1. Introduction

In order to sustain the tritium self-sufficiency for fusion reaction, it is important to minimize the tritium inventory in the blanket. Accordingly, the understanding of tritium migration and release process in ceramic breeder materials is indispensable. The tritium release from ceramic breeder materials is understood to be composed by multiple fundamental processes including diffusion in the bulk, surface processes, migration along grain boundaries, percolation in the interconnected open pores and convection by purge gas. The previous experiments indicate that the tritium release is influenced by temperature, purge gas condition, irradiation, microstructure of materials, etc. However, the influence of material microstructure (grain size, surface condition, and open and closed pores) has not been fully understood.

The objective of the present work is to study the influence of microstructure (grain size, surface condition, and open and closed pores) on the release behavior of hydrogen isotopes in ternary lithium oxides systematically by changing the samples from single crystal to poly crystal. In order to achieve the objective, LiNbO(3) and Li(2)TiO(3) are chosen as the researching materials.

2. Experiment

2.1 Sample preparation

LiNbO(3) single crystal cubic samples (4.7×4.7×4.7 mm3, 1.5×1.5×1.5 mm3, 0.46×0.46×0.46 mm3) are cut from a block by using a diamond cutter. LiNbO(3) single crystal powder samples with size of around 50 μm (av.) and around 5 μm (av.) are prepared by crushing a part of the same block and meshing with a set of sieves: 45 μm and 53 μm sieves for the former, and 5 μm and 6 μm sieves for the latter. LiNbO(3) poly crystal powder sample with size of around 50 μm (av.) is prepared by meshing the as-received material.

Li(2)TiO(3) single crystal plate samples are synthesized by a flux method with B(2)O(3) as the flux agent. Li(2)TiO(3) single crystal powder samples with size of around 50 μm (av.) and around 5 μm (av.) are prepared by crushing single crystal plates and meshing. Li(2)TiO(3) poly crystal pellets are prepared by pressing the starting materials of as-received Li(2)TiO(3) poly crystal powder, PVA (5 wt%) and stearic acid (5 wt%) and sintering at 1173 K, 1223 K, 1273 K, 1373 K, and 1473 K for 12 h in air to obtain five pellet samples with certain apparent density (% of theoretical density): 76% T.D., 83.5% T.D., 86.6% T.D., 88.6% T.D. (1373 K) and 88.6% T.D.(1473 K). These apparent densities were evaluated by measuring the mass and size of pellets.

2.2 Experiment procedure

Samples put in a Ni sample holder are heated at 1073 K for 1 h in 100 Pa D(2)O vapor for deuterium absorption in absorption system. After cooling down and evacuation, samples are transferred to a high-vacuum TDS system without exposure to air by using a transfer rod. In TDS system samples are heated at 5 K/min to high temperatures (1173 K for cubic and plate sample, 973 K for powder and pellet samples) for desorption. The species and amount of gases released from the samples during heating process are monitored by a quadrupole mass spectrometer (QMS). The correlation of QMS signal versus the temperature or time, namely TDS spectrum is obtained.

3. Release behavior of hydrogen isotopes in LiNbO(3) single crystal cubic and powder

The release behavior of hydrogen isotopes in LiNbO(3) single crystal cubic (4.7×4.7×4.7 mm3, 1.5×1.5×1.5 mm(3), 0.46×0.46×0.46 mm(3)) and powder (50 μm av.) samples was studied by TDS. The deuterium thermally sorbed was mainly released in HDO and D(2)O.

In the large-size cubic samples, two peaks were observed for HDO: low-temperature peak (590~600 K) related to recombination desorption of hydroxyl groups chemically adsorbed on the surface (surface hydroxyl groups) and high-temperature peak (> 800 K) related to diffusion of hydroxyl groups in the bulk. With the sample size decreasing, the peak related to bulk diffusion was clearly observed to move to lower temperature regions, which is due to the shorter diffusion path, while the peak related to recombination desorption did not change much.

