High Speed Imaging of Edge Turbulence in NSTX S.J. Zweben, R. Maqueda 1, D.P. Stotler, A. Keesee 2, J. Boedo 3, C. Bush 4, S. Kaye, B. LeBlanc, J. Lowrance 5, V. Mastrocola 5, R. Maingi 4, N. Nishino 6, G. Renda 5, D. Swain 4, J. Wilgen 4 and the NSTX Team Princeton Plasma Physics Laboratory, Princeton, NJ 1 Los Alamos National Lab, Los Alamos, NM 2 West Virginia University, Morgantown, WV 3 UCSD, San Diego, CA 4 Oak Ridge National Laboratory, Oak Ridge, TN 5 Princeton Scientific Instruments Inc, Monmouth Junction, NJ 6 Hiroshima University, Hiroshima, Japan TTF Meeting, Madison Apr. 3, 2003

Outline Goals Fisheye view of NSTX Gas puff imaging diagnostic GPI image and time series analysis Summary Plans

Understand edge turbulence by comparing turbulence measurements with theory + simulation Explain and predict H-mode threshold and pedestal Explain and predict transport through SOL to wall Explain and predict transport of impurities into plasma Explain and predict density limit ? Goals

Fisheye View of NSTX LANL camera, 10 µsec/frame at 1000 frames/sec no filter (mainly D light) Magnetic structure of edge plasma

Gas Puff Imaging Diagnostic Look at He1(578.6 nm) from gas puff I  n o n e f(n e,T e ) View along B field line to see 2-D structure  B GPI view 16x32 cm

Typical GPI Image out sep. n e /10 11 cm -3 T e (eV) n(He o )DEGAS HeI R MP (cm) GPI HeI Use typically 10 µsec exposure time (  ac ≈ 40 µsec) Average HeI light intensity peaked near separatrix PSI camera frame 80 x 160 pixels

GPI Diagnostic Interpretation HeI light emission “I” visible where 5 eV < T e < 50 eV I  n e  T e    where   ≈  0.5  and  ≈  0.7  near center of cloud, with 0.4 <  < 1 and 0 <  < 2  over most of cloud Space-time structure of I similar to n e , but  I/I ≈   n e /n e Fluctuation spectra of I similar to probe and reflectometer Fluctuation level of I consistent with TS data ( ≈10-60%)  GPI light gives approximate structure of edge turbulence

High Speed Imaging of NSTX Edge CCD camera with 100,000 frames/sec at 10 µsec/frame for 28 frames/shot localized structures can move outward at ≈ 10 5 cm/sec

Profiles from Typical GPI Images HeI profiles narrow from OH to L-mode to H-mode  I/I and L pol derived from 28 images for each shot L(cm) A(k)  I/I k pol (cm -1 )

Poloidal Correlation Length and k-spectra L pol ≈ 4 cm or k pol  s ≈ 0.2 (similar to other experiments)  I/I lower in H-mode than L-mode (with much variation)

Time (sec) Time Series of GPI Light Flucutuaions HeI digitized over 1.5 cm diam. chords through images Relative fluctuation level larger as R increases (≈ images)

Statistical Analysis of Typical Chords Autocorrelation times typically 40 ± 20 µsec Frequency spectra broad over ≈ kHz Autocorrelation function Frequency spectrum Probability distribution function

L-H and H-L Transitions L -> H in ≈ 100 µsec with obvious precursor H -> L in 30 µsec with outward radial pulse 100 µs 1 ms

Slow Imaging of H-L Transition H-L (#105710)L-H-L (#105564) Taken with 10 µsec exposures at 1000 frames/sec For this display, outward direction toward lower right

Motion of Coherent Structures Track high intensity “blobs” over 28 frames of movies Broad distribution of velocities and intensities

Summary of Results So Far Images consistent with previous measurements - large fluctuation level in edge - broad frequency and k-spectrum - approx. isotropic structure  B Coherent structures seem to move through edge - “blob- like” look similar to DIII-D IPOs - “wave-like” look similar to EDA, QCM H-mode generally more quiescent than L-mode - considerable variation in behavior - transitions can happen very fast

Plans for Comparison with Theory Using DEGAS-2 or related atomic physics models: Compare GPI with BOUT simulations for H- and L-mode (Xu and Nevins) Compare motion of GPI “blobs” with blob model (D’Ippolito and Myra) Compare with other simulations…

Plans for Additional Measurements Capture H-mode transition with high speed camera Get better data on zonal flows in images and chords Examine turbulence nearer density limit Look during RF heating, e.g. co- vs. ctr. current drive Make systematic scans of q(a), rotation, Z eff, etc. Make quantitative comparisons with other diagnostics

R MP (cm)   HeI (rel.) Alpha Correlation length (rel.) Alpha Fluctuation level (rel.) Alpha (1) (2) (3) Amplitude (rel.) Wavenumber (rel.) (a) (b) (c) (d) (3) (2) (1)

R MP (cm)  I/I (rel.) LCFS #105637

Frequency (kHz) x probe before puff probe during puff o GPI during puff Power (rel.)

Frequency (kHz) x reflectometer before puff reflectometer during puff o GPI during puff Power (rel.)

R MP (cm)  n/n  n /n  n/  n  T/T  T /T  T/  T