- Email: [email protected]
Hydrogen solubility and diffusivity in a barium cerate protonic conductor using tritium imaging plate technique
SOSI-13587; No of Pages 4 Solid State Ionics xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Hydrogen solubility and diffusivity in a barium cerate protonic conductor using tritium imaging plate technique K. Yamashita, T. Otsuka, K. Hashizume ⁎ Interdisciplinary Graduate School of Engineering Science, Kyushu Univ., Kasuga, Japan
a r t i c l e
i n f o
Article history: Received 3 October 2014 Received in revised form 14 February 2015 Accepted 16 February 2015 Available online xxxx Keywords: Tritium Imaging plate technique Proton-conducting oxides Hydrogen solubility Diffusivity
a b s t r a c t A tritium imaging plate technique has been applied to visualize hydrogen distribution and examine hydrogen solubility and diffusivity in a proton-conducting oxide, Y-doped BaCeO3 (BaCe0.9Y0.1O3 − α). Tritium charging of the BaCe0.9Y0.1O3 − α specimens was carried out by a gas absorption method using partially tritiated water vapor (HTO, 3 kPa, T/H ~ 10−6) at temperatures ranging from 673 K to 873 K for a given time. After charging, tritium distributions of the surface and cross section of the halved specimens were visualized using an imaging plate technique. From the tritium concentration and distributions of the surface and cross section, hydrogen solubility and hydrogen (tritium) diffusivity of the BaCe0.9Y0.1O3 − α specimens were determined. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Perovskite-type oxides have received much attention as good protonic conductors under a hydrogen-containing atmosphere at elevated temperatures. These proton-conducting ceramics have been expected to be used as solid electrolytes for various electrochemical devices such as a fuel cells, hydrogen sensors, and hydrogen pumps [1–7]. To date, hydrogen solubility and diffusivity in oxide ceramics have been reported by many investigators using several experimental techniques such as conductivity measurement and thermal analysis (TG-DTA); i.e. hydrogen solubility was first measured by thermal analysis, followed by calculation of diffusivity from the solubility and conductivity data. Although this is a commonly used method, it is not a direct measurement of proton migration; therefore, the proton diffusivity measurement can be affected by other electric carriers in the conductivity measurement and experimental errors of hydrogen solubility in thermal analysis. Alternatively, ion beam microanalysis (SIMS, ERDA and NRA) [8,9] can allow for the measurement of hydrogen solubility and diffusivity directly. Although these methods have superior space resolution of hydrogen in the depth direction at a localized position, the information is generally limited near the surface of the protonic conductors. Therefore, two-dimensional distribution of hydrogen cannot be obtained directly. The tritium imaging plate (TIP) technique uses tritium as a radioactive tracer to directly assess hydrogen solubility and diffusivity from ⁎ Corresponding author at: 6-1 Kasuga-koen, Kasuga-shi, 816-8580, Japan. Tel./fax: + 81 92 583 7544. E-mail address: [email protected] (K. Hashizume).
visualized tritium distribution in tritium-dissolved materials. The imaging plate (IP) technique was originally used in biological and medical fields to measure two-dimensional radioisotope distribution and recently its applications have extended to materials and energy sciences [9–12]. However, this technique has seldom been applied to protonconducting materials such as the perovskite type oxides. In this paper, we report the experimental results of the visualization of hydrogen distribution in BaCe0.9Y0.1O3 − α, which is a well-known perovskite-type protonic conductor, by using the TIP technique. Based on these results, hydrogen solubility and diffusivity were determined for the oxide. 2. Experimental 2.1. Sample preparation and exposure to HTO vapor The BaCe0.9Y0.1O3 − α powder (TYK Co.) was die-pressed into a cylindrical shape under 200 MPa of isostatic-pressure and sintered at 1873 K for 5 h. The densities of the sintered pellets were greater than 98% of the theoretical density. The black surface layers were removed with an abrasive paper. The pellets were 7 mm in diameter and 2 mm in thickness. Fig. 1 shows the experimental apparatus. Using this apparatus, the BaCe0.9Y0.1O3 − α specimens were vacuum annealed at 973 K for 1 h, followed by exposure to ca. 3 kPa of partially tritiated water vapor (HTO, T/H ~ 10−6) at 673–873 K for 5–60 min. After the exposure experiment, the specimens were cooled down to room temperature. To measure the inside of the specimen, its surface was slightly ground with an abrasive paper to remove adsorbed tritium, and the disk was cut in half.
http://dx.doi.org/10.1016/j.ssi.2015.02.008 0167-2738/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
2
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
tritium distribution obtained with the IP technique provides information about tritium existing on and up to 1 μm from the surface. 3. Results and discussion 3.1. Visualization of hydrogen (tritium) distribution
Fig. 1. Schematic of apparatus for HTO vapor exposure of specimens at elevated temperatures.
