Highly sensitive secondary ion mass spectrometric ...

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 9 ( 2 0 1 4 ) 1 1 6 4 e1 1 7 2

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/he

Highly sensitive secondary ion mass spectrometric analysis of time variation of hydrogen spatial distribution in austenitic stainless steel at room temperature in vacuum Tohru Awane a,b,*, Yoshihiro Fukushima a,c, Takashi Matsuo e, Yukitaka Murakami a,d, Shiro Miwa f a Research Center for Hydrogen Industrial Use and Storage, National Institute of Advanced Industrial, Science and Technology (HYDROGENIUS), 744 Moto oka, Nishi-ku, Fukuoka 819-0395, Japan b Kobe Material Testing Laboratory Co., Ltd., 47-13 Niijima, Harima-cho, Kako-gun, Hyogo 675-0155, Japan c Department of Mechanical Engineering Science, Graduate School of Engineering, Kyushu University, 744 Moto oka, Nishi-ku, Fukuoka 819-0395, Japan d International Institute for Carbon-Neutral Energy Research (WPI-I2CNER), Kyushu University, 744 Moto oka, Nishi-ku, Fukuoka 819-0395, Japan e Hydrogen Energy Test and Research Center (HyTReC), 915-1 Tomi, Itoshima, Fukuoka 819-1133, Japan f Cameca Division, AMETEK Co., Ltd., 1-1-30 Shibadaimon, Minato-ku, Tokyo 105-0012, Japan

article info

abstract

Article history:

Hydrogen contained in austenitic stainless steel is classified as diffusible or nondiffusible.

Received 22 April 2013

The hydrogen distribution in austenitic stainless steel changes with time owing to

Received in revised form

hydrogen diffusion at room temperature, and such changes in hydrogen distribution cause

11 October 2013

the mechanical properties of the steel to change as well. It is therefore important to

Accepted 20 October 2013

analyze the time variation of the hydrogen distribution in austenitic stainless steel at room

Available online 22 November 2013

temperature to elucidate the effects of hydrogen on the steel’s mechanical properties. In this study, we used secondary ion mass spectrometry (SIMS), a highly sensitive detection

Keywords:

method, to analyze the time variation of the distribution of hydrogen charged into 316L

Hydrogen

austenitic stainless steel. SIMS depth profiles of hydrogen that were acquired at the three

Secondary ion mass spectrometry

measurement times were analyzed, and the results were compared among the measure-

Time variation

ment times. 1H intensities and distribution of the intensities changed with time due to

Austenitic stainless steel

diffusion of hydrogen in the hydrogen-charged 316L steel sample at room temperature.

Hydrogen embrittlement

Moreover, the time variation of the hydrogen concentration distribution of the hydrogencharged 316L sample was calculated using a one-dimensional model based on Fick’s second law. The time variations of the measured hydrogen intensities and of the calculated values are compared. Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

* Corresponding author. Present address: Research Center for Hydrogen Industrial Use and Storage, Kyushu University (HYDROGENIUS), 744 Moto oka, Nishi-ku, Fukuoka 819-0395, Japan. Tel.: þ81 92 802 3924; fax: þ81 92 802 3894. E-mail addresses: [email protected], [email protected] (T. Awane). 0360-3199/$ e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.10.116


1.

Introduction

Austenitic stainless steel is a metallic material used for fuelcell vehicles or hydrogen infrastructure. When the material is exposed to hydrogen gas, hydrogen atoms enter the material. The hydrogen atoms in the material can degrade its mechanical properties by means of “hydrogen embrittlement,” a phenomenon that has been studied extensively [1e7,22e28]. Murakami et al. analyzed hydrogen contained in a hydrogen-charged 316L austenitic stainless steel using thermal desorption spectroscopy (TDS), and revealed that the hydrogen-charged stainless steel contained both diffusible and nondiffusible hydrogen [3]. Nondiffusible hydrogen is hydrogen that enters the steel at the time of manufacture, and is present even in uncharged (with no hydrogen charging) 316L austenitic stainless steel, with very little diffusion at room temperature. When uncharged 316L stainless steel is exposed to hydrogen gas, hydrogen entering the steel diffuses in the steel at room temperature, so that the concentration distribution of the diffusible hydrogen changes over time. Changes in the concentration distribution of the diffusible hydrogen are likely to cause changes in the mechanical properties of the 316L austenitic stainless steel [1e5,22e28]. It is therefore important to analyze the time variation of hydrogen distribution in 316L austenitic stainless steel to understand the effects of hydrogen on the steel’s mechanical properties. Recently, we have developed a highly sensitive detection method for hydrogen using secondary ion mass spectrometry (SIMS), and used this method to observe the distribution of hydrogen charged into 316L austenitic stainless steel [15]. When hydrogen in a sample is measured with SIMS, not only net hydrogen (HN) in the sample but also backgroundoriginated hydrogen (HBG) is simultaneously detected. The HBG originates from moisture (H2O), hydrocarbons (CxHY), or organic materials (CxHYOZ) existing in the SIMS chamber or on the sample surface [8]. This HBG precludes an accurate measurement of HN because HBG and HN cannot be distinguished in a SIMS profile. Carbon, oxygen, and nitrogen in residual gas in a SIMS chamber or on an inner surface of the chamber are also likely to be background sources when carbon, oxygen, and nitrogen contained in a sample are analyzed with SIMS. There have been several studies of background sources and effective methods to improve

