Alcator C-Mod Weekly Highlights September 18, 2000 Alcator C-Mod continued plasma operations last week with four runs planned and four completed. Startup reliability was 90%, with 86 research grade plasmas produced. Tuesday's run was an ohmic run dedicated to high resolution imaging and probe measurements of the edge fluctuations. On Wednesday a half day RF development run was carried out to study the J-Port antenna in the 2-strap configuration. On Thursday a careful study of the H-mode density threshold for EDA H-modes was conducted, and on Friday ohmic H-modes were produced and measurements of plume formation and edge fluctuations were made. A boronization was carried out on Monday night. Physics: During Tuesday's run the new 2D turbulence imaging diagnostic was used to image an outboard gas puff. A 6 cm (radial) x 6 cm (poloidal) portion of the outboard plasma edge and SOL is imaged with ~1 mm spatial resolution. Fast snapshots (~1 microsecond exposures) are obtained every 16.7 ms. This diagnostic adds to the complement of edge/SOL turbulence diagnostics, which include the scanning probes and a system that images emission from discreet portions of the camera's 2D field onto fast (250 kHz) filtered diodes. The first goal of this run was to correlate the fluctuations in I_sat and/or V_f (as measured by the F-port Scanning Probe) with emission fluctuations measured by the fast diodes viewing a volume in the gas puff that is connected along a flux tube with the Langmuir probe volume. This measurement would establish the physical basis for the emission fluctuations (presently believed to be density fluctuations) and would determine a lower limit on the parallel correlation length. We observed some influence of the FSP scan on the fast diode signals (which image at the A-B limiter approx 180 degrees away toroidally) when the current was between 0.72 and 0.70 MA. Determination of the degree of correlation awaits further analysis. Certainly the emission fluctuations from the three radially adjacent views were correlated and had a noticeable phase difference. The second goal of the run was to observe the turbulence, i.e. the RMS fluctuations in Gas Puff Emission, I_sat, V_floating vs R, and the 2D spatial structure of the emission, as we scaled the ion gyroradius (through Bt), and the collisionality (by changing separately both n_e and parallel connection length). The collisionality is thought to be a key parameter, since recent studies by LaBombard [APS 1999] show that the edge diffusivity increases strongly with local collisionality. We obtained data at each of three toroidal fields. The 2D images appeared to have larger turbulent structures at the lowest field (2.6 T). Good data was also obtained throughout the current scan. The fast diodes show the normalized RMS fluctuation level peaking at current of approximately 0.85 MA. The density scan was carried out to beyond the Greenwald limit, somewhat before which, the plasma detached poloidally. Contrary to our previous experience, this poloidal detachment was quite stable, lasting ~400ms. The highest densities were 1.2-1.25 times the Greenwald limit. Unfortunately, during the poloidal detachment, the gas puff emission was small compared to the large background. The probe, however, saw a strong reduction in fluctuation level during the detachment. Good images of the turbulence up to the time of detachment (when n/n_GW was ~0.75.) were obtained. Re-configuring of the PPPL J-Port ICRF antenna from its normal 4-strap mode to 2-strap operation was completed on Wednesday morning of last week, and conditioning and comparison with the other two antennas was initiated in the afternoon. Initial results did not agree with a similar comparison performed in June because of an assembly mistake in an external feed connection. This system will run again this week with the feed connection properly installed. On Thursday, the role of density on the EDA/ELM-free boundary was studied. We also wished to investigate the ELMs previously observed in low density operation. It was conclusively demonstrated during this run, and well documented, that there is indeed a threshold in target density for getting EDA H-modes. For a target density NL of 0.6e20 m-2 or lower, H-modes were ELM-free; at higher density they were EDA. There seems to be a progressive onset of EDA with increasing density, as evidenced by Dalpha levels, particle transport (rate of ne and radiation rise) and density fluctuations. Very good data was obtained showing the disappearance of coherent fluctuations in ELM-free H-modes and their appearance in EDA; some nice comparisons between the regimes were made. A correlation of transport and mode amplitude and perhaps frequency should be possible after more analysis. The RF performed very well, coupling up to 2.8 MW from D and E port on most shots, and at a variety of densities and requested waveforms. This reliability was required, as the window for ELM-free operation was close to the low density limit for H-mode production, and power levels of about 2 MW of RF were required to exceed the threshold. When the density was lowered to 3 to 4e19 m-2, H-modes became short lived and weak. The ELM-free H-modes showed quite impressive performance, with core Te and Ti both up to 5 keV and pedestal Te ~900 eV. Stored energies were up to 180 kJ, and H-factors ~2.5 to 3. Very fast edge rotation was observed, up to 100 km/sec vs the usual 40 km/sec, from analysis of edge fluctuations. Edge TS measured steep pedestal Te gradients of over 200 keV/m; one might speculate these are related. However, as is usually found during ELM-free operation, impurities tended to accumulate, eroding the pedestal, and the good performance was transient. EDA H-modes were steady state but did have rather lower confinement, with stored energy ~140 kJ and H-factors ~ 1.3-1.4. Since the heating is presumably similar, the cause of this increased energy transport bears investigation. The density threshold was not markedly different at 1 MA vs 800 kA, however the H-mode power threshold was much lower (PRF 1 MW vs 2 MW) at the same density. Several shots used slow power ramps, which should give good data for LH thresholds and studies of pedestal evolution at different densities. One goal of running ELM-free plasmas was to investigate whether they had any large ELMs, as seen on June 13. Despite the high pedestals, no evidence of ELMs or unexplained HL transitions was seen; all H-modes ended due to high radiation. Collisionality was undoubtedly low, but pressure gradients measured by edge TS apparently did not exceed our usual range; a stability analysis should be undertaken. Last Friday's run was