A low-speed diamond saw (Isomet™, Buehler Ltd.) was used to cut the specimens at a cutting speed of approximately 50 rpm, and IsoCut™ Fluid (Buehler Ltd.) was used as the cooling media. Once cut, the specimens were washed with acetone to remove the coolant. 2.2. Imaging plate technique The tritium-charged specimens were placed on an IP in a dark box in order to expose the IP to β− rays emitted from the tritium adsorbed on and dissolved in the specimen. The IP used was TR-2025 (FUJIFILM Co.), which was suitable for measuring the low energy β− rays and had a wide dynamic range for the radiation energy. Considering the tritium content in the specimen, the duration of IP exposure was 3 h for the surface and 24 h for the cross section. After exposure, the IP was set in an IP Reader (BAS-2500, FUJIFILM Co.) to measure the photostimulated luminescence (PSL) intensity. The space resolution of the obtained intensity was 50 μm. Although PSL intensity is an arbitrary unit, it is proportional to the product of the radiation energy emitted from tritium and the IP exposure time allowing for conversion of PSL intensity to hydrogen concentration in the specimen. Tritium (T1/2 = 12.3 y, β− decay) emits low energy electrons, the maximum and mean energies of which are 18.6 and 5.7 keV, respectively and its maximum range is ca. 1 μm in BaCe0.9Y0.1O3 − α. Therefore, the
The hydrogen (tritium) distribution in BaCe0.9Y0.1O3 − α, which was exposed to HTO vapor, was visualized using the IP technique. Fig. 2 shows the IP images of the specimen surface and its cross section at 773 K. From Fig. 2(a), it was found that hydrogen was distributed uniformly on the specimen surface. Fig. 2(b) shows the time evolution of tritium penetration into the specimen. The tritium distribution in Fig. 2(b) seemed to be rather ambiguous due to the relatively low tritium concentration in the cross section. As seen in Fig. 2(a) and (b), the PSL intensities between the surface and near-surface bulk in the cross section largely differed. The reason for this discrepancy can be derived from the tritium density on the surface and in the bulk. The TIP method is a radiation detection technique for measuring the surface and nearsurface tritium. After tritium loading, the amount of tritium adsorbed on the specimen surface was much higher than that dissolved in terms of the atomic ratio of tritium (hydrogen) and ions of the specimen oxide (H/Oxide). Even in an equilibrium state, the tritium and ion ratios between the surface and near-surface bulk were very different (ca. 1/1 and 0.01/1, respectively). Moreover, the β− rays emitted outside from the surface were absorbed in the IP directly, while the β− rays from the bulk were absorbed by the specimen itself, even within the range of the β− ray (ca. 1 μm). Consequently, the PSL intensity of the specimen cross section became weak, compared with that of the surface. Although the PSL intensity of the cross section was weak, it was sufficient for evaluation of diffusivity and solubility. Fig. 2(c) shows the IP image of standard sample autoradiographic [3H] microscales (RPA506 and RPA507, Amersham Biosciences Co.). The 3H (tritium) activity is arranged in eight layers of polymer, the radioactivities of which range from 3.7 Bq/mg to 4070 Bq/mg. The PSL intensity was originally given as numerical data by the IP reader. The PSL intensity is proportional to the tritium activity or tritium number density over a wide range. The PSL intensities obtained from the specimen were calibrated and converted to tritium activities of the surface or bulk based on that of the standard sample. In determining the tritium activity of the specimen, the absorption and range of the β− rays in the standard sample and oxide specimens were taken into consideration. Finally, we converted the tritium activity to hydrogen number density using the hydrogen isotope abundance ratio (T/H ~ 10−6).