1165

detection limits in SIMS analysis of hydrogen, carbon, oxygen, and nitrogen in semiconductor materials [9e14]. To accurately analyze hydrogen charged into 316L stainless steel samples, further improvements to these methods are needed because even uncharged 316L stainless steel contains hydrogen that enters the steel at the time of manufacture, and because the amount of HBG emitted changes with time during SIMS measurements. However, in past studies, gross intensity of hydrogen (total intensity of HN and HBG) has been merely considered in SIMS analysis of hydrogen in metallic material [19e21]. The original highly sensitive detection method developed by the authors consists of a procedure in which a silicon wafer is sputtered by a SIMS primary ion beam near an analyzed area to reduce HBG in SIMS measurement of hydrogen, and a method to determine the intensities of HBG in measurements of a hydrogen-charged sample by estimating the time variation of hydrogen intensities in measurements of an uncharged sample [15]. In the present study, the time variation of the distribution of hydrogen charged into 316L austenitic stainless steel at room temperature in vacuum is revealed by our SIMS method. In addition, the time variation of the hydrogen concentration distribution in the hydrogen-charged 316L sample is calculated using a one-dimensional model on the basis of Fick’s second law. The time variations of the measured hydrogen intensities and of the calculated values are compared.

2.

Experimental procedure

2.1.

Samples [15]

The 316L austenitic stainless steel (iron-base, 0.01 wt% C, 0.53 wt% Si, 0.77 wt% Mn, 0.023 wt% P, 0.001 wt% S, 12.13 wt% Ni, 17.16 wt% Cr, and 2.86 wt% Mo) was used for this study. Fig. 1(a) shows two rod-shaped samples manufactured from the 316L steel with a diameter of 5 mm and length of 40 mm. One sample (H-PRECHARGE-0) was treated by ultrasonic washing with acetone and then exposed to hydrogen gas at 10 MPa and 250 C for 192 h with the hydrogen exposure facility HYDROGENIUS [17]. After hydrogen exposure, the bulk hydrogen concentration was 22.4 mass ppm. Another sample (NON-CHARGE-0) was prepared to estimate HBG in SIMS measurements of H-

Fig. 1 e Dimensions of the 316L austenitic stainless steel samples [15].

1166


PRECHARGE-0. NON-CHARGE-0 was not exposed to hydrogen and had a bulk hydrogen concentration of 1.7 mass ppm. Both the bulk hydrogen concentrations in H-PRECHARGE-0 and NON-CHARGE-0 were analyzed by TDS. A disk of approximately 3 mm thickness was cut from both samples for SIMS measurements. The disks were named H-PRECHARGE-01 and NON-CHARGE-01, respectively. After the cutting, crosssections of H-PRECHARGE-01 and NON-CHARGE-01 were smoothly polished for 44 min with diamond lapping films. After polishing, H-PRECHARGE-01 and NON-CHARGE-01 were washed in acetone using ultrasonic washing for 1 min Fig. 1(b) and (c) shows the shapes and the dimensions of H-PRECHARGE-01 and NON-CHARGE-01 after polishing.

2.2. SIMS data acquisition from H-PRECHARGE-01 and non-charge-01 SIMS analyses in this research were performed with an IMS7f mass spectrometer (Cameca, France). The IMS-7f is a SIMS instrument equipped with a cesium ion source and a doublefocus sector-type mass analyzer. The measured element in the depth profiles with the SIMS was 1H. In measurements of the time variations of 1H intensities of H-PRECHARGE-01, the finish time of the first set of SIMS measurements of HPRECHARGE-01 was defined as 0 h. Then, H-PRECHARGE-01 data were acquired 27.2 h after 0 h (which was the finish time of the second set of SIMS measurements) and at 484.3 h after 0 h (which was the finish time of the third set of SIMS measurements). Measurements of NON-CHARGE-01 were obtained before and after measurements of H-PRECHARGE-01 to investigate the time variation of HBG in the measurements for H-PRECHARGE-01. H-PRECHARGE-01 measurements were carried out at measurement points of different distances from a cross-section edge to estimate the spatial distribution of hydrogen in H-PRECHARGE-01. The time required to obtain one measurement point for H-PRECHARGE-01 was 1934 s, and for NON-CHARGE-01 it was 664 s. Csþ was used as the primary ion in SIMS measurements. The acceleration voltage of the primary ion beam was 15 kV, which corresponded to the difference in voltage between a Cs ion (þ10 kV) and the sample voltage (5 kV). The intensity of the primary ion beam was constantly maintained at 60 nA while measuring all samples. H-PRECHARGE-01 and NON-CHARGE-01 were scanned by the primary ion beam in a range of 50-mm-square, and then an area of 5.6 mm diameter was analyzed. Intensities of 1H were measured with an electron multiplier (EM).

2.3.