dedicated to completion of part 2 of MP 276, "Transport Experiments in the Scrape-Off Layer during Ohmic L- and H-mode using Gas-Injection Plumes." Plumes generated via gas-injection of deuterated ethylene (C2D4) through the F-bottom scanning probe are imaged from two near-perpendicular locations (F-side, F-top) with coherent fiber bundles optically coupled to gated, intensified CCD cameras through beam-splitter optics and bandpass interference filters. The optical system allows for imaging of C^+1 and C^+2 emission simultaneously at each view location. On Friday, an Ohmic H-mode run was carried out at fixed current (I_p = 0.8 MA). A toroidal field ramp was used to get into H-mode, and the density was varied to achieve both ELM-free and EDA phases. Plumes were generated at various depths in the scrape-off layer, ranging from right at the separatrix to 1 cm outside, and at a range of times in the discharge to study various phases of the H-mode. The first part of the run was dedicated to the investigation of plumes during ELM-free H-modes for various injection depths. No clear differences in the cross-field plume structure were seen for these cases. One explanation is that though the electric field is varying strongly as the injection depth varies, the local electron density and temperature, and hence the ionization lifetime for the impurities, is also varying strongly, and thus these two effects on cross-field dispersion are effectively cancelling. In contrast, the parallel structure of the plumes seemed to change with injection depths, and parallel flow reversal was evident for plumes closer to the separatrix. Further analysis is required. The second part of the run was dedicated to studying plumes close to the separatrix for both ELM-free and EDA phases. In general, plumes generated in EDA exhibited less cross-field dispersal than plumes in ELM-free phases. On one shot during the EDA H-mode phase, the cross-field dispersal of the plume was indicative of an inward directed radial electric field. Unfortunately, the probe experienced serious arcing during this time, and the probe data may not be useful. The probe was switched to a double-probe mode for the remainder of the run to alleviate this problem. However, the cross-field behaviour was not reproduced, though there was some evidence of the plumes experiencing shearing due to being generated in a region of radial electric field reversal. One reason the full reversal of the plume dispersion is not being seen consistently has been tracked down to an issue involving camera exposure timing relative to probe injection timing. This issue is being sorted out, and more near-separatrix plume data will be collected to help illuminate what the radial electric field structure is in this region. Engineering: We continue to prepare for the long pulse operation expected to begin at the end of October. Installation of new power supply diagnostics to monitor solid state components continues as does analysis of critical components. A new system to aid the initiation of the glow discharge at low fill pressures is nearing completion. An RF voltage will be imposed on the electrodes together with the high voltage DC usually used for breakdown. During the next up-to-air, modifications are planned to the cryostat, cylinder, and vacuum vessel. These changes will provide a more reliable seal of the cryostat, new feedthroughs for cabling entering the cryostat, the ability to remove horizontal ports much more easily and reliably and with minimal amounts of disassembly, and possibly new access ports for RF systems and diagnostics. 3D models of all these components are being developed in preparations for analysis, design, and eventual fabrication. Detailed design of the new inner divertor upgrade is nearing completion. Vendors for tiles, girdle backing plates, and c-plates are being selected. Tools for installation have been designed and are being fabricated. Models of critical components will be produced using new rapid prototyping techniques and evaluated before actual hardware for installation into C-Mod is fabricated. The reliability of the diagnostic neutral beam was improved last week. During the previous week, there were a number of instances of apparent arcdowns during C-Mod shots. These were in fact nuisance faults caused by sags in the HV input to the MODREG during C-Mod shots. They were reduced by first recognizing the actual problem, and then simply increasing the HV input to the MODREG. This issue will continue to be pursued during this week's run period. Spectroscopic observations of the beam were made to the red side of H-alpha. These measurements are intended to benchmark the BES and MSE observations. The signal-to-noise is low, but the available data is being analyzed. Perveance measurements were made using the BES diagnostic to acquire the width of the beam at the plasma edge. The initial analysis is promising. Continued analysis will likely suggest the need for some additional experiments at other plasma conditions. Pablo Acedo (Carlos III University, Madrid, Spain) ended his several week stay last week. During his time at MIT he worked with Rick Murray on the tangential two-color interferometer. New electronics were added to separately measure the phase shift of the 532 Angstrom green laser beam. We had been measuring only the subtraction signal derived from the frequency doubled IR and fundamental green signals mixed together in the I/Q detector electronics. In addition, electronics to use this new signal to control the alignment of the interferometer by driving a PZT controlled mirror were developed and and made operational. These additions should help determine the cause of a slow drift in the interferometer subtraction signal. Travel and Visits: Ben Carreras of ORNL is visiting this week and next, working on analysis of density and temperature at the LH transition, as well as edge fluctuations. Stew Zweben was at MIT on Tuesday to participate in an experiment on the imaging of edge turbulence incorporating his new 2D imaging diagnostic. Norton Bretz was at MIT Wednesday and Thursday in an attempt to obtain spectra of the Doppler-shifted light emitted by the University of Texas diagnostic neutral beam. Only extremely weak signals were observed. David Winslow (UT-FRC) visited to continue development of the Texas Turbulence Probe. Ron Bravenec (UT-FRC) visited to work on measurements of the beam perveance with the BES diagnostic and to assist in operating the DNB. Last week Martin Greenwald was in Oak Ridge to participate in a meeting of the ESnet Steering Committee. The principal item of business was the preparation of the ESnet program plan.