low
a) The surface image (773K, 15min.)
b) The cross section image (773K, 15min., 30min. & 45min.)
high
c) [3H]MISROSCALES (Left; RPA506, Right; RPA507)
Fig. 2. IP images of (a) specimen surface, (b) cross section, and (c) autoradiographic [3H] microscales.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
3.2. Hydrogen surface concentration, solubility, and diffusivity Fig. 3 shows the hydrogen (tritium) concentration on the specimen surface exposed to HTO vapor at 773 K for 5–60 min. The data points in Fig. 3 were determined by averaging the PSL intensities over the entire surface with error bars that included both scattering of the PSL intensities on the surface and dispersion in the radiation detection. The IP is a radiation detection technique for measuring the surface and nearsurface tritium. Therefore, the surface concentration of hydrogen exceeded the number of the ions in the specimen (ca. 1015/cm2). The surface concentration reached a saturated value over a short exposure time (5 min). This suggests that the hydrogen concentration near the specimen surface also increased and reached a saturated value at elevated temperatures. Fig. 4 shows the tritium concentration gradient of the specimen cross section at 773 K for 15 min (as shown in Fig. 2(b)) and its fitted curve. The data points in Fig. 4 were fitted to the curve by the following equation,
8
4x10
8
3x10
8
2x10
8
1x10
0
0
0.2
0.4
0.6
0.8
x (mm)
1
ð1Þ
where c is the tritium concentration, DT is the tritium diffusivity, x is the distance from the surface, and t is the diffusion time. This equation is an analytical solution of the diffusion equation based on a onedimensional, semi-infinite medium model in which the tritium concentration on the surface is constant [13]. As a result of curve fitting, the tritium diffusivity was found to be DT = 7.0 × 10−7 cm2/s at 773 K. In a similar manner to that described above, the diffusivities were determined at 673–873 K. The tritium diffusivity obtained from the study was, 0:34ðeVÞ 2 −4 DT cm =s ¼ 1:0 10 exp − ; kB T
ð2Þ
where kB is the Boltzmann constant and T is the absolute temperature. The diffusivities were plotted as a function of reciprocal temperature in Fig. 5, along with literature data [14,15]. Although the activation energy of diffusivity (0.34 eV) was slightly low, it was comparable to that of the literature data for proton. The diffusivity obtained was smaller than the literature data for proton. Although this discrepancy in the diffusivities could be partly due to the isotope effect between proton pffiffiffi and tritium (DT =DH 1= 3), and/or the difference in the experimental method, the details of the discrepancy were unclear at the present stage. Hydrogen solubility in BaCe0.9Y0.1O3 − α was obtained from the IP data by converting the tritium concentration to hydrogen concentration, which was plotted as a function of HTO exposure temperature,
Hydrogen concentration (/cm 2)
3
Fig. 4. Tritium distribution of the specimen cross section exposed at 773 K for 15 min, and its fitted curve.
! x erfc pffiffiffiffiffiffiffiffi ; 2 DT t
18
1.2x10
18
1x10
17
8x10
17
6x10
shown in Fig. 6. Although the hydrogen solubility data of the specimen was scattered, it was found to be ca. 1.8 mol.% and nearly constant from 673 K to 873 K. The temperature dependence of hydrogen solubility was in agreement with previous thermal analysis data [16], but the solubility itself was slightly smaller than that (~8 mol.%). The reason for the low solubility and scattering might be ascribed to tritium leakage from the specimens. The value of tritium diffusivity at room temperature extrapolated from Eq. (2) was on the order of 10−10 cm2/s: therefore, tritium existing near the surface of the specimen could diffuse out over time between cutting and IP exposure. A preliminary successive measurement by IP of the specimen cross section actually showed a decrease in the PSL intensity after a few days. Thus, the influence of the tritium leakage on the IP measurement could be reduced by decreasing the time between the specimen cutting and IP exposure, allowing for evaluation of and the true solubility by extrapolation. Moreover, the discrepancy would not be derived from the IP technique itself, but from the hydrogen behavior of the specimen. However, in order to improve the reliability of the solubility measurement, further studies are being carried out. The present study focused on the visualization of hydrogen distribution in a typical proton conductor, BaCe0.9Y0.1O3 − α. It was demonstrated
10
Diffusion coefficient (cm2/s)
cðx; t Þ ¼ cx¼0
Tritium concentration (/cm 2)
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
-4
M. Oishi et al. (proton)
10
-5
T. Schober et al. (proton)
0.46 eV 0.48 eV
10
-6
present study (tritium) 0.34 eV
17
4x10
10 0
10
20
30
40
50
60
70
time (min) Fig. 3. Time evolution of hydrogen concentration on the specimen surface at 773 K.