In the small-size powder (50 μm av.) sample, two peaks were observed for HDO: (i) around 350 K due to desorption of molecules physically adsorbed on the surface; (ii) around 635 K, with H2O peak observed at the same temperature, due to recombination desorption of hydroxyl groups chemically adsorbed on the surface. The bulk diffusion in single crystal was modelled by using the simple diffusion model. It was shown that the rate-controlling step is the bulk diffusion at high temperatures (> 800 K) in the large-size cubic sample, and moves to surface processes at low temperature region (590~635 K) in the small-size powder sample.

4. Release behavior of hydrogen isotopes in Li(2)TiO(3) single crystal plate and powder

The release behavior of hydrogen isotopes in Li(2)TiO(3) single crystal samples was studied by TDS. Similar to LiNbO(3), the deuterium thermally sorbed in Li(2)TiO(3) was mostly released in HDO and D(2)O. In single crystal plate sample (thickness of 0.2~0.4 mm), three peaks were observed for HDO:

Peak (1): due to desorption of water molecules physically adsorbed on the surface (around 345 K)

Peak (2): due to recombination desorption of surface hydroxyl groups (around 625 K)

Peak (3): due to recombination desorption of hydroxyl groups diffused from the bulk (around 1000 K)

In small-size powder samples (50 μm av., 5 μm av.), no obvious peak (3) was observed, which is considered to be due to (i) the sample amount is too small to observe this peak, (ii) the rate-controlling step has moved from the bulk diffusion to surface processes.

A broad peak was observed at 400~600K in 5 μm (av.) powder sample:

Peak (4): the peak between peak (1) and peak (2) (400~600K), which is considered to be the release of molecules (HDO, D(2)O) physically adsorbed on the surface of inner open pores created by agglomeration through gas migration along narrow open pore channels with different shapes and sizes.

By comparing the TDS spectra of LiNbO(3) and Li(2)TiO(3), it was found that surface hydroxyl groups have the similar environment in LiNbO(3) and Li(2)TiO(3) and both bonded to Li atoms. The diffusivity of hydroxyl groups in the bulk of grains is smaller in Li(2)TiO(3) than that in LiNbO(3).

5. Influence of surface condition on hydrogen isotopes release behavior in LiNbO(3) cubic (single crystal) and powder (single and poly crystal)

The release behavior of hydrogen isotopes in LiNbO(3) single and poly crystal samples with different surface condition was studied by TDS. The surface morphology was observed by SEM and the specific surface area of powder samples was measured by BET.

In single crystal cubic sample (1.5×1.5×1.5 mm3), HDO, D(2)O and H2O peaks related to recombination desorption of surface hydroxyl groups (peak (2)) disappeared after high-temperature pretreatment (at 1233 K for 1 h in vacuum and followed at 1073 K for 1 h in air), while the peak related to bulk diffusion (peak (3)) did not largely change. Meanwhile, the surface roughness of cubic sample created by cutting was greatly smoothed by high temperature pretreatment. The decrease in the amount of surface hydroxyl groups is considered to result from the decrease of the surface area and recovery of surface defects induced by high-temperature pretreatment.

In 50 μm (av.) single and poly crystal powder samples, the amount of surface hydroxyl groups per gram is proportional to the specific surface area measured by BET. While in 5 μm (av.) single crystal powder sample, smallest amount of surface hydroxyl groups was observed in contrast of the largest theoretical and BET specific surface areas, which is considered to be related to agglomeration of this powder sample, for example the access of hydroxyl groups to the surface was decreased due to agglomeration.

By assuming that the surface area for one hydroxyl groups (-OH or -OD) is half of the surface area for one water molecule with Van der Waals radius of around 1.4 A and considering the distance between two oxygen atoms of around 2.7~3 A in LiNbO(3) crystallographic frame, the amount of hydroxyl groups per m2 adsorbed in the monolayer was calculated (calculation result). The experiment results for the amount of hydroxyl groups per m2 were estimated from the TDS results and BET measurement. It was found that the experiment results are around two orders of magnitude larger in the 50 μm (av.) powder samples and one order of magnitude larger in the 5 μm (av.) single crystal powder sample by comparing with the calculation result. It reveals that hydroxyl groups are densely formed on the surface region.

For peak (2) at 590~635 K related to recombination desorption of surface hydroxyl groups, H2O peaks have almost two orders of magnitude higher level than HDO and D(2)O, which is mainly due to the -OH formed by isotope exchange reaction between -OD on the sample surface and H2O residue in the system.