Temperature of the SIMS sample stage

In this study, H-PRECHARGE-01 and NON-CHARGE-01 were set in the SIMS chamber vacuum, and then the time variation of the hydrogen distribution at room temperature was analyzed with SIMS. The temperature inside the laboratory where the SIMS is installed is constantly maintained at approximately 21 C by an air conditioner. The temperature of the sample stage in the SIMS chamber becomes constant at approximately 26 C due to the heat emitted from a vacuum gauge.

2.4. Correction method for sensitivities for hydrogen changing with time When SIMS measurements for an element that stably and uniformly exists in a sample are conducted over time, detected intensities of the element may vary even though the sample is analyzed under identical measurement conditions. Differences in the detected intensities are mainly attributed to variation in the sensitivity of the EM used as the detector for secondary ions. In this research, to correct differences of detection sensitivities for 1H among the 0, 27.2, and 484.3 h measurements, the following method was applied. A silicon wafer implanted with hydrogen (HeSi) [15] was measured with a depth profile method before the measurements of HPRECHARGE-01 and NON-CHARGE-01 at each measurement time. The implanted hydrogen atoms (HIMP) exist at a certain depth in the HeSi. Since the HIMP hardly shifts in the silicon at room temperature, the hydrogen concentration in the HIMP layer is always constant [29]. In this study, a peak intensity in the depth profile of 1H emitted from the HIMP layer at 0 h was used as a reference to correct detection sensitivities for 1H at 27.2 and 484.3 h. Correction coefficients for 27.2 and 484.3 h were obtained by dividing the peak intensity at 0 h by the peak intensities at 27.2 and 484.3 h, respectively. The correction coefficient was 1.48 for 27.2 h and 1.19 for 484.3 h.

2.5. Techniques for HBG reduction and HBG correction in SIMS measurements [15] To reduce HBG in the measurements of H-PRECHARGE-01 and NON-CHARGE-01, the silicon sputtering method was conducted before acquiring SIMS measurements of hydrogen contained in H-PRECHARGE-01 and NON-CHARGE-01 at each measurement time. In the silicon sputtering method, a wafer made from highly pure silicon is sputtered by a primary ion beam near an analyzed area. Since sources of the HBG, such as moisture or hydrocarbon existing on a sample surface or in the SIMS chamber, are covered with the sputtered silicon, HBG emission from the source is inhibited. HeSi was used for the silicon sputtering method. The sputtering time of silicon was 54,137 s at 0 h, 7974 s at 27.2 h, and 15,368 s at 484.3 h. Moreover, a cold trap using liquid nitrogen was used to reduce the HBG during all measurements. The HBG in SIMS measurements of H-PRECHARGE-01 were acquired by the following procedure. The measurements for NON-CHARGE-01 before and after the measurements for HPRECHARGE-01 were conducted to investigate time-variation of the HBG in the measurements for H-PRECHARGE-01. The average value of the 1H intensities for NON-CHARGE-01, which were acquired before the measurements of H-PRECHARGE-01, was calculated (I1). The average value of the 1H intensities for the NON-CHARGE-01, which were acquired after the measurements of H-PRECHARGE-01, was also calculated (I2). It is assumed that the time-variation of the HBG in the measurements of H-PRECHARGE-01 varied along the power approximation curve between the I1 and the I2 on the plot of time versus secondary ion intensity. The intensities of HBG in the measurements of H-PRECHARGE-01 were determined by inserting the finish times of the measurements of HPRECHARGE-01 into the approximation curve.


3.

Results and discussion

3.1. Depth profile of 1H of H-PRECHARGE-01 and NONCHARGE-01 with SIMS Fig. 2(a)e(c) shows the depth profiles of 1H that were acquired from H-PRECHARGE-01 measurement points using SIMS (Fig. 2(a): at 0 h, Fig. 2(b): at 27.2 h, Fig. 2(c): at 484.3 h). The intensities of 1H detected from the near-surface region were comparatively high mainly due to hydrocarbons and moisture attached to the cross-section surface at any measurement time. The profiles increased approximately linearly from the depths of the minimal intensities to depths of approximately 10 mm for all measurement times. The intensities were almost constant at regions deeper than approximately 10 mm at 0 h (Fig. 2(a)). We deduced that the hydrogen contained in the region from the surface to approximately 10 mm deep diffused due to frictional heat generated by cutting and polishing HPRECHARGE-01, so that most hydrogen in that region escaped from the cross-section. The depth profiles of 1H other than the H-B7 at 27.2 h (Fig. 2(b)) show trends similar to those at 0 h: the 1H intensities in the depth profiles at 27.2 h became almost constant at regions deeper than approximately 10 mm. The depth profiles of 1H at 484.3 h (Fig. 2(c)) increased approximately linearly from the depths of the minimal intensities to the bottom of the sputter crater. The depth profiles of 1H at 484.3 h show a concentration gradient of hydrogen that occurred with the progression of hydrogen diffusion with time in the measured depth range (0 to approximately 25 mm) of H-PRECHARGE-01.