-7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
-1
1000/T (K ) Fig. 5. Arrhenius plot of tritium diffusivity in BaCe0.9Y0.1O3 − α.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
1.6
4
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
0.03
873 K. The tritium diffusivity obtained was somewhat smaller than in the literature data for proton. One reason for this discrepancy can be attributed to isotope effect. The activation energy of tritium diffusivity was 0.34 eV, which was relatively close to previous data for proton. Hydrogen solubility in the specimen converted from tritium concentration was evaluated to be ca. 1.8 mol.% and nearly constant over the temperature range. The temperature dependence of the solubility was in agreement with literature data of thermal analysis, but the solubility itself was lower.
H mol/1mol of BCY10
0.025 0.02 0.015 0.01
Acknowledgments
0.005 0 650
700
750
800
850
900
Temperature (K) Fig. 6. Hydrogen solubility in BaCe0.9Y0.1O3 − α (BCYO10).
that the IP technique is a powerful tool for studying hydrogen dissolution and migration in various protonic conductor materials. 4. Conclusion The hydrogen (tritium) distribution in a BaCe0.9Y0.1O3 − α specimen exposed to the partially tritiated water vapor (ca. 3 kPa) was visualized using a TIP technique, allowing for determination of surface concentration, solubility, and diffusivity of hydrogen. From the visualized tritium distribution on the specimen surface exposed to HTO at 773 K, it was found that hydrogen was uniformly distributed on the specimen surface and its concentration reached a saturated value after a few minutes. The tritium diffusivity and solubility in the BaCe0.9Y0.1O3 − α specimen were determined from its concentration distribution on the cross section of the specimen at a temperature range of 673 K to
This work was supported by the Japan Atomic Energy Agency under the Joint Work contract #24K389 as a part of Broader Approach activities.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
H. Iwahara, T. Esaka, H. Uchida, N. Maeda, Solid State Ionics 3/4 (1981) 359–363. H. Iwahara, Solid State Ionics 28–30 (1988) 573–578. N. Bonanos, K. Knight, B. Ellis, Solid State Ionics 79 (1995) 161–170. T. Yajima, K. Koide, H. Takai, N. Fukatsu, H. Iwahara, Solid State Ionics 79 (1995) 333–337. E.C.C. de Souza, R. Muccillo, Mater. Res. 13 (2010) 385–394. M. Zhang, J. Xu, G. Ma, J. Mater. Sci. 46 (2011) 4690–4694. N. Nguyen, H. Yoon, J. Power Sources 231 (2013) 213–218. T. Ishigaki, S. Yamaguchi, K. Kishio, K. Fueki, H. Iwahara, Solid State Ionics 21 (1986) 239–241. E. Ruiz-Trejo, J.A. Kilner, Solid State Ionics 130 (2000) 313–324. K. Hashizume, J. Masuda, T. Otsuka, T. Tanabe, Y. Hatano, Y. Nakamura, T. Nagasaka, T. Muroga, J. Nucl. Mater. 367–370 (2007) 876–881. K. Hashizume, T. Tanabe, Fusion Sci. Technol. 60 (2011) 1135–1138. K. Hashizume, K. Ogata, M. Nishikawa, T. Tanabe, S. Abe, S. Akamaru, Y. Hatano, J. Nucl. Mater. 441 (2013) S880–S884. J. Crank, The Mathematics of Diffusion, Clarendon Press, 1975. T. Schober, W. Grover Coors, Solid State Ionics 176 (2005) 357–362. M. Oishi, S. Akoshima, K. Yashiro, K. Sato, J. Mizusaki, T. Kawada, Solid State Ionics 179 (2008) 2240–2247. K.D. Kreuer, Solid State Ionics 125 (1999) 285–302.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
Contents lists available at ScienceDirect
Solid State Ionics journal homepage: www.elsevier.com/locate/ssi
Hydrogen solubility and diffusivity in a barium cerate protonic conductor using tritium imaging plate technique K. Yamashita, T. Otsuka, K. Hashizume ⁎ Interdisciplinary Graduate School of Engineering Science, Kyushu Univ., Kasuga, Japan
a r t i c l e
i n f o
Article history: Received 3 October 2014 Received in revised form 14 February 2015 Accepted 16 February 2015 Available online xxxx Keywords: Tritium Imaging plate technique Proton-conducting oxides Hydrogen solubility Diffusivity
