1. Scaling experiments of perturbative impurity transport in NSTX D. Stutman, M. Finkenthal Johns Hopkins University J. Menard, E. Synakowski, B. Leblanc,R. Bell, S. Kaye, V. Soukhanovskii, D. Darrow Princeton Plasma Physics Laboratory C. Bourdelle CEA, Cadarache M. Gilmore University of California, Los Angeles

2. Motivation Field and momentum input effect on impurity transport in NSTX Part of larger experiment aimed at dimensionless scaling Impurity transport is independent probe of the ion channel: -  i from power balance still uncertain (D. Gates invited talk) - electron channel strongly dominates Tools Brief, non-perturbing Neon puff into beam heated discharges Ultrasoft X-ray (USXR) imaging + high resolution spectroscopy Atomic physics + transport modeling

3. ∫ n e dl (10 15 cm -2 ) T e0 (keV) P rad r/a ≈ 0.8 (W/cm 2 ) I p (kA) reference shot Injection experiments 1.5 MW NBI 5 ms Neon puff Neon injected in L-mode, MHD-free (q 0 > 1), DND discharges Injection is non-perturbing (n Ne /n e ≈ 0.5%) Fast puff enhances contribution of diffusive term t (s)

4. 0.2 0.4 0.6 0.8 1 500100015002000 He-,H-like E (eV) Fully stripped Li-, Be-like Transmission Three diode arrays for peripheral, mid and core Ne charge states Neon contribution from consecutive, reproducible shots Average emissivity from the up/down profiles (symmetric) Inclusion of peripheral charge states (P rad ) improves D, V estimate P rad Be-, Li-like He-, H-like Fully stripped E > 0.4 keV E > 1.4 keV USXR diagnostic

5. r/a D MIST (m 2 /s) He-, H-like Fully stripped E > 0.4 keV E > 1.4 keV Neon emissivity (mW/cm 3 ) puff R (cm) t (ms) Neon penetration at 4.5 kG/1 MA Slow core penetration despite fast rise in peripheral Neon density Best fit modeling (MIST) indicates core D in the neoclassical range No significant pinch velocity (V < 0.5 m/s) Microstability computations predict ITG turbulence intrinsically suppressed in NSTX and not ExB shear effect (C. Bourdelle NF 02)

6. T e (keV n e (cm -3 ) R (cm) t = 250 ms t (ms) Chord position He-, H-like Fully stripped midplane periphery 280.000260.000 260240 t puff 3 kG 0.7 MA 4.5 kG 1.0 MA 3.75 kG 0.85 MA Both peripheral and core charge states penetrate less at higher field, despite very similar electron profiles B t scan at fixed B t /I p reveals strong effect

7. D MIST (m 2 /s) r/a Peripheral turbulence correlation length also strongly decreases (see following talk by M. Gilmore) Comparable effect also observed with B t scaling at fixed I p Note that B t /I p is ‘true’   scaling in a ST: - since B t in >> B t out, varying B t or I p separately, changes   in and   out in different proportions 3.0 kG 0.7 MA 4.5 kG 1.0 MA 3.75 kG 0.85 MA Neon diffusion decreases at higher field

8. Turbulence correlation length also decreases with I p (M. Gilmore) Threshold effect around 1 MA ? Large D decrease for only 20% I p increase Chord position He-, H-likeFully stripped t (ms) midplane periphery D MIST (m 2 /s) r/a 0.9 MA 1.1 MA Global confinement: W tot and  E do not scale with I p /B t W electron /W tot ‘frozen’ at ≈ 0.35-0.40 Changes in Neon transport and edge turbulence not accompanied by changes in W th ion, T i profiles ? (T i data in progress)

9. Possible explanation T i > T e in beam heated NSTX discharges  i  i neoclassical, while  e >>  i Ion power balance: thermal ion profiles governed in fact by the balance between P input, Q i-e and  e  0 H-mode 109070 TiTi TeTe R. Bell, PPPL < ~ S. Kaye, PPPL

10. Summary Sensitive technique for perturbative transport developed Further evidence of ‘naturally’ low particle transport in the NSTX core Initial scaling experiments suggest both particle transport and ion turbulence decrease with  * Global confinement and ion energy content do not scale similarly; negligible ion and large electron conduction a probable reason I p scaling hints at threshold effects