核融合研究
Online ISSN : 1884-9571
Print ISSN : 0451-2375
ISSN-L : 0451-2375
45 巻, 4 号
選択された号の論文の3件中1~3を表示しています
  • 秋宗 秀夫
    1981 年 45 巻 4 号 p. 324-366
    発行日: 1981年
    公開日: 2011/03/04
    ジャーナル フリー
    A new type of negative deuterium ion source for the high energy neutral beam injector is proposed. The D-ions are formed via two steps of processes, which are the dissociation of D2 and the electronic transition. The dissociation occurs when deuteriums are solved into the Si-semiconductor on a side where D2 gas is filled at about one atmospheric pressure. The atomic deuteriums are almost in ionized state at inner sites far from the surface according to the Fermi-Dirac distribution function. However, the deuteriums at just inner sites from the surface and ones emerging from the surface are considered to be almost in neutral state, because the energy gap between the ionized stat and neutral state is enlarged due to the image force on D+ ions in the vicinity of the surface. The D-ions are formed with the resonance transition of electron to D (ls) emerging from the surface of Si-semiconductor at the vacuum side. The hot electrons for the resonance transition to take place, are produced with the avalanche arising in the p-n junction in the Si-semiconductor. The energy distribution of hot electrons depends on the parameters relating to the junction, which are chosen as to minimizing unfavourable electron flux emerging from the surface. The probability of the resonance transition, D (ls) +e→D- (ls ls') is obtained with its matrix element which is calculated with use of the simple wave function for the electrons of D- (ls ls'), φ1s1s' (r1, r2) =α3/2 β 3/2 /π exp (-αr1-βr2), where α =1.04/ao, β=O.28/ao, ao is the Bohr radius. The effective transition rate can be calculated with the transition probability and the energy distribution of hot electron in the semiconductor, assuming the suitable forms of potential energy arising from the interaction between particles and the surface of semiconductor. The deuteriums stuck on the surface by the van der Waals force are easily converted to D- ions with the electronic transition, assuming that D2 formation due to the recombination of D (ls) 's is slower process. Esimations are also made on the deuterium flux diffusing through the D- ion formation element which consists of a bi-layer of Si-semiconductor and vanadium, V. The calculations are made with use of the following values; the total electron density in conduction band, 1 X1019/ cm3; the diffusion coefficients of deuterium in Si-semiconductor and vanadium, Dsi= 1.8×10-8 cm2 /sec and Dv. 4×10-5cm2/sec respectively at 150°C, the deuterium concentration in vanadium at the Si-V boundary, 1×1022 /cm3. The results show that the obtainable current density of D-ion is about 27 mA/cm2 with the D- ion formation element of which Si-layer thickness is 9 μm and V-layer thickness is 0.1 mm. The neutral component leaving from the surface is about 18 percent of D- ion flux. On the basis of results, a conceptual design is made on the D- ion source with output current of about 20 A.
  • 市川 行和
    1981 年 45 巻 4 号 p. 368-371
    発行日: 1981年
    公開日: 2011/03/04
    ジャーナル フリー
  • 折居 茂夫
    1981 年 45 巻 4 号 p. 372-385
    発行日: 1981年
    公開日: 2011/03/04
    ジャーナル フリー
    An ion loss from a magnetic mirror under the effect of ambipolar potential is considered in a pure decay. An ion distribution function in velocity space at a given instant of time is estimated with results of numerical calculation of a full set of two dimensional Fokker-Planck equation. The estimated velocity distribution function is in good agreement with the results of the numerical calculation. And using this function, we obtained the ion loss-time from the two dimensional Fokker-Planck equation which is assumed that the velocity distribution of field particles is isotropic. In this calculation, it may also be explained that the longitudinal loss can not be negligible in comparison with the transverse ones on a large potential region. Finally, it is showed that the velocity shift Vs is necessary to determine the loss-time (see Fig. 10, 11).
feedback
Top