a b s t r a c t A tritium imaging plate technique has been applied to visualize hydrogen distribution and examine hydrogen solubility and diffusivity in a proton-conducting oxide, Y-doped BaCeO3 (BaCe0.9Y0.1O3 − α). Tritium charging of the BaCe0.9Y0.1O3 − α specimens was carried out by a gas absorption method using partially tritiated water vapor (HTO, 3 kPa, T/H ~ 10−6) at temperatures ranging from 673 K to 873 K for a given time. After charging, tritium distributions of the surface and cross section of the halved specimens were visualized using an imaging plate technique. From the tritium concentration and distributions of the surface and cross section, hydrogen solubility and hydrogen (tritium) diffusivity of the BaCe0.9Y0.1O3 − α specimens were determined. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Perovskite-type oxides have received much attention as good protonic conductors under a hydrogen-containing atmosphere at elevated temperatures. These proton-conducting ceramics have been expected to be used as solid electrolytes for various electrochemical devices such as a fuel cells, hydrogen sensors, and hydrogen pumps [1–7]. To date, hydrogen solubility and diffusivity in oxide ceramics have been reported by many investigators using several experimental techniques such as conductivity measurement and thermal analysis (TG-DTA); i.e. hydrogen solubility was first measured by thermal analysis, followed by calculation of diffusivity from the solubility and conductivity data. Although this is a commonly used method, it is not a direct measurement of proton migration; therefore, the proton diffusivity measurement can be affected by other electric carriers in the conductivity measurement and experimental errors of hydrogen solubility in thermal analysis. Alternatively, ion beam microanalysis (SIMS, ERDA and NRA) [8,9] can allow for the measurement of hydrogen solubility and diffusivity directly. Although these methods have superior space resolution of hydrogen in the depth direction at a localized position, the information is generally limited near the surface of the protonic conductors. Therefore, two-dimensional distribution of hydrogen cannot be obtained directly. The tritium imaging plate (TIP) technique uses tritium as a radioactive tracer to directly assess hydrogen solubility and diffusivity from ⁎ Corresponding author at: 6-1 Kasuga-koen, Kasuga-shi, 816-8580, Japan. Tel./fax: + 81 92 583 7544. E-mail address: [email protected] (K. Hashizume).
visualized tritium distribution in tritium-dissolved materials. The imaging plate (IP) technique was originally used in biological and medical fields to measure two-dimensional radioisotope distribution and recently its applications have extended to materials and energy sciences [9–12]. However, this technique has seldom been applied to protonconducting materials such as the perovskite type oxides. In this paper, we report the experimental results of the visualization of hydrogen distribution in BaCe0.9Y0.1O3 − α, which is a well-known perovskite-type protonic conductor, by using the TIP technique. Based on these results, hydrogen solubility and diffusivity were determined for the oxide. 2. Experimental 2.1. Sample preparation and exposure to HTO vapor The BaCe0.9Y0.1O3 − α powder (TYK Co.) was die-pressed into a cylindrical shape under 200 MPa of isostatic-pressure and sintered at 1873 K for 5 h. The densities of the sintered pellets were greater than 98% of the theoretical density. The black surface layers were removed with an abrasive paper. The pellets were 7 mm in diameter and 2 mm in thickness. Fig. 1 shows the experimental apparatus. Using this apparatus, the BaCe0.9Y0.1O3 − α specimens were vacuum annealed at 973 K for 1 h, followed by exposure to ca. 3 kPa of partially tritiated water vapor (HTO, T/H ~ 10−6) at 673–873 K for 5–60 min. After the exposure experiment, the specimens were cooled down to room temperature. To measure the inside of the specimen, its surface was slightly ground with an abrasive paper to remove adsorbed tritium, and the disk was cut in half.
http://dx.doi.org/10.1016/j.ssi.2015.02.008 0167-2738/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
2
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
tritium distribution obtained with the IP technique provides information about tritium existing on and up to 1 μm from the surface. 3. Results and discussion 3.1. Visualization of hydrogen (tritium) distribution
Fig. 1. Schematic of apparatus for HTO vapor exposure of specimens at elevated temperatures.
