High-speed imaging of the SSPX plasma

Transcription

1 High-speed imaging of the SSPX plasma Carlos A. Romero-Talamás, Paul M. Bellan, SSPX team * California Institute of Technology 1200 E. California Blvd. Mail Stop Pasadena, CA, U.S.A * Lawrence Livermore National Laboratory Livermore, CA, U. S. A. 1

2 Abstract A specially designed high-speed imaging system has been installed at the Sustained Spheromak Physics Experiment (SSPX), located at the Lawrence Livermore National Lab (LLNL). We have observed bright patterns close to the electrodes during SSPX current sustainment and ramp-down. In particular, bright circular patterns are seen on the end face of the cylindrically shaped cathode. These circular patterns are azimuthally segmented, approximately concentric with the cathode axis, and appear at discrete radial positions. The anode, in contrast, has patterns near the divertor region that are not as well defined as those at the cathode. The patterns on both electrodes evolve in time. Since the high-speed imaging system can take up to two images for every SSPX shot, it is possible to obtain information about how the patterns evolve. The cathode pattern morphology and position did not vary significantly in less than 0.1 ms. At times greater than 0.1 ms, cathode patterns elongate toroidally in a constant direction. It has been possible to image the open flux surfaces linking the hole of the torus, and measure the diameter of these flux surfaces. We conjecture that the patterns on the cathode are the intersection of the open magnetic flux surfaces with the cathode. 2

3 Location of camera at SSPX The SSPX chamber has an opening in the flux conserver and several viewports for instrumentation and measurements Vessel Bellows ICCD Figure 1 3

4 Camera viewports on SSPX Viewport for photography (22.5 ) Viewport for photography (270, shared with electron gun) Viewport for photography (157.5 ) Figure 2. TOP VIEW OF SSPX CHAMBER 4

5 Re-entry ports and windows Re-entry ports designed to go through the opening of the flux conserver, with a window and a shutter mechanism. Re-entry ports are mounted on a movable bellows to retract them when no in use. Shutters close to protect the windows from being coated by gettering. Three units built and mounted on available ports at SSPX (see Fig. 2). Ball drive actuator Flux conserver Macor shutter 6 nipple Vacuum vessel wall Window Bellows Figure 3. Bellows with ball drive linear actuator Figure 4. Re-entry port mounted on bellows. Shutter closed and port retracted (upper images), and shutter open and window ready for viewing inside flux conserver (lower images). 5

6 The camera DiCam-Pro camera features: 12-bit dynamic range Fast shutter down to 3 ns High resolution sensor 1280 x 1024 (VGA) ICCD Cooled down to -12 C Two distinct shots with an interframing time of 500 ns 25mm MCP-image intensifier Spectral sensitivity UV to NIR Figure 5. DiCam-Pro camera (from Cooke Corp.) 6

7 Magnetic shielding The manufacturer of the camera specified that the field around the camera should not exceed ~10-4 Tesla. However actual field is ~ 10-2 Tesla. We designed a shield made of mu-metal, with inner layers of steel foil (separated by layers of plastic). The shield achieves the required attenuation of ~100. Figure 6 Figure 7 7

8 Challenging optical access Small opening in flux conserver (5 cm), strong magnetic environment and harsh plasma conditions made it necessary to have the camera outside of the flux conserver and vessel. High FOV required (~ ) To meet requirements three options were considered: Fiber bundle. Easiest to implement and flexible (literally!), but worst resolution. Screen concept. Partially tested (successfully) at SSPX; intermediate resolution. Relay lens. Best resolution, hardest to implement (hard to align and position lenses). Figure 8. Fiber bundle (from Schott Co.) Melting Figure 9. Screen concept. Figure 10. First re-entry port and shutter system built failed after hundreds of plasma shots. It was replaced with a retractable system. Erosion of ceramics from plasma impingement 8

9 Optics Relay lens Figure 11 Figure 12. Interior view of the flux conserver as seen through the relay lens. Figure 13 Figure 14. Test of relay lens at Caltech Relay lens consists of two achromat lenses and the wide angle lens that is placed close to the window of the viewport. Armature allows adjustment of distance between elements (required length is 90 cm). The image is projected directly onto the ICCD. 9

10 Optics Double branch fiber bundle Allows to take two simultaneous views from different angles, using the same ICCD. Flexible, so it can be easily changed between ports. High-speed optics, but lower resolution than single relay lens. Figure 17. Two distinct images seen (using a magnifying lens) on common end of double fiber bundle Figure 15. Double fiber bundle, 4.5m length. Resolution: 50 lp/mm. Optical area of each branch: 6 x 6 mm^2 (built by Schott Optics). Figure 16. Common end of double fiber bundle. Figure 18. Fiber bundle with special relay optics. 10

11 The images Images can be obtained and precisely timed by synchronizing the trigger of the camera with the start of plasma breakdown. Two images can be obtained at each shot. All images have been obtained in the visible spectrum, including all wavelengths (color added to illustrate light intensity levels). 11

12 So, what do we look for in the images? Formation dynamics Quality of confinement Validate theory and numerical simulations. Troubleshoot malfunctions in equipment. Figure 19. Concept of particle trajectories on open flux surfaces at SSPX. Images taken during sustainment of gun current, Ig, suggest that current channels formed on open flux surfaces remain essentially unchanged. 12

13 Plasma stages at SSPX Breakdown and plasma ejection? Sustainment Decay I (A) Figure 20. Gun current trace for a typical SSPX plasma shot 13