1167

Fig. 2(d)e(f) shows the depth profiles of 1H that were acquired from NON-CHARGE-01 measurement points using SIMS (Fig. 2(d): at 0 h, Fig. 2(e): at 27.2 h, Fig. 2(f): at 484.3 h). The intensities of 1H became almost constant at a depth of approximately 3 mm for all measurement times. The average intensity of 1H from 3 to 6 mm deep in NON-CHARGE-01 (0 h: 26.8e74.5 cps, 27.2 h: 19.1e35.2 cps, 484.3 h: 30.0e55.8 cps) was lower than the minimal intensity of 1H acquired from HPRECHARGE-01 (0 h: 462.2e1140.1 cps, 27.2 h: 207.4e430.2 cps, 484.3 h: 57.8e230.1 cps) at each measurement time. Thus it is found that HBG could effectively be sufficiently reduced to estimate the significant difference of the intensities of 1H between H-PRECHARGE-01 and NON-CHARGE-01 at any measurement time.

3.2. Determination of the intensities of HN in Hprecharge-01 Table 1(a)e(c) shows the average intensities of 1H from 10 to 16 mm deep, and the pressures in the vacuum system just before the measurements for H-PRECHARGE-01 together with the averages (Ave), standard deviations (SD), and variation coefficients (VC ¼ SD/Ave). Table 1(d)e(f) shows the average intensities of 1H from 3 to 6 mm deep, and the pressures in the vacuum system just before the measurements for NONCHARGE-01 together with the Ave, SD, and VC values. The 1 H intensities at 27.2 and 484.3 h were corrected by multiplying the correction coefficients for the hydrogen sensitivities, which were 1.48 and 1.19, respectively. The chamber vacuum temperature remained almost constant through all

Fig. 2 e Depth profiles of 1HL acquired from H-PRECHARGE-01 and NON-CHARGE-01 using SIMS. (a) H-PRECHARGE-01 at 0 h [15]. (b) H-PRECHARGE-01 at 27.2 h (c) H-PRECHARGE-01 at 484.3 h (d) NON-CHARGE-01 at 0 h [15]. (e) NON-CHARGE-01 at 27.2 h (f) NON-CHARGE-01 at 484.3 h.

1168


Table 1 e Average intensities of 1HL and pressures in the vacuum system just prior to SIMS measurements of HPRECHARGE-01 and NON-CHARGE-01, together with averages (Ave), standard deviations (SD), and variation coefficients (VC [ SD/Ave). (a) H-PRECHARGE-01 data at 0 h, together with distances from the cross-section edge [15]. (b) H-PRECHARGE-01 data at 27.2 h, together with distances from the cross-section edge. (c) H-PRECHARGE01 data at 484.3 h, together with distances from the crosssection edge. (d) NON-CHARGE-01 data at 0 h [15]. (e) NONCHARGE-01 data at 27.2 h (f) NON-CHARGE-01 data at 484.3 h. The 1HL intensities at 27.2 and at 484.3 h were corrected by multiplying the correction coefficients for the hydrogen sensitivities, which were 1.48 and 1.19, respectively. (a) Distance (mm) 10e16 mm deep Chamber vacuum ( 1010 Torr) 1 H (cps) H-A1 H-A2 H-A3 H-A4 H-A5 H-A6 Ave SD VC

0.07 0.54 1.01 1.40 1.67 2.46

1128 1160 2342 4314 2879 3275 2516 1135.4 0.451

2.6 2.6 2.8 2.8 3.1 2.8 2.78 0.167 0.060

10e16 mm deep

Chamber vacuum ( 1010 Torr)

(b) Distance (mm)

1

H-B1 H-B2 H-B3 H-B4 H-B5 H-B6 H-B7 Ave SD VC

H (cps) 1.48

0.14 0.42 0.81 1.34 1.73 1.58 2.00

678 644 821 959 900 2350 876 1033 548.4 0.531

2.7 2.5 2.5 2.6 2.5 2.4 2.4 2.51 0.099 0.039

Distance(mm)

10e16 mm deep


(c)

H (cps) 1.19

1

H-C1 H-C2 H-C3 H-C4 H-C5 H-C6 Ave SD VC

1.46 0.16 0.73 1.52 1.31 1.82

591 224 239 711 199 829 466 254.7 0.547

2.9 2.9 3.1 2.9 3.0 3.0 2.97 0.075 0.025

(d) 3e6 mm deep H (cps)

1

U-A1 U-A2 U-A3

66 75 44

Chamber vacuum ( 1010 Torr) 2.6 2.5 2.6

Table 1 (continued) (d) 3e6 mm deep H (cps)

1


Ave. SD Variation coefficient

62 (Id1) 12.8 0.207

2.57 0.047 0.018

U-B1 U-B2 U-B3 U-B4 Ave. SD Variation coefficient

27 28 30 44 32 (Id2) 6.8 0.212

3.0 3.0 2.8 2.8 2.90 0.100 0.034

3e6 mm deep


(e)

1

H (cps) 1.48

U-C1 U-C2 U-C3 Ave. SD Variation coefficient

52 39 28 40 (Ie1) 9.7 0.246

2.4 2.4 2.6 2.47 0.094 0.038

U-D1 U-D2 U-D3 Ave. SD Variation coefficient

32 28 33 31 (Ie2) 2.1 0.067

2.4 2.3 2.5 2.40 0.082 0.034

3e6 mm deep


(f)