A low-speed diamond saw (Isomet™, Buehler Ltd.) was used to cut the specimens at a cutting speed of approximately 50 rpm, and IsoCut™ Fluid (Buehler Ltd.) was used as the cooling media. Once cut, the specimens were washed with acetone to remove the coolant. 2.2. Imaging plate technique The tritium-charged specimens were placed on an IP in a dark box in order to expose the IP to β− rays emitted from the tritium adsorbed on and dissolved in the specimen. The IP used was TR-2025 (FUJIFILM Co.), which was suitable for measuring the low energy β− rays and had a wide dynamic range for the radiation energy. Considering the tritium content in the specimen, the duration of IP exposure was 3 h for the surface and 24 h for the cross section. After exposure, the IP was set in an IP Reader (BAS-2500, FUJIFILM Co.) to measure the photostimulated luminescence (PSL) intensity. The space resolution of the obtained intensity was 50 μm. Although PSL intensity is an arbitrary unit, it is proportional to the product of the radiation energy emitted from tritium and the IP exposure time allowing for conversion of PSL intensity to hydrogen concentration in the specimen. Tritium (T1/2 = 12.3 y, β− decay) emits low energy electrons, the maximum and mean energies of which are 18.6 and 5.7 keV, respectively and its maximum range is ca. 1 μm in BaCe0.9Y0.1O3 − α. Therefore, the
The hydrogen (tritium) distribution in BaCe0.9Y0.1O3 − α, which was exposed to HTO vapor, was visualized using the IP technique. Fig. 2 shows the IP images of the specimen surface and its cross section at 773 K. From Fig. 2(a), it was found that hydrogen was distributed uniformly on the specimen surface. Fig. 2(b) shows the time evolution of tritium penetration into the specimen. The tritium distribution in Fig. 2(b) seemed to be rather ambiguous due to the relatively low tritium concentration in the cross section. As seen in Fig. 2(a) and (b), the PSL intensities between the surface and near-surface bulk in the cross section largely differed. The reason for this discrepancy can be derived from the tritium density on the surface and in the bulk. The TIP method is a radiation detection technique for measuring the surface and nearsurface tritium. After tritium loading, the amount of tritium adsorbed on the specimen surface was much higher than that dissolved in terms of the atomic ratio of tritium (hydrogen) and ions of the specimen oxide (H/Oxide). Even in an equilibrium state, the tritium and ion ratios between the surface and near-surface bulk were very different (ca. 1/1 and 0.01/1, respectively). Moreover, the β− rays emitted outside from the surface were absorbed in the IP directly, while the β− rays from the bulk were absorbed by the specimen itself, even within the range of the β− ray (ca. 1 μm). Consequently, the PSL intensity of the specimen cross section became weak, compared with that of the surface. Although the PSL intensity of the cross section was weak, it was sufficient for evaluation of diffusivity and solubility. Fig. 2(c) shows the IP image of standard sample autoradiographic [3H] microscales (RPA506 and RPA507, Amersham Biosciences Co.). The 3H (tritium) activity is arranged in eight layers of polymer, the radioactivities of which range from 3.7 Bq/mg to 4070 Bq/mg. The PSL intensity was originally given as numerical data by the IP reader. The PSL intensity is proportional to the tritium activity or tritium number density over a wide range. The PSL intensities obtained from the specimen were calibrated and converted to tritium activities of the surface or bulk based on that of the standard sample. In determining the tritium activity of the specimen, the absorption and range of the β− rays in the standard sample and oxide specimens were taken into consideration. Finally, we converted the tritium activity to hydrogen number density using the hydrogen isotope abundance ratio (T/H ~ 10−6).
low
a) The surface image (773K, 15min.)
b) The cross section image (773K, 15min., 30min. & 45min.)
