|
内容記述 |
The radial electric fields in magnetically confined toroidal plasmas are<br> considered to play an important role in plasma confinements. For<br>example, they change the neoclassical ripple loss in the low collisionality<br>regime in helical devices, and they are considered to be the important<br>parameter to determine the L/H transition and the anomalous transport<br>characteristics of L/H-modes in tokamak plasmas. Therefore many efforts<br>to study the radial electric fields experimentally have been performed.<br>One method for this study is the spectroscopic measurements of the<br>rotations of impurity ions. However, toroidal effects on the plasma<br>rotations have never been studied experimentally. The coupling of<br>toroidal and poloidal rotations caused by the toroidal effect to satisfy the<br>poloidal flow conservation condition is the most important basis of<br> neoclassical transport theory and is also important for understanding the<br>supersonic (with M<small>p</small>~1 where M<small>p</small> is the poloidal Mach number) plasma<br>flows in tokamak H-mode plasmas. Therefore many related theoretical<br> studies have been made.<br /> To study this problem experimentally is to compare poloidal flux on<br> the inside and outside of the magnetic surfaces. In the poloidal rotation<br>measurements in many tokamaks, the poloidal rotation velocities only in<br> the outside were measured, since it is difficult to install the observation<br>chords viewing vertically the inside of the torus. Another severe<br>difficulty is the calibration of mechanical wavelength offset~0.5 Å) of <br>spectrometers with the accuracy for the plasma rotation measurements.<br> The study of the inside/outside asymmetry of poloidal rotation velocity requires<br> the accuracy of absolute wavelength of ~0.01 Å. To measure<br> the absolute value of the rotation velocity canceling this offset, it needs<br>the observation along opposite viewing directions. In past plasma rotation<br>measurements using the observation from one direction only, some<br>assumptions or approximations about the plasma rotation velocity profiles<br>were used. For example, the average of the poloidal rotation velocities in<br> the inside and the outside was used as poloidal rotation 'velocity' in<br> Heliotron E.<br /> In the present work, I have carried out the measurement of the profiles<br> of the poloidal rotation velocity, the temperature and the density of <br>impurity ions using bidirectional charge exchange spectroscopy (CXS) in<br>the Compact Helical System (CHS). For the purpose mentioned above, this<br>measurement system uses two fiber arrays to view vertically the beam<br> line from up and down sides simultaneously at one vertically elongated<br>section. In Heliotron/Torsatron devices like CHS, the strong parallel<br>viscosity reduces the parallel ion flow velocity which is necessary for<br> incompressible flow conservation when the perpendicular ion flow exists<br> in low aspect ratio tori. This damping is strong in peripheral region<br>where the helical ripple becomes large. However, the poloidal rotation of<br> impurity tons mainly driven by radial electric field determined by the<br>ambipolar condition of the electron and ion fluxes is also large in this <br>peripheral region. Therefore the compensation of the asymmetry of<br>inside and outside perpendicular flows by the parallel flows becomes<br>difficult in this region. When the electrostatic potential is the surface<br>quantity and the poloial rotation of ions is mainly the E×B drift, the<br> flow, especially of the impurity ions having low pressure, should be<br> compressible. Otherwise the electrostatic potential is not the surface<br>quantity or the poloidal rotation of impurity ions is not E × B drift. <br> Investigating this problem is easier in low aspect ratio devices. Therefore<br>this measurement in CHS with the lowest aspect ratio R<small>0</small>/a=5 in helical<br>devices will give the new information about the plasma rotations.<br /> The preliminary measurements of plasma rotations using this system<br>clarified some technical problems in multi-channel CXS. The most<br>important problem was the apparent wavelength shift caused by<br> the spectral fine structure of hydrogen-like ions used in CXS. This structure<br>is the red-side/blue-side asymmetric splitting of the lines due to a <br>relativistic effect and thus cause the red-side/blue-side asymmetry of the<br>Doppler broadened spectral profile. Because of this asymmetry, the<br> wavelength given by single Gaussian least square fitting shows the<br> apparent shifts which depends on Doppler widths. The observed apparent<br> shifts of CVI lines, not due to plasma rotation, in the plasma peripheral<br> region (Ti~100eV) and in the after-glow recombining phase (Ti~<br>30eV) are always red-shifts regardless the direction of plasma rotation. <br> The magnitude corresponds to the velocity error of a few km/s. This<br>direction and magnitude are consistent with the calculation using the<br>collisional l-mixing model. This value is not negligible in CHS plasmas, <br>and thus should be corrected.<br /> The density profile of the fully ionized impurity ions can be measured<br>using the intensity of the charge exchange spectral lines. For this purpose, <br>the initial beam density profile without attenuation was also measured in<br>the torus using H α from the beam. The measured density profile was a<br>broad and inside shifted profile compared with the calculated one. This<br>result means the possibility to measure the parameters on inside of the<br> torus with CXS. However, the calculation of the beam attenuation<br>required that the average electron densities should be less than 2 ×<br>10<sup>13</sup> cm<sup>-3</sup> to avoid the ambiguity of beam attenuation calculation and the<br> degradation of signal level on the inside.<br /> The measurements of the asymmetry of the poloidal flux of fully<br> ionized carbon tons on the inside and outside of the torus were carried<br> out for the magnetic surface configurations with different magnetic axis<br> positions. In inward shifted configurations, the gradients of surface<br> function (dψ/dR) on the inside and outside of the section are almost<br> symmetric. It becomes asymmetric in outward shifted configurations<br> and the strength of the radial electric field will become asymmetric in<br> these configuration.<br /> The asymmetry of the Doppler shifts of the CVI line(Δn=8-7, λ<br>=5290 Å) on the inside and outside of the torus was successfully <br>measured. In outward shifted configurations, the electrostatic potential<br> calculated from this velocity using the momentum balance equation is the<br> surface quantity. The measured density of impurity ions has a hollow<br> profile and is higher on the inside of the magnetic surfaces compared<br> with that on the outside. This inside/outside asymmetry of the density profile<br> can be explained by the poloidal flow conservation on both sides<br> under the damping of toroidal rotation.<br /> In the inward shifted configurations, the density profile is a flat or<br>peaking profile and the inside/outside asymmetry is not clear. The<br>quantitative comparison or the electrostatic potential and the poloidal flow<br> on both sides is difficult in inward shifted configurations because of<br>the intense back and radiation at the inside of the magnetic axis. It<br>causes the degradation of signal/noise ratio of spectrum after subtracting<br>background spectrum. However, this change in the density asymmetry is<br>consistent with the past measurement of the toroidal rotation damping and<br>suggests the poloidal rotation accompanying the inside/outside asymmetric<br>toroidal flow. Therefore the measurement of inside/outside asymmetry of<br> the toroidal rotation velocity is an interesting future theme. |