Alcator C-Mod Quarterly Progress Report - July 2001

The primary activities at Alcator C-Mod during the third quarter of FY01 were:

assessment of the performance of the J-port antenna and of the DNB/diagnostics.

Science Results

Core Confinement and Transport

Enhanced D_a H-mode/Quasi-coherent Mode

Ohmic EDA H-mode plasmas, which are more suitable for investigation with the fast

scanning probes, were produced in the usual manner, by ramping the toroidal field down to induce an L-H transition and then bringing the field back up. Specifically, the goals were: to look for q-resonance effects; to compare the particle (and energy) flux convected by the fluctuations with local particle and power balance; and to verify the extent and radial structure of the quasi-coherent mode.

The motivation for the q-resonance study stems from the tentative identification of the QC mode with the family of resistive ballooning modes, which would be expected to be localized to rational q-surfaces. The experiment varied the programming of the toroidal field ramps to produce a slow scan of the edge q, looking for reproducible changes in the QC mode amplitude, frequency, or wave number. Phase Contrast Interferometry (PCI) was the principal diagnostic for this experiment, while the scanning probe was used to collect detailed fluctuation data. No clear q resonance effect was seen. Changes in mode amplitude and frequency were observed, but the strongest correlation was with sawteeth. There are other excursions which are not correlated with sawteeth and these need to be carefully compared to the q time histories. Data were obtained from the probes, edge Thomson scattering, and from the Lyman alpha array for the transport studies. On one shot, the scanning probes apparently traversed the entire radial extent of the mode,

which measured 1-2 mm, as had been found in earlier experiments.

EDA H-mode plasmas were produced at 800kA and 1MA, with RF power between

2.5 and 3.5MW to investigate the effect of edge current density on ELM behavior in high b_N (1.2<b_N<1.6) C-Mod plasmas. These discharges were in the regime where "grassy" ELM activity begins to be seen on the H-alpha and other diagnostics, along with the QC mode typical of the EDA. The edge current profile was transiently altered by imposing current ramps, raising and lowering the surface loop voltage by about 2V. The goal was to test the theoretical prediction that the consequent increase in edge current density would make the peeling modes more unstable, leading to increased ELM activity, while decreasing J_edge with reverse current ramps would stabilize the modes. This experiment was technically challenging, as it was difficult to maintain good RF coupling during these transients; there were also some perturbations to the equilibrium shapes during the current ramps, which may complicate the analysis. However, the preliminary result is that the observed changes in ELM behavior were opposite to the expectation: increasing the current seemed to suppress pre-existing ELM activity, while reverse ramps resulted in short bursts of ELMs in a previously quiescent discharge. A number of shots have been

identified for detailed analysis.

Density Limit

A half-day run was devoted to study of the density limit. The goal was to collect data on the changes in turbulence and transport which have been linked to the density limit. In the previous experiments it was observed that the region of large coherent fluctuations, which normally exists in the far scrape-off, moves inward to the separatrix and intrudes into the closed field line regions as the density limit is reached. In this experiment, we inserted the probe more deeply into these plasmas at high densities in order to follow

the evolution of the profiles and fluctuations, to assess how far into the main plasma this phenomenon reaches. With the probe inserted about 50 msec before the density-limit disruption, we observed a break in profiles, turbulence and autocorrelation function about 1cm inside the LCFS. The temperature profile was seen to shrink from the plasma edge as the limit was approached even with P_rad << P_input.

Turbulence/Li Pellet Imaging

The ultra-fast-framing camera (see description in Diagnostics section) has been used to follow the evolution of turbulence found in the edge of the C-Mod plasmas. We have found previously that the edge turbulence is "bursty", that is occurs with a higher incidence of large density regions than would by expected from a normal distribution. These bursts have a lifetime of ~10-50 microseconds. With the camera we are able to observe the growth, movement, and disappearance of the emission from the high-density bursts. To date we have done this using a radial view where they appear as toroidal filaments. We find preliminarily that at least some of the filaments move poloidally with a speed of ~200 m/s. We are going to move the view so that we can resolve both poloidal and radial motion. Fig. 1 shows side views of turbulent filaments.

