embrittlement properties. (Online version in color.)

In this context, Kushida and Kudo have reported that Ni, Cr, and Mo addition to martensitic steel increases C0, which is inversely proportional to D, thereby resulting in the promotion of hydrogen embrittlement.7,8,9) In the alloys considered here, we assume that the substitutional elements play a similar detrimental role. Meanwhile, Pressouyre and Berstein et al. have suggested the dual role of Ti in hydrogen embrittlement process.21,22,23,24,25,26) In the absence of tensile stress, both solute Ti atoms and TiC particles are beneficial for the inhibition of intergranular hydrogen-induced cracking, particularly in the early stages of hydrogen charging.21,22) On the contrary, tensile tests with pre-hydrogen charged specimens showed that increasing the solute Ti degrades the ductility whereas TiC addition is still beneficial.24,25) Solute Ti atoms act as both innocuous hydrogen sinks and detrimental hydrogen sources depending on the test conditions. During tensile testing, reversible traps due to solute Ti act as hydrogen sources that release hydrogen to moving dislocations or flaws and aid in the nucleation of cracks. On the contrary, TiC particles are considered to be strong irreversible hydrogen traps that do not release hydrogen atoms, thereby preventing hydrogen from reaching flaws under every test condition. The solute Ti atoms in this study showed a tendency similar to that reported in the aforementioned studies.

4.4. Effects of Grain Boundary and Grain Size

The preceding discussion considered the effects of substitutional alloying elements. Here, we consider the grain-size and grain-boundary effects of the investigated alloys.

Grain boundaries can act as hydrogen-trapping sites. In this regard, Todaka experimentally demonstrated that 11-mass-ppm hydrogen is absorbed in pure Fe with ultra-fine grains (0.25 μm),27) wherein the sample was prepared by means of a high-pressure torsion (HPT) process and subsequent annealing heat treatment. This process enables the preparation of samples with ultra-fine grains with no dislocation. Under the assumption that all the hydrogen is trapped along the grain boundaries, the absorbed hydrogen in a sample is proportional to the total grain boundary area, which is inversely proportional to the square of the grain size. Therefore, the trapped-hydrogen concentration can be predicted by means of the sample grain size. In our study, we calculated the trapped-hydrogen concentrations at the grain boundary to be 3 × 10−4 mass ppm in the 50-μm-grain-size Ni- and Mn-bearing alloys, and 8 × 10−8 mass ppm in the 3-mm-grain-size base pure Fe. These values are sufficiently lower than the measured values of hydrogen concentration (horizontal axis) in Fig. 10. Therefore, hydrogen trapping at the grain boundaries can be ignored in the present study.

Concerning strengthening effects, as discussed previously, we posit that the main mechanism is solute-solution strengthening in the investigated alloys. However, Ni- and Mn-bearing can exhibit a grain-refinement effect, because they have fine grains (~50 μm) than other alloys. It is well-known that grain refinement increases the alloy strength according to the Hall–Petch relationship:28,29)   

σ y =0.1+6× 10 -4 / d 1/2 (4)
Here, σy denotes the yield strength (GPa) and d the average grain size (m). Parameter σy as per Eq. (4) is 185 MPa for Ni- and Mn-bearing alloys with a grain size of 50 μm, whereasσy is 111 MPa for pure Fe with a grain size of 3 mm. Therefore, the difference in grain size among the investigated alloys can affect their strengths.

As regards susceptibility to hydrogen embrittlement, all the tested samples exhibited transgranular cracking, which indicated that the grain size negligibly affects hydrogen embrittlement susceptibility.

5. Summary

We examined the effects of substitutional alloying elements on the microstructure, hydrogen diffusivity, and tensile properties of Fe–X binary ferritic alloys in air and under hydrogen charging. Our main findings are as follows:

(1) The addition of alloying elements, except for Si and Co, expanded the ferrite lattice. This increment in the lattice constant was more significantly enhanced by the addition of Mo, W, and Ti than other alloying elements.

(2) The addition of alloying elements led to a reduction in hydrogen diffusion coefficient D as a function of the alloy concentration. In particular, the decrement in D was more significantly enhanced by Ti, Mn, and Cr than other alloying elements. This variation in D could not simply be explained based on the lattice expansion effect of the alloying elements. Our results implied that D is influenced by both the hydrogen-trapping effect at the expanded internal lattice spaces adjacent to substitutional solute atoms and the chemical interactions between hydrogen and the solute atoms.

(3) The addition of alloying elements, except for Al and Co, afforded a strengthening of the alloy when subjected to tensile tests in air. Under hydrogen charging, the addition of alloying elements, particularly Ti, Mn, and Cr, decreased the fracture elongation, which indicates that such elements increase the alloy susceptibility to hydrogen embrittlement. The elongation loss due to hydrogen did not depend on the strengthening effects, but exhibited a strong correlation with the decrement in D and increment in surface hydrogen concentration C0, which is inversely proportional to D. This correlation indicates that substitutional alloying elements act as reversible hydrogen-trapping sites (detrimental hydrogen sources) that supply hydrogen to developing cracks.

Acknowledgements

The authors wish to thank Nippon Steel Corporation for allowing the publication of this paper. The assistances of co-workers in Nippon Steel Corporation are gratefully acknowledged.

References
 
© 2021 The Iron and Steel Institute of Japan.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs license.
https://creativecommons.org/licenses/by-nc-nd/4.0/
feedback
Top