6. Release behavior of hydrogen isotopes in Li(2)TiO(3) poly crystal pellet

The release behavior of Li(2)TiO(3) poly crystal pellet samples with different microstructure was studied by TDS. The grain size and pore morphology were observed by SEM. The open porosity, open pore size distribution and specific surface area of open pores were measured by mercury intrusion porosimetry.

The amount of molecules physically adsorbed (peak (1)) and hydroxyl groups chemically adsorbed (peak (2)) on the surface are not totally proportional to the specific surface area of open pores measured by mercury intrusion porosimetry, which could not reflect the real surface area of pellet samples due to that pores smaller than 168 nm could not be detected. More precise method should be used to measure the specific surface area of pellets.

The peak (4) related to the release of molecules physically adsorbed on the surface of inner open pores was observed at 400~650 K and similar to that was observed in 5 μm (av.) LiNbO(3) and Li(2)TiO(3) single crystal powder samples with obvious agglomeration. It is considered that the release of peak (4) is mainly due to the Knudsen diffusion and interaction between gas molecules and walls of narrow open pore channels with several (tens) nm.

A new peak - peak (5) was observed at 650~900 K in high-density (≧ 83.5% T.D.) pellet samples, for which pores with dead end were observed by SEM. It is due to the release of hydroxyl groups absorbed in the bulk of grains with size of several (tens) μm, which was delayed due to the existence of closed pores. The "effective grain size" is defined to be an effective diffusion length needed to reach the surface (or open pores) and calculated from temperature of peak (5) by using the simple diffusion model. It is found that the "effective grain size" is several (tens) times larger than the real grain size decided by SEM observation. It is considered that the hydroxyl groups absorbed in the bulk of high density pellet samples have to undergo bulk diffusion in multi-grains and trapping in closed pores before arriving at the surface or open pores.

7. Conclusion

As the sample size of single crystal decreases from several hundreds μm to 50 μm and smaller, the rate-controlling step moves from the bulk diffusion at high temperatures (> 800 K) to surface processes at lower temperature region (590~635 K).

The amount of surface hydroxyl groups, which can influence the tritium release behavior by acting as tritium trapping, is found to decrease with the surface area decreasing induced by the smoothing of surface roughness at high temperatures. It was revealed that isotope exchange reaction between surface hydroxyl groups and H2O residue in system occurs easily.

The release of hydroxyl groups absorbed in the bulk of grains with size of several (tens) μm is delayed by the existence of closed pores to high temperature region (650~900 K). Before arriving at the surface or open pores, these hydroxyl groups have to undergo bulk diffusion in multi-grains and trapping in closed pores several times.

The migration of gas along narrow open pore channels occurred at relatively high temperature of 450~600 K due to the Knudsen diffusion and interaction between gas molecules and walls of narrow open pore channels of several (tens) nm.

核融合炉燃料サイクルを確立するにはブランケットにおけるトリチウムインベントリーを小さくすることが重要である。そのためにリチウム酸化物を用いる固体ブランケットにおいては、トリチウム移行、放出プロセスを正しく知る必要がある。リチウム酸化物固体増殖材料において生成されたトリチウムは、バルク内拡散、及び表面脱離、結晶粒界にそっての移動、連結された空孔に沿ってのパーコレーション、パージガスによる脱離などの素過程から構成されている。既往の研究ではトリチウム放出挙動は温度、パージガスの化学組成、照射、材料のミクロ構造などによって影響されることが示されている。この中で材料のミクロ構造(結晶粒サイズ、表面状態、通気孔、孤立気孔)の影響については十分に研究が行われていないとの認識のもとに、本論文の研究目的は、三元系リチウム酸化物における水素同位体放出挙動に及ぼすミクロ構造の影響を、試料を単結晶から多結晶と変化させること等によって系統的に明らかにすることとしている。