high
c) [3H]MISROSCALES (Left; RPA506, Right; RPA507)
Fig. 2. IP images of (a) specimen surface, (b) cross section, and (c) autoradiographic [3H] microscales.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
3.2. Hydrogen surface concentration, solubility, and diffusivity Fig. 3 shows the hydrogen (tritium) concentration on the specimen surface exposed to HTO vapor at 773 K for 5–60 min. The data points in Fig. 3 were determined by averaging the PSL intensities over the entire surface with error bars that included both scattering of the PSL intensities on the surface and dispersion in the radiation detection. The IP is a radiation detection technique for measuring the surface and nearsurface tritium. Therefore, the surface concentration of hydrogen exceeded the number of the ions in the specimen (ca. 1015/cm2). The surface concentration reached a saturated value over a short exposure time (5 min). This suggests that the hydrogen concentration near the specimen surface also increased and reached a saturated value at elevated temperatures. Fig. 4 shows the tritium concentration gradient of the specimen cross section at 773 K for 15 min (as shown in Fig. 2(b)) and its fitted curve. The data points in Fig. 4 were fitted to the curve by the following equation,
8
4x10
8
3x10
8
2x10
8
1x10
0
0
0.2
0.4
0.6
0.8
x (mm)
1
ð1Þ
where c is the tritium concentration, DT is the tritium diffusivity, x is the distance from the surface, and t is the diffusion time. This equation is an analytical solution of the diffusion equation based on a onedimensional, semi-infinite medium model in which the tritium concentration on the surface is constant [13]. As a result of curve fitting, the tritium diffusivity was found to be DT = 7.0 × 10−7 cm2/s at 773 K. In a similar manner to that described above, the diffusivities were determined at 673–873 K. The tritium diffusivity obtained from the study was, 0:34ðeVÞ 2 −4 DT cm =s ¼ 1:0 10 exp − ; kB T
ð2Þ
where kB is the Boltzmann constant and T is the absolute temperature. The diffusivities were plotted as a function of reciprocal temperature in Fig. 5, along with literature data [14,15]. Although the activation energy of diffusivity (0.34 eV) was slightly low, it was comparable to that of the literature data for proton. The diffusivity obtained was smaller than the literature data for proton. Although this discrepancy in the diffusivities could be partly due to the isotope effect between proton pffiffiffi and tritium (DT =DH 1= 3), and/or the difference in the experimental method, the details of the discrepancy were unclear at the present stage. Hydrogen solubility in BaCe0.9Y0.1O3 − α was obtained from the IP data by converting the tritium concentration to hydrogen concentration, which was plotted as a function of HTO exposure temperature,
Hydrogen concentration (/cm 2)
3
Fig. 4. Tritium distribution of the specimen cross section exposed at 773 K for 15 min, and its fitted curve.
! x erfc pffiffiffiffiffiffiffiffi ; 2 DT t
18
1.2x10
18
1x10
17
8x10
17
6x10
shown in Fig. 6. Although the hydrogen solubility data of the specimen was scattered, it was found to be ca. 1.8 mol.% and nearly constant from 673 K to 873 K. The temperature dependence of hydrogen solubility was in agreement with previous thermal analysis data [16], but the solubility itself was slightly smaller than that (~8 mol.%). The reason for the low solubility and scattering might be ascribed to tritium leakage from the specimens. The value of tritium diffusivity at room temperature extrapolated from Eq. (2) was on the order of 10−10 cm2/s: therefore, tritium existing near the surface of the specimen could diffuse out over time between cutting and IP exposure. A preliminary successive measurement by IP of the specimen cross section actually showed a decrease in the PSL intensity after a few days. Thus, the influence of the tritium leakage on the IP measurement could be reduced by decreasing the time between the specimen cutting and IP exposure, allowing for evaluation of and the true solubility by extrapolation. Moreover, the discrepancy would not be derived from the IP technique itself, but from the hydrogen behavior of the specimen. However, in order to improve the reliability of the solubility measurement, further studies are being carried out. The present study focused on the visualization of hydrogen distribution in a typical proton conductor, BaCe0.9Y0.1O3 − α. It was demonstrated
10
Diffusion coefficient (cm2/s)
cðx; t Þ ¼ cx¼0
Tritium concentration (/cm 2)
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
-4
M. Oishi et al. (proton)
10
-5
T. Schober et al. (proton)
0.46 eV 0.48 eV
10
-6
present study (tritium) 0.34 eV
17
4x10
10 0
10
20
30
40
50
60
70
time (min) Fig. 3. Time evolution of hydrogen concentration on the specimen surface at 773 K.
