I. INTRODUCTION

A common fabrication method for MgB2 wire or tape is the Powder in Tube (PIT) process. This conventional process is based on the powder metallurgy (PM) accompanying densification and shrinkage after sintering. Giunchi et al made an adventurous attempt with a process that reported a successful result for MgB2 superconductor wire [1]. In this process, a composite billet composed of a steel pipe internally lined with Nb tube is filled with a coaxial internal pure Mg rod and B powder. However, their Jc values are not so high at 4.2 K and large magnetic fields probably due to the high heat treatment temperature of 850-900 oC which decreases the upper critical field [2]. Very recently, they also tried C-doping to the wire and obtained the increase of Birr from 10 to 16 T at 4.2 K [3]. Togano et al fabricated MgB2/Fe wires with a similar method using Mg-Li alloy tube or rod, which is more easily mechanically cold worked than pure Mg [4-5]. It showed that the reacted MgB2 layer has dense structure. However, the Li was diffused into MgB2 layer and Jc values of MgB2 layer showed a decrease as a result.

In this research, critical current density using a classical diffusion process without any pressure was achieved as much as that of Nb3Sn. One of the remarkable properties of an internal Mg diffusion process (IMD) is quite hard dense structure. Heat treatment temperature dependences of critical current (Ic), cross sectional area of the reacted layer, critical current density (Jc) and Vickers hardness with each different sheath of Fe and Ta were systematically investigated. The fabrication and characterization of MgB2 wire, both monocore and multicore, with non-doped and 10 mol% SiC doped samples were also carried out.

II. FABRICATION OF WIRES

The process of this research is basically the same as that reported by Togano et al [4]. The schematic of the IMD processed multicore wire is shown in figure 1. The prepared sheath material is pure Fe or Ta tube with an outer diameter of 6 mm, an inner diameter of 4 mm and a length of 50 mm. A pure Mg rod with a diameter of 2 mm was placed at the center of Fe or Ta tube, and the empty space between the Mg rod and the Fe or Ta tube was filled with the amorphous B powder (99.99%, -300 Mesh) or the powder mixture of B and 5 or 10 mol% SiC (a few tenths nanometer size) as shown in figure 1. The weight ratio of Mg rod to packed powders was about 3:1 (for B+5 or 10 mol% SiC mixed powder). The Mg rod is prepared extra amount in order to compensate the loss by oxidation and evaporation above the Mg melting point (650 oC). In monocore case, the composite performed initially swaging until a rod shape with an outer diameter of 4 mm and then wedged the rod into the Cu-20wt%Ni sheath with an inner diameter of 4 mm and an outer diameter of 6 mm. The prepared sample was groove-rolled into a rod shape with 2.3 x 2.3 mm2 cross section and then drawn into a wire of 1.3 mm in diameter.

In multicore case, seven pieces of the 7 monocore wires of 1.3 mm diameter and 47 mm length of Fe or Ta were bundled and inserted into the 4 mm hole of the Cu-20%Ni tube of 6 mm outer diameter. Then the composite was cold worked into a wire of 1.3 mm outer diameter by the same process for the monocore wire. The samples cut at a length of 40 mm were heat treated between 600 and 800 oC from 0.25 to 10 hr under Ar gas atmosphere. During heat treatment, the samples were wrapped with a zirconium foil in order to minimize oxidation.

III. RESULTS AND DISCUSSION

A. Microstructure for monocore and multicore

Figure 2 (a) and (b) shows optical micrographs of the polished transverse and longitudinal cross sectional area of the Ta sheathed 10 mol% SiC doped multicore after cold working. The composite of the pure Mg/[B(-SiC) powder]/Ta/Cu-20wt%Ni tube with an outer diameter of 6 mm was successfully cold worked into a 1.3 mm wire at room temperature without any breakage of Mg in spirit of very poor workability of pure Mg. It supposes that workability of pure Mg is overcame without any breakage by IMD process because of confinement by boron powder with Ta or Fe sheath of good workability.

