Alcator C-Mod Weekly Highlights Sept. 13, 1999 Plasma operations continued at Alcator C-Mod last week. Three run-days were scheduled and completed. A total of 52 plasma shots were produced. This week's runs were in support of the continuing ICRF commissioning activity. Physics and Analysis --------------------- A quasi-coherent mode in the frequency range from 50 kHz up to 300 kHz is often observed by both reflectometry and phase contrast imaging (PCI) during EDA H-modes. This mode was believed to be located on the EDA H-mode pedestal; however, no direct experimental evidence had been reported due to the limitation of both diagnostics. After configuring the reflectometer to take fluctuation data in four channels, (the channel with highest frequency, 110 GHz, has the critical density of 1,5e20 m^-3), the location of the mode has been roughly identified. Evidence shows that in some shots with low H-mode pedestal height, the mode seen from the 110 GHz channel can appear or disappear dependent on whether the critical density is on or inside the pedestal. The density profile used for the comparison is from the visible bremsstrahlung measurement. Comparison among different reflectometry channels also indicates that the mode starts to appear at the top of the pedestal and soon fills the whole barrier region. Analyses of measurements from recent experiments with high density ohmic plasmas have shown that the divertor bypass flaps do not directly affect the threshold for divertor detachment. Open bypass flaps provide for a faster density increase (for the same gas puff) than closed flaps. Thus, while the density at which detachment is observed (ne_bar~2.2e20 m^-3) does not depend on the state of the bypass flaps, that density is reached earlier when the flaps are open. In addition nitrogen impurity puffing into these higher density discharges showed that the detachment threshold and depth of detachment were modulated by the amount of nitrogen puffed. This is not a new result. However, new images of the detachment in D_alpha and D_gamma light show the detachment modulation clearly. A region of recombination is observed to move from the outer strike point up to the x-point along the outer leg as the detachment deepens. After the nitrogen puff, the recombination region retreats back to near the outer strike point. In the cases of deepest detachment the D_gamma emission reaches closed field lines inside the separatrix. Observation of ohmic H-modes with a large electron density gradient in the core of the plasma was reported. The plasma core and edge density profiles are measured using Thomson scattering diagnostics. In case of "normal" H-modes (both ohmic and RF induced) the profile is characterized by steep edge density gradients and more or less flat core profiles. The long H-modes observed during two recent ohmic runs (05/27/99 and 08/31/99) developed a sharp core density gradient in addition to the edge pedestal. This core density gradient develops gradually during the H-mode and is not a continuation of the edge pedestal but starts well inside the separatrix at r_mid of about 0.87 m while the edge pedestal is located around the separatrix at r_mid of about .89 m. The resulting core density profile can be strongly peaked, with the central density as high as 6e20 m^-3 and the pedestal height around 2e20 m^-3, and is similar to the profiles observed when internal transport barriers are formed in the plasma. Further investigations of these phenomena are under way. Prior observations with the BES viewing optics had not seen any evidence of poloidal propagation of D-alpha fluctuations, which we attributed to our views (almost strictly toroidal) not being parallel to the field lines in the edge. The chord-integrated signals were therefore averaged over many flux tubes. The fiber holder was rotated in a direction which effectively increased the view angle, after which we saw dramatically increased phase shifts between nearby channels. The phase shifts were largest during Ohmic plasmas and dramatically decreased during RF. This implies the poloidal ExB velocity dramatically increases during RF. We are presently working on quantifying these observations.We had not seen any evidence of poloidal propagation of D-alpha fluctuations, which we attributed to our views (almost strictly toroidal) not being parallel to the field lines in the edge. The chord-integrated signals were therefore averaged over many flux tubes. Matt rotated the fiber holder in a direction which effectively increased the view angle, after which we saw dramatically increased phase shifts between nearby channels. The phase shifts were largest during Ohmic plasmas and dramatically decreased during RF. This implies the poloidal ExB velocity dramatically increases during RF. We are presently working on quantifying these observations. ICRF System ------------ Three experimental runs were dedicated to RF conditioning. The focus was continuing the conditioning for J-port, and bringing up D and E-port antennas. The J-port antenna continued to show improvement: heating was observed with various phases and 1.5 MW injected power was obtained routinely. The relative strap current phase (varies the antenna wave spectrum) was scanned on J-port. Intial analysis suggests that both the heating and impurity influx vary with phase. For added experimental flexibility with the J-port antenna, a design for the external antenna connection loops giving an additional 180 degree current strap phase shift has been started. This will permit operation with the four straps phased 0-pi-0-pi, which is not accessible with the present connections. ICRF power up to 1 MW from the D-port antenna (FMIT#1) has also been added, both in parallel with the J-port power and separately. EDA H-mode conditions were obtained with D-port operated alone at the 1MW power level. An arc in the transmission line to E-port (FMIT#2) was located and repaired. We anticipate operation of this antenna into plasma this week. Work on eliminating a high frequency oscillation in FMIT#1 and #2 has centered upon changing the FPA screen voltage to -1500 V from -960 V and the FPA bias voltage to -600 V from -500 V. This has resulted in a higher gain final stage. The solution to the high frequency oscillation suggests that the cause was high harmonic generation from a clipping drive component; this would result from the power feedback loop attempting to overcome loss in output stage gain due to excessive screen supply droop. Advice and circuit drawings on adding screen supply regulation have been obtained from the PPPL RF engineering group. The temporary high voltage modifications to FMIT#1 and #2 were insufficient in extremely humid weather as experienced last Friday. The high voltage connection in the output cavity was arcing to such an extent that these two transmitters could not be operated for part of the run. All crowbar protection circuitry operated normally, and there was no damage to any of the hardware. Purging of the cavity with dry N2 gas improved the situation. Design modifications based on FMIT#3 and #4 are being pursued to improve high voltage operation. Travel and Visitors ------------------- Miklos Porkolab attended the FESAC meeting last week on Wednesday and Thursday in Gaithersburg, Md. Gary Taylor returned from PPPL and worked on the GPC2 ECE instrument. He made good progress in reducing the pickup which has been experienced during operation of the new J-port RF antenna. Matt Sampsell (U.Texas) visited C-Mod last week to make some adjustments to the BES viewing geometry, as detailed above. David Winslow (U. Texas) began a visit to CMOD to start driven divertor probe experiments and explore divertor fluctuation measurements. R. Chatterjee (U. Texas) began a visit to continue work on temperature fluctuations with the ECE radiometer. Jesus Ramos attended the US-Japan workshop on high-beta systems at GA. He gave a talk on "MHD-stability and current-drive studies for advanced tokamaks".