2015 Volume 55 Issue 1 Pages 335-337
Hydrogen uptake in a ferritic steel was investigated through secondary ion mass spectrometry (SIMS) at 83 K, where hydrogen diffusion is sufficiently suppressed. Additionally, the SIMS was operated with cold trap and Si sputtering to reduce the back ground effect. Thanks to the suppression of hydrogen diffusion during the measurements, the cryogenic SIMS could demonstrate reproducible results which showed a significant difference in hydrogen content between hydrogen-charged and uncharged specimens. Namely, hydrogen in the ferritic steel was successfully detected similarly to austenitic steels.
Hydrogen has been drawing attention as a next-generation clean energy carrier that addresses concerns regarding the exhaustion of existing oil resources and environmental conservation. Practical uses of hydrogen as an energy carrier have been explored, and already achieved in a number of energy-related applications.1,2) Since various parts in the hydrogen systems, e.g. the hydrogen gas tank, pipe etc. are exposed to hydrogen gas, the metallic parts have a potential risk in terms of hydrogen embrittlement (HE).3,4,5) In particular, this negative effect of hydrogen is known to be remarkable in BCC steels compared to FCC materials such as austenitic steels.3,4,6)
An important factor affecting HE is distribution of hydrogen. More specifically, the presence of diffusible hydrogen has a key role on the HE.7) Diffusible hydrogen detection has been conducted through microprint technique,8) silver decoration method,9) and scanning Kelvin probe.10,11) Although these methods have a great advantage which enables to visualize microstructure-scale hydrogen distribution, these detect hydrogen only on a surface. Namely, the detected hydrogen has already had a complex history of diffusion before the hydrogen appears on the surface.
An alternative technique for detecting hydrogen is secondary ion mass spectrometry (SIMS), which visualizes the hydrogen distribution through scanning the specimen surface with Cs ions and the subsequent detection of the secondary ions ejected. SIMS enables direct hydrogen detection in deep regions from the surface. However, due to the low specific weight and solubility of hydrogen compared to typical solute elements such as carbon in steels, the analytical precision is negatively affected by the hydrogen drifting in the device. This is termed as the background (BG) effect.12,13) Reducing the BG effect was realized by using a cold trap and a stage cooling system.14,15) The cold trap absorbed the drifting hydrogen atoms, and the stage cooling system prevented the hydrogen diffusion, enabling a highly accurate hydrogen analysis by SIMS.
The previous works on the hydrogen distribution by using SIMS were conducted in austenitic steels,13,16) or martensitic steel which contains a considerable amount of boundary defects and dislocations.17) In contrast, to our best knowledge, the use of SIMS to detect diffusible hydrogen in ferritic steels has a less number of reports18) owing to the larger diffusion coefficient and lower solubility of hydrogen than those in austenitic steels and martensitic steels. Moreover, one hydrogen-charged specimen cannot show reproducible data, since hydrogen in interstitial sites and weekly trapped hydrogen diffuses out more or less during a measurement which needs a few hours. We believe that cryogenic SIMS using the stage cooling system solves these problems indeed. In this study, we aimed at detecting weakly trapped hydrogen in a ferritic steel through SIMS with the following novel techniques.
1) The cold trap and the stage cooling system were applied to detect hydrogen in mildly hydrogen-charged ferritic steel, which is applicable for practical issues.
2) Cryogenic SIMS was conducted several times in an identical location to demonstrate the reproducibility of the analysis.
3) Raster size was changed to prove existence of a small amount of hydrogen.
A Fe-0.13C-0.39Mn-0.22Si-0.01P-0.02S steel (mass%) with a 10% pre-strain in tension was used in this study. A cylindrical bar with a diameter of 22 mm was a low-carbon steel annealed at 1173 K for 1 hour and subsequently cooled in a furnace. Then, cuboidal specimens with dimensions of 3 × 4 × 6 mm3 were cut mechanically. The central part of the 4 × 6 mm2 side was used for the SIMS-based analysis. This side was mechanically polished to remove the work-hardened layer and facilitate surface observation.
The specimens were charged with hydrogen through immersion into a 20% NH4SCN aqueous solution at 313 K for 48 hours in an isothermal bath. The hydrogen contents of the specimens were measured by thermal desorption spectrometry (TDS) from 313 K to 953 K at a heating rate of 0.3 K s–1. Figure 1 shows hydrogen desorption rates plotted against temperature. The total hydrogen amounts of the hydrogen-charged and uncharged specimens were measured to be 0.142 and 0.019 wt. ppm, respectively. In addition, the cumulative desorbed hydrogen amount from 313 to 673 K, which corresponds to diffusible hydrogen content, was measured to be 0.140 wt. ppm in the hydrogen-charged specimen. These results indicate that the total hydrogen content in the hydrogen-charged specimen is accounted for much by the diffusible hydrogen which exists at interstitial site, vacancy, dislocation, and grain boundary. The diffusible hydrogen amount is much smaller than that in conventional SIMS works (for example17)).
TDS profiles of hydrogen-charged and uncharged specimens.
The SIMS system was an IMS-7f spectroscope (Cameca, France). This system has a Cs ion source, a double-focus sector-type mass analyzer, and a stage cooling system. Cs ion was used as primary ion. An acceleration voltage of the primary ion beam was 15 kV which was the difference of voltages between Cs ion voltage of +10 kV and a specimen voltage of –5 kV. The primary ion beam intensity was constantly maintained at 60 nA during SIMS measurements of all specimens. This stage can cool specimens down to 83 K, limiting the hydrogen diffusion even in the ferritic steel. Immediately after the hydrogen charging, the specimen side to be analyzed was mechanically polished with a diamond wrap. During the SIMS-based analysis, hydrogen-charged and uncharged specimens were placed together in the analysis chamber. Then, cooling started after 3 hours from the hydrogen charging. Note that the liquid-nitrogen-based cold trap and the Si sputtering techniques19) were applied to suppress the BG effect.