Figure 2 (c) and (d) shows transverse and longitudinal cross sectional area of this multicore wire heat treated at 645 oC for 1hr, respectively. Figure 2 (e) and (f) shows transverse cross sectional area of Fe and Ta sheathed 10 mol% SiC doped monocore wire heat treated at 640 oC for 1 hr. Although remarkable shrinkage for both monocore and multicore wire occurred after sintering, the diffused MgB2 layer showed uniform deformation without porosity inside the microstructure. During the heat treatment, liquid Mg seems to infiltrate into the boron layer to form MgB2 layer along the inner wall of the Fe or Ta sheath, which is seen as a ring on the cross section. As the structure is the same even when the heat treatment temperature is below the Mg melting point, Mg may be melted by heat generated by the exothermic reaction forming MgB2 at the B/Mg interface.

Figure 3 shows the SEM image of reacted MgB2 layer compared with that of PIT processed wire. The PIT-processed wire shows granular MgB2 microstructure, which is typical for PIT processed MgB2. On the other hand, IMD processed wire shows much denser structure than that of the PIT processed wire, and there is not porosity. It depends on the volume faction of raw boron powder. Being isolated from Mg, denser packed boron powder in IMD process caused denser MgB2 structure. In conventional in-situ PIT process, boron powder mixed with Mg forms porous microstructure of low density. The amount of remaining Mg decreases with the increasing in the heat treatment temperature and dwell time. However, we need to take into account of the evaporation and/or leakage of Mg. Such evaporation and/or leakage of Mg are suppressed for longer wires, and higher Jc can be expected.

B. Fe or Ta sheathed monocore

The heat treatment temperature dependence of transport critical current (Ic), cross sectional area of the reacted layer, critical current density (Jc), and Vickers hardness (HV) for dwell time of 0.5 hr are shown in figure 4. The Ic, and Jc of 10 mol% SiC doped samples are higher than those of non-doped samples. It is well known Hc2 enhancement caused by the fact that the C atoms from the SiC powder can effectively substitute on the B site [6]. Also, the maximum values of Ic and Jc for non-doped and 10 mol% SiC doped samples occur at a heat treatment temperature of 640 oC. The origin of the maximum is considered to be a competition between Jc increase by the progress of MgB2 forming reaction and Jc decrease by the MgB2 grain growth with increasing heat treatment temperature. The Ic for monocore in Ta sheath is almost the same or slightly lower than Fe sheathed monocore samples. Kovac et al. reported that the higher MgB2 core density corresponds to the stronger metallic sheath [7]. So, excellent Jc is obtained from Ta sheathed monocore, considering that HV of Ta sheath (~HV 204) is stronger than that of Fe sheath (~HV 118). In fact, a comparison of density relative to HV in Ta and Fe sheath is difficult because of improved grain connectivity, nano-sized MgB2 grain, solid solution strengthening by decomposed C and Si. The other reasons of hardening expect to the hardening by SiC or processing are taken account into the age hardening or affection of secondary phase such as MgO.

In the previous work of Husek et al., they reported a good correlation between density and measured HV [8]. The superior value of HV of typical MgB2 wire by an ex-situ PIT process was about HV 400 due to the numerous pores and the microstructure. HV 1700, the highest hardness in MgB2 bulk, was achieved with the hot isotropic press (HIP) method by Takano et al [9]. As shown in figure 4, the highest value of HV for our samples reaches 1700, which is 4 times higher than that for conventional PIT-processed wire and is almost the same as that of bulk sample. From the above comparison, it is found that wires fabricated by this IMD process bring quite high density, which should be the origin of the high Jc.

As a result, for Ta sheathed 10 mol% SiC doped monocore sample heat treated at 640 oC for 0.5 hr, Jc at 4.2 K and 7 T reached the value of 300k A/cm2, which is 3 times higher than that of recently reported conventional in-situ PIT processed wires.

C. Ta/Cu-20wt%Ni sheathed monocore and multicore

The critical current density (Jc) for 10 mol% SiC doped monocore and 7 multicore wires with Ta sheath inserted into the Cu-20wt%Ni sheath show similar trends for heat treatment temperature at the dwell time of 1 hr as shown in figure 5. Although the area of the reacted layer in monocore increases with increasing temperature, the area in multicore is almost constant. This means that the reaction between B and Mg to form MgB2 is finished at even the lowest heat treatment temperature of 600 oC for 1 hr in the multicore sample, because of sufficiently shorter diffusion length. As a result, Ic for the multicore is higher than that for the monocore. The larger Ic is important especially for the practical use.