1

H (cps) 1.19

U-E1 U-E2 U-E3 Ave. SD Variation coefficient

66 45 42 51 (If1) 10.7 0.209

2.8 2.9 3.0 2.90 0.082 0.028

U-F1 U-F2 U-F3 Ave. SD Variation coefficient

58 40 36 44 (If2) 9.6 0.215

2.7 2.8 3.0 2.83 0.125 0.044

the measurements of H-PRECHARGE-01 and NON-CHARGE-01 as shown in Table 1(a)e(f) Therefore, we concluded that the change in HBG during the measurements of H-PRECHARGE-01 and NON-CHARGE-01 was not caused by moisture or hydrocarbons in the chamber vacuum. We estimated the distributions of HN in H-PRECHARGE-01 using the average intensities of 1H from 10 to 16 mm deep in Table 1(a)e(c) However, since the average intensities in Table 1(a)e(c) included intensities of HBG as well as the intensities of HN, it is necessary to subtract the intensities of HBG from the average intensities (the gross intensity of 1H) for determinations of HN. The intensities of

1169


HBG in the measurement of H-PRECHARGE-01 were determined by estimating the time variation of hydrogen intensities in the measurements of NON-CHARGE-01, which were acquired before and after the measurements of H-PRECHARGE-01 for each measurement time [15]. It is assumed that the time variations of HBG in the measurements for HPRECHARGE-01 varied along power approximation curves between Id1 and Id2 at 0 h (Table 1(d)), between Ie1 and Ie2 at 27.2 h (Table 1(e)), and between If1 and If2 at 484.3 h (Table 1(f)) on the plots of time versus secondary ion intensity. The intensities of HBG (IHBG) in the measurements of H-PRECHARGE01 were determined by inserting the finish times of the measurements of H-PRECHARGE-01 into the approximation

Table 2 e The determined intensities of HBG (IHBG), the determined net intensities of 1HL (IHN), and the distances from the edge in the measurements of H-PRECHARGE-01, together with averages (Ave), standard deviations (SD), and variation coefficients (VC [ SD/Ave). (a) HPRECHARGE-01 data at 0 h, (b) H-PRECHARGE-01 data at 27.2 h, (c) H-PRECHARGE-01 data at 484.3 h.

curves. The intensity of HN (IHN) at each measurement point of H-PRECHARGE-01 at each measurement time was determined by subtracting IHBG from the average intensity of 1H from 10 to 16 mm deep. Table 2(a)e(c) shows the determined IHBG values, the determined IHN values, and the distances from the edge of the SIMS measurement points of H-PRECHARGE-01, together with the Ave, SD, and VC values.

3.3. Time variation of the distribution of HN in HPRECHARGE-01 Fig. 3(a)e(c) shows the relationships between IHN and the distance from the cross-section edge of H-PRECHARGE-01. The trends observed in the plots for the three measurement times differ, which is mainly attributed to the diffusion of hydrogen over time in H-PRECHARGE-01 at room temperature. Such

5000 At 0h

a

At 27.2h

b

At 484.3h

c

4000

(a) Distance (mm)

10e16 mm deep

3000

IHN (cps)

IHBG (cps) 0.07 0.54 1.01 1.40 1.67 2.46

35 34 34 33 33 33 34 1.0 0.028

1093 1126 2308 4281 2847 3242 2483 1136.2 0.458

10e16 mm deep

IHN (cps)

(b) Distance (mm)

IHBG (cps) H-B1 H-B2 H-B3 H-B4 H-B5 H-B6 H-B7 Ave SD VC

0.14 0.42 0.81 1.34 1.73 1.58 2.00

32 32 32 32 31 31 31 32 0.4 0.012

646 612 789 927 869 2319 845 1001 548.6 0.548

10e16 mm deep

IHN (cps)

2000 1000 0 2500

2000

I HN (cps)

H-A1 H-A2 H-A3 H-A4 H-A5 H-A6 Ave SD VC

1500 1000 500 0 1000 800 600

(c) Distance(mm)

IHBG (cps) H-C1 H-C2 H-C3 H-C4 H-C5 H-C6 Ave SD VC

1.46 0.16 0.73 1.52 1.31 1.82

45 45 45 45 45 44 45 0.3 0.006

400 200

546 179 194 667 155 785 421 254.8 0.605

0 0

0.5

1

1.5

2

2.5

Distance from the edge (mm) Fig. 3 e Relationship between distance from the edge of the sample and the net intensity of 1HL (IHN). (a) At 0 h [15]. (b) At 27.2 h. (c) At 484.3 h.