-7
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
-1
1000/T (K ) Fig. 5. Arrhenius plot of tritium diffusivity in BaCe0.9Y0.1O3 − α.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008
1.6
4
K. Yamashita et al. / Solid State Ionics xxx (2015) xxx–xxx
0.03
873 K. The tritium diffusivity obtained was somewhat smaller than in the literature data for proton. One reason for this discrepancy can be attributed to isotope effect. The activation energy of tritium diffusivity was 0.34 eV, which was relatively close to previous data for proton. Hydrogen solubility in the specimen converted from tritium concentration was evaluated to be ca. 1.8 mol.% and nearly constant over the temperature range. The temperature dependence of the solubility was in agreement with literature data of thermal analysis, but the solubility itself was lower.
H mol/1mol of BCY10
0.025 0.02 0.015 0.01
Acknowledgments
0.005 0 650
700
750
800
850
900
Temperature (K) Fig. 6. Hydrogen solubility in BaCe0.9Y0.1O3 − α (BCYO10).
that the IP technique is a powerful tool for studying hydrogen dissolution and migration in various protonic conductor materials. 4. Conclusion The hydrogen (tritium) distribution in a BaCe0.9Y0.1O3 − α specimen exposed to the partially tritiated water vapor (ca. 3 kPa) was visualized using a TIP technique, allowing for determination of surface concentration, solubility, and diffusivity of hydrogen. From the visualized tritium distribution on the specimen surface exposed to HTO at 773 K, it was found that hydrogen was uniformly distributed on the specimen surface and its concentration reached a saturated value after a few minutes. The tritium diffusivity and solubility in the BaCe0.9Y0.1O3 − α specimen were determined from its concentration distribution on the cross section of the specimen at a temperature range of 673 K to
This work was supported by the Japan Atomic Energy Agency under the Joint Work contract #24K389 as a part of Broader Approach activities.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]
H. Iwahara, T. Esaka, H. Uchida, N. Maeda, Solid State Ionics 3/4 (1981) 359–363. H. Iwahara, Solid State Ionics 28–30 (1988) 573–578. N. Bonanos, K. Knight, B. Ellis, Solid State Ionics 79 (1995) 161–170. T. Yajima, K. Koide, H. Takai, N. Fukatsu, H. Iwahara, Solid State Ionics 79 (1995) 333–337. E.C.C. de Souza, R. Muccillo, Mater. Res. 13 (2010) 385–394. M. Zhang, J. Xu, G. Ma, J. Mater. Sci. 46 (2011) 4690–4694. N. Nguyen, H. Yoon, J. Power Sources 231 (2013) 213–218. T. Ishigaki, S. Yamaguchi, K. Kishio, K. Fueki, H. Iwahara, Solid State Ionics 21 (1986) 239–241. E. Ruiz-Trejo, J.A. Kilner, Solid State Ionics 130 (2000) 313–324. K. Hashizume, J. Masuda, T. Otsuka, T. Tanabe, Y. Hatano, Y. Nakamura, T. Nagasaka, T. Muroga, J. Nucl. Mater. 367–370 (2007) 876–881. K. Hashizume, T. Tanabe, Fusion Sci. Technol. 60 (2011) 1135–1138. K. Hashizume, K. Ogata, M. Nishikawa, T. Tanabe, S. Abe, S. Akamaru, Y. Hatano, J. Nucl. Mater. 441 (2013) S880–S884. J. Crank, The Mathematics of Diffusion, Clarendon Press, 1975. T. Schober, W. Grover Coors, Solid State Ionics 176 (2005) 357–362. M. Oishi, S. Akoshima, K. Yashiro, K. Sato, J. Mizusaki, T. Kawada, Solid State Ionics 179 (2008) 2240–2247. K.D. Kreuer, Solid State Ionics 125 (1999) 285–302.
Please cite this article as: K. Yamashita, et al., Solid State Ionics (2015), http://dx.doi.org/10.1016/j.ssi.2015.02.008