IV. CONCLUSION

In this thesis, it was found that the internal Mg diffusion (IMD) process achieved much improved MgB2 core density and, hence, much enhanced critical current density (Jc) values. The highest Jc value was obtained for Ta sheathed 10 mol% SiC doped 7 multicore wires and was exceedingly over 300k A/cm2 at 7 T in 4.2 K and about over 3 times higher than those of in-situ powder in tube (PIT) processed wires (about 90k A/cm2) and was reached to almost practical level of Jc values as shown in figure 6.

Figure 1 Schematic illustration of the multicore wire fabrication process

Figure 2 Optical micrographs polished (a), (c) transverse (b), (d) longitudinal cross sectional area of Ta sheathed 10 mol% SiC doped multicore wire (a), (b) after cold working and (c), (d) after heat treatment at 645 oC for 1 hr. Transverse cross sectional area of (e) Fe , (f) Ta sheathed 10 mol% SiC doped monocore wire heat treated at 640 oC for 1 hr.

Figure 3 SEM image of microstructure of (a) IMD processed wire and (b) PIT processed wire.

Figure 4 Ic, cross sectional area, Jc at 7 T and 12 T and Vickers hardness of IMD-processed wire heat treated at 600 oC, 640 oC and 680 oC for 0.5 hr ; black (non-doped) and red (10 mol% SiC doped).

Figure 5 Averaged Ic, cross sectional area and averaged Jc at 7 T and 4.2 K for Ta sheathed 10 mol% SiC doped multicore and monocore wires heat treated at various temperature for 1 hr; black closed triangle (multicore) and red closed diamond (monocore).

Figure 6 Applied magnetic field dependence of transport Jc for reported highest Jc values of IMD processed Ta/Cu-20wt%Ni sheathed 10 mol% SiC doped MgB2 wire compared with applied highest Jc values of conventional PIT processed wire, Nb-Ti and Nb3Sn superconductor wire or tape.

本論文では、MgB2線材の作製において、Mg棒からB粉末への拡散後の反応を使うことによって、Mg粉末とB粉末のその場での反応を使う一般的なPIT(Powder in Tube)法に比べ、臨界電流を数倍大きくすることに成功した。光学および電子顕微鏡観察とVickers硬さ試験により、その理由が緻密な組織であることを明らかにした。さらに、より最適なシース材料や熱処理条件とその理由を明らかにした。シースは、ボロンと反応し難く、より緻密な組織を得るために硬い材料が良い。熱処理条件は、温度上昇や時間増加による緻密化と粒成長の競合により、臨界電流が最大となる温度と時間が存在する。論文は10章からなる。

第1章は序論であり、超伝導についての一般的性質と超伝導体の応用例について述べている。超伝導の臨界温度、臨界磁場、臨界電流密度、マイスナー効果、第I種および第II種超伝導体、ジョセフソン効果について説明している。超伝導の大規模な応用として送電線やリニヤーモーターカーがあり、電子デバイスへの応用として各種センサーや検出器があることを述べ、特に大規模な応用には臨界電流密度を高くすることが必要であることを指摘している。

第2章は背景であり、MgB2超伝導体の位置付けやMgB2線材に関する従来の研究について概観し、本研究の目的を述べている。特に、直前に行われた戸叶等のMg-Li合金棒を使った研究との違いを明確にしている。MgB2の有利な特徴として、臨界温度が高く歪みに鈍感であること、結晶の配向が必要無いこと、低価格で軽量であることを挙げている。ただし、これまでに報告されてきたMgB2線材の臨界電流は、実用線材であるNb-TiやNb3Snに及ばないことが問題であるとしている。これまでに実施されてきた臨界電流向上の試みとしては、ドーパントの添加、新しいプロセスの適用、シース材料の探索を挙げている。次に、従来の最も一般的なMgB2線材作製プロセスであるPIT法と最近試みられているIMD(Internal Mg Diffusion)法について述べている。比較的高い臨界電流が得られているIn-situ PIT法では、Mg粉末とB粉末を押し固めてから反応させるため、押し固めにより生じる空隙と反応による体積収縮により、MgB2の充填率は50%以下になる。一方、IMD法では、押し固めたB粉末中に中心の棒からMgが拡散してMgB2を生じるため、中心に大きな穴は生じるが、より緻密な組織が得られる。ただし、戸叶等の研究では、Mgの供給源として、中心の棒材として加工性に優れたMg-Li合金を使ったため、作製されたMgB2中にLi化合物が混入した。また、シース材料のFeとB粉末が反応してFeB2も生成した。以上から、本研究は、IMD法を用いて、実用線材であるNb-TiやNb3Snと同程度の臨界電流を持ったMgB2線材を作製することを目的とした。中心の棒材としては純Mgを用い、これまでで最も有望なSiCをドーパントとし、シース材としてはFeの他により硬いTaや外シースとしてCu-Ni合金を試し、熱処理温度と保持時間を変えて最適な条件を探した。また、単芯の線材と共に、実用化を念頭に7芯の多芯線材の作製も行った。