First, SIMS analyses were performed over a fixed square raster size (50×50 μm2), which was the scan field of the primary ion beam. The analyzed area was a circle with a diameter of 5.6 μm at the central part of the raster area. The target elemental ion was 1H–. The SIMS profiles of the hydrogen-charged and uncharged specimens were then compared. Next, for the sake of clarity, SIMS analyses were performed while changing the raster size from a square with side of 150 μm to one with side of 50 μm in the same primary ion current; the analyzed area remained a circle in diameter of 5.6 μm at the central part of the raster area. Since primary ion current density depends on raster size, a decrease in raster size increases primary ion current and raises sputter rate. And, it is known that a secondary ion intensity originating from a specimen is a linear function of the primary current density.20) Thus, secondary ion density originating from the specimen rises with a decrease in raster size. Secondary ion density of a certain element originating from background never rises with a decrease in raster size.21) Therefore, the increment of 1H– intensities due to the reduction in raster size must more clearly demonstrate a difference in hydrogen content between hydrogen-charged and uncharged specimens.
Figure 2(a) shows the SIMS profiles with a fixed raster size in the hydrogen-uncharged specimen. There are 4 data items for 1H– and the primary ion beam intensities. The primary ion beam intensities were shown to confirm the degree of stability of the analysis conditions. Any significant differences were not observed among the primary ion beam intensities. The chamber pressure and stage temperature during the analysis were 2.7±0.5×10–10 mbar and 83.6±0.2 K, respectively. These stable experimental conditions indicate that the SIMS-based analyses were performed using the stable sputter rate. 1H– decreased with time. Then, it became stable at around 102 cps. The decrease of intensity is attributed to contaminations such as hydrocarbon.19) Namely, the stable 1H– secondary ion intensity of 102 cps is considered to correlate with the total amount of true hydrogen content (0.019 wt. ppm) and background-originated hydrogen.
SIMS profiles and primary beam intensities of (a) hydrogen-uncharged and (b) charged specimens.
Figure 2(b) shows the SIMS profiles of the hydrogen-charged specimen. There are 4 data items for 1H– and the primary ion beam intensities. The chamber pressure (1.85±0.15×10–10 mbar), stage temperature (83.6±0.3 K), and primary ion beam intensity (Fig. 2(b)) during the analysis were stable, indicating that the sputter rate for the SIMS-based analyses was stably conducted. Here, it is clearly shown that the 1H– secondary ion intensity in the hydrogen-charged specimen is higher than that in the hydrogen-uncharged specimen. As the TDS clarified, most of hydrogen in the hydrogen-charged specimen is diffusible. Although hydrogen in interstitial sites would diffuse out even in cryogenic temperature1, the cryogenic SIMS analysis is considered to detect hydrogen weakly trapped at lattice defects such as dislocation even in the mildly hydrogen-charged ferritic steel. The differential between the 1H– intensities of the hydrogen-charged and the uncharged specimen provides an intensity of the charged hydrogen (0.142 – 0.019 = 0.123 wt. ppm). In this case, it is presumed that the intensity of the background-originated hydrogen is lower than that of the charged hydrogen. From this fact, it is clear that the background-originated hydrogen could be effectively reduced by the cold trap, the cryogenic stage, and the silicon sputtering method. Note that the intensity level did not show a significant difference among the multiple measurements as long as the measurement area does not include characteristic hydrogen trap sites. These results indicate that the cryogenic SIMS analysis can provide quantitatively-reproducible data in a measurement, which is an important advantage particularly in ferritic steels.
Next, a SIMS-based analysis was performed while changing the raster size from a square with sides of 150 μm to one with sides of 50 μm. Figure 3 shows the SIMS profiles of the hydrogen-charged and uncharged specimens. For both the specimens, the raster size was decreased over a period of 1200 s, and then increased again to 150 μm over a period of 1400 s. Since the sputter rate increased with decreasing raster size, the high sputter rate associated with the decrease in raster size can amplify only the signal of 1H– in the specimen. As a result, the 1H– intensity increased with decreasing raster size in both specimens. In case of the raster size of 50 μm (1200–1300 s), the ratio of hydrogen intensity in the hydrogen-charged specimen (≈400 cps) to that in the uncharged specimen (≈100 cps) is approximately 4. On the other hand, in case of the raster size of 150 μm (1000–1100 s), the ratio of hydrogen intensity in the hydrogen-charged specimen (≈100 cps) to that in the uncharged specimen (≈50 cps) is approximately 2. This difference between the ration is attributed to the fact that the higher sputtering rate resulted from the smaller raster size improved the sensitivity for the true hydrogen contained in the specimens as compared to the larger raster size. Therefore, we conclude that the cryogenic SIMS with considerations of raster size enables to discuss true hydrogen uptake in ferritic steels.
Secondary ion intensities rose with a decrease in raster size from 150 μm to 50 μm. The increment of the detectable hydrogen amount in the hydrogen-charged specimen (ΔIH ≈ 420 cps) is four times higher than that in the hydrogen-uncharged specimen (ΔIH ≈ 110 cps).
The major part of this study was conducted as part of the “Fundamental Research Project on Advanced Hydrogen Science” funded by the New Energy and Industrial Technology Development Organization (NEDO).
