Fast-framing movies of TFTR plasma events

A fast framing camera (Kodak Ektapro EM1012) running at up to 2000 frames per second was used at TFTR to observe fast edge phenomena and disruptions. In these short clips, digital sequences have been turned into MPEG movies. For improved clarity, white/black has been inverted...... so darkness is represented by white. The horizontal (vertical) resolution is 239 pixels, while the vertical (horizontal) resolution is either 192 pixels ("square" frame) or 96 pixels ("rectangular" frame).

The fast-framing Kodak system is used to image different portions of the view above. This image of the interior of TFTR's chamber was obtained with a fish-eye lens and a different imaging system.

This web page serves as a link to the clips referenced in the following two papers:

"Fast Imaging of visible phenomena in TFTR", R. J. Maqueda and G. A. Wurden, submitted to Nuclear Fusion.
"Images of disruption effects in the Tokamak Fusion Test Reactor (TFTR)", R. J. Maqueda and G. A. Wurden, submitted to IEEE Transcations on Plasma Science.

where you can find additional information on the diagnostic and physics.

If you need a viewer for the MPEG clips follow this link.

Disruption caused by a "locked mode" or SMP (discharge 103681)

Clip(2.4 MB) - Waveforms

(2000 frames per second with 30 microseconds exposure of each frame, no interference filter.)

Clip shows the inner third of the toroidal plasma column with the armored inner wall of TFTR's plasma chamber at the left. (This inner wall is also called the bumper limiter.) The outer wall of the toroidal chamber is not shown. The tiles on the bumper limiter are nominally 10 cm x 10 cm in size.

This disruption that takes place ~2.425 s after discharge initiation was caused by a rotating m=4 n=1 mode that locks to the vessel wall. A ~0.6 MA toroidal current remains after the disruption. This current is mostly carried by runaway electrons created by the strong electric fields present during the quench of the toroidal (ohmic) plasma current. These high energy (~20 MeV) electrons collide into plasma facing components causing damage.

The damage manifest itself by a (horizontal) shower of debris that "falls" over the bumper limiter above the device midplane. Although not as spectacular as the example shown here, flying debris is almost always seen after disruptions. The diameter of the hot-glowing flying particles is estimated to be up to 1 cm and their speed is of the order of 100 m/s. It is speculated that the debris seen in these images is part of the Faraday shield for the RF heating antenna (IBW) that was located approximately opposite to the section of bumper limiter shown and that was noticed missing a few weeks later.

Lithium pellet injection (discharge 92977)

Clip(340 kB) - Waveforms

(500 frames per second with 100 microseconds exposure of each frame, no interference filter.)

Clip shows the inner bumper limiter of TFTR's plasma chamber and, on the right, portion of the toroidal plasma column.

Two lithium pellets are injected into the torus at the end of the discharge (current ramp-down) to condition the bumper limiter for the next experiment. These pellets are injected from the left of the images at 4.1 and 4.3 s after discharge initiation. As the pellets ablate, a short (one frame) flash is seen illuminating the bumper limiter on the left of the image. After this flash, the free electrons created during the ionization of the pellet material enhance the bremsstrahlung radiation from the plasma that is observed as a toroid of light.

Shallow lithium injection

A new experiment on plasma-wall interaction was performed close to the end of TFTR's operation in which a lithium aerosol was injected into the edge plasma by directing a 30 Hz pulsed YAG laser onto a cup filled with molten lithium. This experiment, termed DOLLOP (Deposition Of Lithium by Laser Outside the Plasma), yielded significant improvements in plasma performance. Two clips obtained during the injection of the lithium aerosol are shown below.

While the individual sub-millimeter aerosol particles can be observed with either a neutral lithium interference filter or no filter (discharge 104023), the use of a Li+ filter at 548.5 nm allows the ablation cloud that extends toroidally to dominate (discharge 104298).

Discharge 104023: Clip(4.4 MB)

(1000 frames per second with 30 microseconds exposure of each frame, no interference filter.)

Clip shows the bottom edge of the bumper limiter close to the location of the DOLLOP source.

The neutral beam injection (60% co-current) that is turned on at 4.0 s after discharge initiation to heat the plasma (~18 MW), and its associated momentum imparted on the toroidal column, causes the "DOLLOPs" to move only in one toroidal direction: away from the camera. This behavior lasts until 5.0 s when the auxiliary heating is turned off.

Discharge 104298: Clip(520 kB)

(1000 frames per second with 200 microseconds exposure of each frame, Li+ interference filter @ 548.5 nm.)

Clip shows the toroidal plasma column and the bumper limiter (left). The horizontal on this image is tilted ~45 deg. clockwise.

Edge turbulence (discharge 103782)

Clip(570 kB)

(2000 frames per second with 20 microseconds exposure of each frame, no interference filter.)

Clip shows the bumper limiter above the midplane (straigh-on view). Note some hot spots on the edge of the tiles that cover the limiter.

Although the framing speed of the Kodak system is not enough to follow individual filaments from one frame to another due to their autocorrelation time of less than 150 microseconds, the fast imaging allows the evolution of the characteristic wavelength of this turbulence to be followed as a pellet is injected into this auxiliary heated (~14 MW) discharge.

The pellet, injected into the plasma at ~3.0 s from discharge initiation, produces a momentary saturation of the images. Soon after injection, the filaments can not be observed due to either a much shorter autocorrelation time (<20 microseconds) or to their actual suppression by pellet modification of the plasma profiles. Once the turbulent filaments reappear, the poloidal spacing (i.e., "wavelength") between filaments is apparently shorter than before the injection. At around 3.037 s there is some indication of a very fine structured filament pattern, particularly closer to the midplane, with poloidal spacing smaller than 7 cm.

Ricardo J. Maqueda and Glen A. Wurden
Los Alamos National Laboratory, P-24 Plasma Physics, Los Alamos, NM 87545, USA
Oct. 1998