第3章は実験方法であり、本論文の第4章から7章までに共通する線材の評価方法と熱処理について述べている。評価には、光学顕微鏡、X線回折、走査型電子顕微鏡、エネルギー分散型X線スペクトルを用いている。臨界電流密度とVickers硬さの求め方について詳細を述べている。章毎に異なる線材の作製方法は、各章の始めの部分で述べている。

第4章は、Feシース単芯線材についての結果である。加工性の悪い純Mg棒材を使ったにもかかわらず、全体を直径6mmから1.2mmまで室温で線引きしてもMgは切れ切れになることは無く一様に細くなっており、その理由を加工性の良いシース材による束縛の効果であるとしている。熱処理後、中心に大きな穴は開くが、熱処理温度がMgの融点(650℃)より高い場合は1時間以上で、融点より低くても3時間以上でB層に完全に拡散し、MgB2の緻密な組織が得られている。臨界電流密度は、他の条件が同じPIT法で作製された線材に比べて大きく上昇したことが示されている。

第5章は、FeまたはTa内シースとCu-20wt%Ni外シースを使った単芯線材の結果である。Cu-20wt%Ni外シースは、安定化と半田付けを可能にするために使ったとしている。最大の臨界電流密度は、Ta内シースを使い640℃で0.5時間熱処理した場合に、4.2K、7TでPIT法による過去の報告の最大値の約3倍である300kA/cm2を得ることに成功している。Vickers硬さは熱処理温度とともに緻密化により上昇し1000から1700となったが、これは通常のPIT法線材の値である約400よりもはるかに高く、高圧合成したバルク材の値に近いとしている。熱処理条件は、温度上昇や時間増加による緻密化と粒成長の競合により、臨界電流が最大となる温度と時間が640℃で0.5時間になったとしている。FeよりTa内シースで臨界電流密度が高くなった理由は、Taの方が硬いため室温での線引き後のB粉末の充填率がより高く、Mgとの反応後もより緻密な組織が得られたためと考えられるとしている。

第6章は、Feシースの多芯線材の結果である。多芯線材の最終的な線引き後の直径は1.3mmと単芯線材(1.2mm)とほぼ同じであり、B粉末層の厚さが小さいためにMgの拡散距離が短く、臨界電流が最大となる温度と時間が600℃で1時間と、より低く短い時間になったとしている。

第7章は、Ta内シースとCu-20wt%Ni外シースを使った多芯線材の結果である。多芯線材と単芯線材の臨界電流密度の熱処理温度依存性は、ほぼ同じであるが、熱処理温度がMgの融点より低くなるほど1時間の熱処理では単芯線材のMgB2層の断面積は小さくなるが、B粉末層の厚さが小さくMgの拡散距離が短い多芯線材では常に完全に拡散・反応が終了し断面積はほぼ一定である。このため、低温熱処理では多芯線材は単芯線材より臨界電流を大きくすることに成功している。同じ径の線材で臨界電流が大きくなることは、実用的には重要な結果である。

第8章は今回のIMD法における拡散機構のまとめ、第9章は本論文の結論、第10章は今後の展望である。

なお、本論文第4、5、6、7章は、熊倉浩明、和田 仁、松本明善、戸叶一正、木村 薫との共同研究であるが、論文提出者が主体となって測定および解析を行ったもので、論文提出者の寄与が十分であると判断する。

以上本論文は、IMD法によるMgB2線材の最適な作製条件を検討し、従来のPIT法に比べ臨界電流密度を大幅に改善することに成功し、その理由を明らかにしており、物質科学および材料工学の発展に寄与するところが大きく、よって博士(科学)の学位を授与できると認める。