1170


Fig. 4 e SEM micrographs of typical microstructures of the 316L stainless steel [15]. (a) Microstructures of the NON-CHARGE01. (b) Magnified picture of Fig. 4(a).

et al. [18]. It is assumed that H-PRECHARGE-01 is an infinite plane with a thickness of 2a. At 0 h (Time: t ¼ 0), it is assumed that hydrogen atoms are uniformly distributed in H-PRECHARGE-01 at a concentration C0. When hydrogen atoms in HPRECHARGE-01 diffuse toward the cross-sectional surface from t ¼ 0, and the thickness direction of the H-PRCHARGE-01 corresponds to the x-axis, the hydrogen concentration distribution can be expressed as the solution of the following onedimensional diffusion equation on the basis of Fick’s second law:

vC vt

¼

D

x

v2 C vx2

(1) t

where D is the diffusion coefficient of hydrogen. To solve Equation (1), the following initial and boundary conditions (Equations (2)e(4)) are given.

3000

Average I HN (I ) (cps)

changes in the concentration distribution of the diffusible hydrogen are likely to cause changes in the mechanical properties of the 316L austenitic stainless steel, such as hardness [1,2], fatigue crack growth rate [3e5] or tensile property [22e28]. On the other hand, the IHN values of the inner region from 1.0 to 2.5 mm are higher than the IHN values of the region near the cross-section edge from 0 to 1.0 mm at any measurement time. The relative relationship of the IHN values between the inner region and the region near the crosssection edge at 0 h was maintained at 27.2 and 484.3 h. This is mainly attributed to the fact that the amounts of escaping hydrogen were nearly consistent for the entire range of depths from 10 to 16 mm. The peak IHN value at 0 h occurred at 1.40 mm (H-A4) and the peak IHN value at 27.2 h occurred at 1.58 mm (HeB6). These results might suggest that hydrogen trapping sites existed in the region from 1.40 to 1.58 mm. In the previous study, we observed d-ferrite grains (length: 50e100 mm, width: several micrometers) in microstructure of the 316 L stainless steel using SEM as shown in Fig. 4 [15]. M.L. Luppo et al. studied effects of d-ferrite on hydrogen embrittlement of austenitic stainless steel welds using hydrogen microprint technique, so that they concluded that the dferrite/austenite interfaces act as hydrogen trapping sites [30]. If content of d-ferrite grain contained in the area of 1.40e1.58 mm of H-PRECHARGE-01 was more than those in the other areas, it is likely that hydrogen concentration in the area of 1.40e1.58 mm was higher those in the other areas. Fig. 5 shows the time-dependent curve of the average of IHN values acquired at all the measurement points of H-PRECHARGE-01 for each measurement time. The average IHN values decreased with time: at 0 h the average was 2482.7 cps, 1001.0 cps at 27.2 h, and 420.9 cps at 484.3 h. This decrease in IHN is attributed to the fact that hydrogen atoms at depths of 10e16 mm diffused, so that the hydrogen concentration in that region decreased. The time variation of the hydrogen concentration distribution as a function of thickness of H-PRECHARGE-01 was calculated using a one-dimensional model proposed by Naoe

2500

I(t) = 1304.1 t -0.15396

2000 1500 1000 500 0 0.001

0.01

0.1

1

10

100

1000

Time (t ) (hr) Fig. 5 e Time-dependent curve of average IHN values acquired at all measurement points of H-PRECHARGE-01 for each measurement time.


Cðx; 0Þ ¼ C0

(2)

where it is assumed that hydrogen atoms diffusing from the inner region immediately convert to the gaseous phase at the cross-sectional surface and that gaseous hydrogen is emitted from the cross-sectional surface. Cð0; tÞ ¼ 0

(3)

where gaseous hydrogen is emitted from both sides of the plane of H-PRECHARGE-01 (x ¼ 0 and x ¼ 2a). The position x ¼ a, which is the half-thickness position of H-PRECHARGE-01, is assumed to be a reflection point for diffusing hydrogen atoms: vcða; tÞ ¼0 vx

(4)

The solution of Equation (1) solved with the initial and boundary conditions of Equations (2)e(4) can be expressed as follows: Cðx; tÞ ¼

N 4C0 X 1 ð2n þ 1Þpx p2 2n þ 1 sin exp 2 Dt 2a 2 p n¼0 2n þ 1 a (5)

The time variation of the hydrogen concentration distribution in H-PRECHARGE-01 was calculated with Equation (5). Since the thickness of H-PRECHARGE-01 is 2.94 mm as shown in Fig. 1, a ¼ 1.47 103 m. Sakamoto et al. reported that the hydrogen diffusion coefficient of a type 304 single-phase austenitic stainless steel at room temperature (30 C) is 2.63 1016 m2/s [16], and this coefficient was substituted into Equation (5). The ratio of the hydrogen concentration at 27.2 h (t ¼ 97,920 s), C27.2, to the initial hydrogen concentration C0 (C27.2/C0) and the ratio of the hydrogen concentration at 484.3 h (t ¼ 1,743,480 s), C484.3 to C0 (C484.3/C0) were calculated for x ¼ 3, 5, 10, and 15 mm. For comparison, a variation of the average IHN at 27.2 h (I27.2 ¼ 1001.0 cps) to the average IHN at 0 h (I0 ¼ 2482.7 cps) was calculated (I27.2/I0 ¼ 40.3%). A variation of the average IHN at 484.3 h to the average IHN at 0 h (I0 ¼ 2482.7 cps) was also calculated (I484.3/I0 ¼ 17.0%). Fig. 6 shows the relationship between elapsed time and variations

Fig. 6 e Changes in average IHN values and in calculated hydrogen concentrations at various H-PRECHARGE-01 sample depths with increasing time. (C0 is defined as the initial hydrogen concentration.)