14 Plasma evolution after breakdown Initial bias field Gas puffed Breakdown (t=0) Ballooning out of gun (t ~ 35 µs) Ballooning reaches bottom of flux conserver (t ~ 40 µs) Transient central column formation (t ~ 50 µs) Figure 21. Evolution of B field inside SSPX flux conserver 14

15 Theater curtains coming down t = 30 µs t = 35 µs Figure 22. Images of shot 8027 shortly after breakdown. Expansion front merges resembling descending theater curtains. Axial speed of expansion front ~ 30 km/s Figure 23. From top to bottom: Gun current (amperes), gun voltage (volts), edge of flux conserver poloidal and toroidal B fields at mid-plane (tesla). Dashed lines correspond to time of images. As plasma expands resembles descending theater curtains. Features merge as the plasma moves axially and toroidally. Magnetic field in the flux conserver increases as plasma fills the volume. Edge magnetics indicate correlation of bright features with magnetic field intensity. 15

16 Two different central columns Shortly after breakdown (< 100 µs) a short-lived (< 10 µs) central column is observed. The column disappears abruptly (< 1 µs). During sustainment of gun current, a distinct central column is observed. This column survives as long as current is sustained. 16

17 Transient central column The central column exhibits reduction in diameter as time evolves. Plasma moves helicoidally around the column. Eventually the column bends impulsively, and seems to disappear. t = 70 µs t = 73 µs Figure 24. Filaments wrapping around central column. Pitch angle of helicoidal filament ~20. Gas: helium. Shot t = 80 µs t = 85 µs Figure 25. Column bending and disappearing. Gas: helium. Shot Figure 26. Current, in amps, (upper) and voltage (lower) traces for shot Dotted lines correspond to time of images. 17

18 Stereoscopic view and tomography Experimental setup Double branch fiber bundle Lens Lens t = 50 µs t = 55µs Flux conserver region Figure 27. Experimental setup and irregular grid used for 2D tomography computations (shown with coarse cells for clarity). Shot Views are apart. 18

19 Tomography reconstruction Only 2 projections available, but high resolution. 2D tomography is based on back projection (at midplane) and maximum entropy techniques. Accuracy of reconstructed profile is limited 3D tomography is work in progress. Back projection + Maximum entropy Figure 29. 3D tomography is work in progress t = 50 µs Figure 28. 2D Tomography at midplane. 19

20 Concentric structures on cathode surface During current sustainment, we observe bright patterns on the electrodes Sustainment stage Figure 30. Digital stretching of the cathode region. Digital mapping (stretching) of the image helps visualize patterns as if seen from the bottom of the flux conserver Figure 31. L 2 (right) is the resulting dimension and apparent position of the original image (left) with dimension L 1, after digital stretching. 20

21 We can measure the pattern elongation Figure 32. Red marks indicate pattern elongation after 0.2 ms. The patterns change in time Figure 33. (Animation.) Figure 34. Pattern elongation at the cathode versus time 21 Figure 35. Pattern elongation velocity at the cathode versus time

22 Intensity changes much faster than patterns A correlation of light intensity was made at different radii, between pairs of images. Variation in intensity is much faster than change in pattern shape. Figure 36. Angular shift of highest peak (per discrete annular region) of neutral light emission versus radius. Figure 37. Example of annular region for angular shift measurements (top). Average of all images, showing bands where patterns are most likely to appear (middle). Traces how the circular bands in the average image correspond to the most prominent peaks in light intensity shift plot shown in Figure 36 (bottom). 22

23 What causes the cathode patterns, and what makes them elongate? We conjecture the patterns are caused by current-carrying filaments intercepting the cathode. The filaments go through the hole of the spheromak. E x B and curvature drifts are good models for the speed of elongation observed. Figure 38. Estimation of radius of curvature of the magnetic field near the cathode, from images. Figure 39. Typical rendering of the sustainment stage using the MHD equilibrium code CORSICA. Reproduced from B. W. Stallard et al. Phys. Plasmas, 10, 2912 (2003). 23

24 For Figure 40. Coordinate system used for the 1-D calculation of the electric field across the ionization sheath near the cathode. then 24

25 Reversing the B field, changes elongation direction 25

26 Discrete flux surfaces? Sustainment central column Figure 43. Dotted lines mark the boundary of the faint central column. A nested column is also apparent above. Gas used: hydrogen. Shot Figure 42. Nested flux surfaces. The outermost flux surface is indicated with an arrow, and is also offset from the electrode. Compare with purple line of CORSICA rendering on Figure 44. Shot t = 3.3 ms Figure 44. Sample CORSICA output. Purple line marks boundary between open and closed flux surfaces 26

27 Image measurements vs. CORSICA Figure 45. Comparison of plasma column radius (open flux surfaces) between experiment and numerical reconstruction using CORSICA. 27

28 Filamentation after sustainment At sufficiently low values of gun current, the column breaks off into filaments. t = 3800 µs t = 3900 µs Figure 46. Filamentation of the central column at low gun current values (Ig < 50 ka). 28

29 Summary High-speed imaging helps understand formation dynamics. Cathode pattern elongation consistent with E x B and curvature drift theory. Study of electrode patterns relevant to plasma-wall interactions and SOLs. Measurements of the spheromak hole useful in validating and further constraining MHD equilibrium simulations (CORSICA). High-speed imaging system yields information that can be used to troubleshoot malfunctions in SSPX. 29