論文は7章より構成されている。第1章は研究の背景と目的が記されている。

第2章は実験方法について記されている。試料として、LiNbO(3)単結晶の立方体試料(4.7x4.7x4.7mm(3), 1.5x1.5x1.5mm(3), 0.46x0.46x0.46mm(3))、及びLiNbO(3)単結晶粉末(平均粒径50μm,5μm)、LiNbO(3)多結晶体粉末(平均粒径50μm)、Li(2)TiO(3)単結晶板(厚み0.2-0.4mm)、Li(2)TiO(3)単結晶粉末(平均粒径50μm,5μm)、Li(2)TiO(3)多結晶粒子から焼結作成されたLi2TiO3多結晶体ペレットなど、多くの異なるミクロ構造を持つ試料を用いていることが本研究の大きな特徴である。実験ではまず1073Kで1時間、100PaのD2O蒸気にさらすことにより重水を吸収させる。その後試料を室温まで冷却後、眞空中でTDS(thermal desorption spectroscopy)装置に移し、5K/minの昇温速度で1173Kまで加熱し放出される水素同位体をQMSで観測することが示されている。

第3章はLiNbO(3)単結晶立方体試料およびLiNbO(3)単結晶粉末試料についての実験結果が示されている。大きなサイズの単結晶立方体試料では化学吸着された表面水酸基の再結合脱離によるピーク(590-600K)とバルク中での水酸基の拡散によるピーク(>800K)が観察されている。拡散によるピークは単結晶立方体試料の大きさと拡散係数から計算されるピーク温度と一致するという興味深い結果が示されている。LiNbO(3)単結晶粉末試料では物理吸着された重水の脱離ピーク(約350K)と化学吸着された表面水酸基の再結合脱離ピーク(635K)が観察されたとしている。

第4章はLi(2)TiO(3)単結晶板及びLi(2)TiO(3)単結晶粉末での実験結果が示されており、物理吸着された水蒸気の脱離(約345K)、及び表面水酸基の再結合脱離(約625K)、バルクでの拡散による脱離(約1000K)の3つのピークが観察されたとしている。また、5μmの単結晶粉末試料においては新たに400-600Kにブロードな放出ピークが観察されており、これは内部の通気孔表面に吸着された重水蒸気の細孔チャンネルを通しての移行によるものと結論づけている。また、LiNbO(3)単結晶試料とLi(2)TiO(3)単結晶試料のTDS結果を比較し、表面重水酸基の存在状態はいずれの材料においてもLi原子に結合したものであると考えている。

第5章はLiNbO(3)試料について表面状態の影響に着目しての実験結果を示している。ここではSEMによる表面観察や、BETによる表面積測定結果との比較も行われておる。LiNbO(3)単結晶立方体試料(1.5x1.5x1.5mm(3))で、高温前処理(1233K、1h)した後は、表面水酸基再結合によるピークは小さくなった。これは、高温前処理により表面が平滑になったことによると結論している。また、50μmの単結晶および多結晶粉末試料では表面水酸基の再結合放出量はBET表面積に比例することが示されている。しかし、5μm単結晶粉末試料では、放出量は表面積から予測されるものよりも小さいことが示され、これは微小粒子の凝集によると結論している。また、表面水酸基の再結合によると考えられる600K付近のピークがH(2)Oによるものが主になっておること、及びその量が表面積から推定される量よりも大きいことについて検討を加え、実験システム内でのH2Oとの交換反応、及び表面層内での水酸基の存在とそれとの交換反応によるものと考察している。

第6章は、Li(2)TiO(3)多結晶体ペレットについてのTDS実験結果を示すとともに、SEM表面観察、水銀注入法による空孔率測定結果を用いて検討している。その結果、表面物理吸着水蒸気の脱離、表面水酸基の再結合脱離ピーク以外に、400-600Kに内部の通気孔表面に吸着された重水蒸気の細孔チャンネルを通しての移行によるブロードなピーク、及び650-900Kに結晶粒内に吸収された重水酸基の多回の拡散、脱離、細孔移動によるブロードなピークが観察されたとしている。

第7章は結論と今後の研究の展望が述べられている。

以上要するに、本研究は核融合炉固体増殖材料である三元系リチウム酸化物について、水素同位体の移行・放出の基礎過程を、結晶状態、試料形状、前処理条件などを変えることによるミクロ構造の観点から体系的に調べ明らかにしたものである。これらは学術上重要な知見を与えるものであるとともに原子力工学、特に核融合材料工学への貢献も大きい。

よって本論文は博士(工学)の学位請求論文として合格と認められる。