1171

of average IHN values together with variations of calculated concentrations. The time variations of the average IHN values were much different from the time variations of the calculated concentrations at x ¼ 10 and 15 mm, which were included in the depth range of the SIMS measurements (x ¼ 10e16 mm), whereas the time variations of the averaged IHN values were in good agreement with the time variations of the calculated concentrations at x ¼ 5 mm. Calculations of the hydrogen concentration distributions were performed under the assumption that the hydrogen concentration was uniform in H-PRECHARGE-01 at 0 h (t ¼ 0). However, the measured hydrogen intensities in the depth range from 0 to 10 mm in HPRECHARGE-01 were lower than the measured hydrogen intensities in the depth range from 10 to 16 mm. The fact that the time variations of the averaged IHN values were in good agreement with those of the calculated concentrations for x ¼ 5 mm is attributed to the difference between the assumed initial hydrogen concentration distribution and the actual initial concentration distribution.

4.

Conclusion

The time variation of the distribution of hydrogen in crosssection samples of hydrogen-charged 316L austenitic stainless steel at room temperature in vacuum was analyzed with SIMS, a highly sensitive hydrogen detection method. A sample rod of 316L stainless steel was exposed to hydrogen gas at 10 MPa and 250 C for 192 h to charge hydrogen into the sample. Then, a cross-section sample (H-PRECHARGE-01) was cut from the hydrogen-charged rod. The 1H intensities of HPRECHARGE-01 and of an uncharged cross-section sample (NON-CHARGE-01) were measured using SIMS. In measurements of the time variations of 1H intensities of H-PRECHARGE-01, the finish time of the first set of SIMS measurements of the hydrogen-charged sample was defined as 0 h. Then data of the hydrogen-charged sample were acquired at 27.2 h (the finish time of the second set of SIMS measurements) and at 484.3 h (the finish time of the third set of SIMS measurements). The time variation of the hydrogen concentration distribution of H-PRECHARGE-01 was calculated using a one-dimensional model on the basis of Fick’s second law, and the calculated values were compared with the experimentally determined values. Our conclusions are summarized as follows. (1) The 1H intensities of NON-CHARGE-01 were lower than the minimal intensities of 1H of H-PRECHARGE-01 for all measurement times. Thus, we could sufficiently reduce HBG to estimate the amount of hydrogen charged into the 316L sample at any measurement time. (2) Changes in the net intensity of 1H (IHN) with increasing distance from the cross-section edge of H-PRECHARGE-01 differed with increasing measurement time. The differences among the three measurement times were mainly attributed to diffusion of hydrogen in H-PRECHARGE-01 at room temperature in vacuum. (3) The IHN values of the inner region from 1.0 to 2.5 mm are higher than those of the region near the cross-section edge from 0 to 1.0 mm for any measurement time. The relative

1172


relation of the IHN between the inner region and the region near the cross-section edge at 0 h was maintained at 27.2 and 484.3 h. (4) The average IHN value decreased with time: at 0 h the average value was 2482.7 cps, 676.4 cps at 27.2 h, and 353.7 cps at 484.3 h. (5) The time variation of the hydrogen concentration distribution in H-PRECHARGE-01 was calculated from the onedimensional model of Naoe et al., which is based on Fick’s second law. The experimentally measured time variations of the average IHN values at depths of 10e16 mm were in good agreement with the time variations of the calculated concentrations at a depth of 5 mm from the sample surface.

Acknowledgments This research has been supported by the NEDO project ‘‘Fundamental Research Project on Advanced Hydrogen Science (2006e2012).” The author gratefully acknowledges the support of the International Institute for Carbon-Neutral Energy Research (WPI-I2CNER) supported by the Japanese Ministry of Education, Culture, Sport, Science and Technology.

references

[1] Godoi W, Kuromoto NK, Guimaraes AS, Lepienski CM. Effect of the hydrogen outgassing time on the hardness of austenitic stainless steels welds. Mater Sci Eng 2003;A354:251e6. [2] Kumar P, Balasuramaniam R. Determination of hydrogen diffusivity in austenitic stainless steels by subscale microhardness profiling. J Alloy Compd 1997;25:130e4. [3] Murakami Y, Kanezaki T, Mine Y, Matsuoka S. Hydrogen embrittlement mechanism in fatigue of austenitic stainless steels. Metall Mater Trans A 2008;39A:1327e39. [4] Murakami Y, Kanezaki T, Mine Y. Hydrogen effect against hydrogen embrittlement. Metall Mater Trans A 2010;41A:2548e62. [5] Kanezaki T, Narazaki C, Mine Y, Matsuoka S, Murakami Y. Effects of hydrogen on fatigue crack growth behavior of austenitic stainless steels. Int J Hydrogen Energy 2008;33:2604e19. [6] Kanno M, Okada H, Itoh G. Detection of gasses evolved from metallic materials during deformation. J Jpn Inst Metals 1995;59:296e302. [7] Shoda H, Suzuki H, Takai K, Hagiwara Y. Hydrogen desorption behavior of pure iron and Inconel 625 during elastic and plastic deformation. Tetsu-to-Hagane 2009;95:573e81. [8] Magee CW, Botnik EM. Hydrogen depth profiling using SIMS e problems and their solutions. J Vac Sci Technol 1981;19:47e52. [9] Homma Y, Ishii Y. Analysis of carbon and oxygen in GaAs using a secondary ion mass spectrometer equipped with a 20 K-cryopanel pumping system. J Vac Sci Technol A 1985;3:356e60. [10] Kobayashi J, Nakajima M, Ishida K. The sources for contaminants in the trace analysis of carbon in GaAs by secondary ion mass spectrometry. J Vac Sci Technol A 1988;6:86e92.

[11] Kobayashi J, Nakajima M, Ishida K. Study of the background sources in the trace analysis of carbon using secondary ion mass spectrometry. J Vac Sci Technol A 1989;7:2542e8. [12] Meuris M, Vandervorst W, Borghs GK. Improved quantification and detection limits for oxygen analysis in Alx Ga1x As/GaAs multilayers with secondary ion mass spectroscopy. J Vac Sci Technol A 1989;6:1663e72. [13] Pivovarov AL, Stevie FA, Griffith DP. Optimization of secondary ion mass spectrometry detection limit for N in SiC. J Vac Sci Technol A 2003;21:1649e54. [14] Miwa S, Nomachi I, Kitajima H. High sensitivity analysis of atmospheric gas elements. Appl Surf Sci 2006;252:7247e51. [15] Awane T, Fukushima Y, Matsuo T, Matsuoka S, Murakami Y, Miwa S. Highly sensitive detection of net hydrogen charged into austenitic stainless steel with secondary ion mass spectrometry. Anal Chem 2011;83:2667e76. [16] Sakamoto Y, Katayama H. The electrochemical determination of diffusivity and solubility of hydrogen in an austenitic type 304 steel. J Jpn Inst Metals 1982;46:805e14. [17] http://unit.aist.go.jp/hydrogenius/. [18] Naoe S, Torikai Y, Akaishi K, Penzhorn RD, Watanabe K, Matsuyama M. Tritium release behavior from SS316 under vacuum. Annu Rep Hydrogen Isot Res Cent Toyama Univ 2008;28:23e31. [19] Kawamoto K, Oda Y, Nuguchi H, Fujii H, Izumi T, Itoh G. Jpn Soc Mech Eng 2007;73:36e43. [20] Saintier N, Awane T, Olive JM, Matsuoka S, Murakami Y. Analyses of hydrogen distribution around fatigue crack on type 304 stainless steel using secondary ion mass spectrometry. Int J Hydrogen Energy 2011;36:8630e40. [21] Kossoy E, Khoptiar Y, Cytermann C, Shemesh G, Katz H, Sheinkopf H, et al. The use of SIMS in quality control and failure analysis of electrodeposited items inspected for hydrogen effects. Corrosion Sci 2008;50:1481e91. [22] San Marchi C, Michler T, Nibur KA, Somerday BP. On the physical differences between tensile testing of type 304 and 316 austenitic stainless steels with internal hydrogen and in external hydrogen. Int J Hydrogen Energy 2010;35:9736e45. [23] San Marchi C, Somerday BP, Tang X, Schiroky GH. Effects of alloy composition and strain hardening on tensile fracture of hydrogen-precharged type 316 stainless steels. Int J Hydrogen Energy 2008;33:889e904. [24] Tsay LW, Lee WC, Luu WC, Wu JK. Effect of hydrogen environment on the notched tensile properties of T-250 maraging steel annealed by laser treatment. Corrosion Sci 2002;44:1311e27. [25] Huang JH, Ho CS. Effect of hydrogen gas on the mechanical properties of a zirconium alloy. Mater Chem Phys 1994;38:138e45. [26] Wang S, Ohnuki S, Hashimoto N, Chiba K. Hydrogen effects on tensile property of pure iron with deformed surface. Mater Sci Eng A 2013;560:332e8. [27] Boitsov IE, Grishechkin SK, Malkov IL, Khabarov YA, Yukhimchuk AA. Physical and mechanical characteristics of EP741 and EP99 high-temperature nickel alloys in highpressure hydrogen gas. Int J Hydrogen Energy 1999;24:919e26. [28] Michler T, San Marchi C, Naumann J, Weber S, Martin M. Hydrogen environment embrittlement of stable austenitic steels. Int J Hydrogen Energy 2012;37:16231e46. [29] Fink D, Krauser J, Nagengast D, Almeida Murphy T, Erxsmeier J, Palmetshofer L, et al. Hydrogen implantation and diffusion in silicon and silicon dioxide. Appl Phys A 1995;61:381e8. [30] Luppo MI, Hazarabedian A, Ovejero-Garcia J. Effects of delta ferrite on hydrogen embrittlement of austenitic stainless steel welds. Corrosion Sci 1999:4187e5103.