Fig.1

The fast-framing camera has also been used to measure the radial profile of the poloidal magnetic field. This quantity is especially crucial for the "advanced tokamak" research planned for Alcator C-Mod, where control of this profile is needed. The measurement technique employs viewing the ablation light from an injected lithium pellet. As the pellet traverses the plasma radially, the Li⁺ in the rapidly ionizing ablation cloud is constrained to be parallel to the local magnetic field and is "cigar-shaped" when viewed radially. The tilt of the "cigar" thus measures the field pitch, which rotates in response

to the changing poloidal field as the pellet crosses the plasma. The fast-framing camera measures this cigar tilt and its change during the ~250 microseconds it takes for the pellet to traverse the plasma. Fig. 2 shows the changing tilt, where each frame integrates for

20 microseconds.

Fig.2 Images of Li cigar

Neo-classical Tearing Modes

A comparison was made between the collisionality calculated at the q=1.5 and q=2 surfaces in C-Mod with several scalings for Neoclassical Tearing Mode (NTM) thresholds in DIII-D and ASDEX-Upgrade, for a large number of high b shots (1.2<b_N<1.7). In a few of these shots, large low frequency MHD modes were observed that increase in amplitude with increasing beta during the discharge and then finally lock at an amplitude of about 50 G measured at the wall and lead to a beta collapse. Most of the C-Mod data lie at substantially higher collisionality than the DIII-D and ASDEX-Upgrade NTM data, but the two discharges with large MHD modes fell close to each of the DIII-D and ASDEX-Upgrade scalings in terms of dimensionless parameters b_N/r*_i and n*_i. The C-Mod r*_i values for these discharges fell in the range of 0.004 to 0.007 and the collisionalities ranged from 0.01 to about 2.0. While the proximity of the threshold from these other machines to these two shots with large MHD does not prove they are NTMs, it is further evidence suggesting that these higher beta discharges at low

collisionality are at least approaching the NTM threshold.

Disruption Studies

An attempt was made to extend the encouraging results of last year's neutral point experiments by programming the hybrid plasma control system (PCS) to jump to specialized feedback programming at the thermal quench, with the goal of extending the period of post-quench vertical stability. Thermal quench disruptions were induced by firing silver-doped lithium pellets into established discharges. We started off by reconfirming last year's neutral point result that post-thermal-quench plasmas show

enhanced vertical stability when run about 3 cm above the midplane. Then we proceeded to try out two different PCS programming setups, which we hoped would promote vertical stability for even longer periods after the thermal quench. The alternate programming showed some promise; these experiments will be revisited in a future campaign.

Edge/Divertor Physics

Continuum Emission Studies

The purpose of these experiments was to obtain high resolution images of D_g and continuum emission in the divertor at various densities, determine the spatial relation of the D_g and continuum emissivities, and test the hypothesis of molecular recombination as the cause of the "anomalous" continuum emission in the divertor. Plasmas were run

with the same density programming in deuterium, and with strong helium puffing, to vary the possible contribution of molecular deuterium to the observed continuum emission between 420 nm and 430 nm. The continuum signal was observed to decrease by about 25% during the mostly helium shots, while the deuterium lines dropped by approximately a factor of 4 to 10. This modest decrease in the continuum when compared to the significant deuterium decrease leads us to conclude that the continuum is not dominated by molecular recombination.

Secondary Separatrix

This study consisted of a series of shots in which the separation between the separatrix associated with the upper x-point, normally the secondary separatrix, and that associated with the lower x-point, which normally defines the plasma boundary, was varied dynamically during the discharges. In some cases the ramp was extended until the plasma went from the usual lower single null to an upper single null. This was done to determine

whether or not the emission band seen near the inner wall is associated with the inboard-most (secondary) separatrix. All results seem to indicate that the emission follows the inboard-most separatrix, suggesting different radial transport characteristics for the two regions separated by the inboard-most separatrix. A plasma with a secondary separatrix is shown in Fig.3.

Fig.3 Plasma with a secondary separatrix.

RF Research

Installation of the J-Port antenna was completed this quarter after significant modification to the antenna strip line components, front tiles, and back plate. Due to arc damage found after the last campaign, the radial strip lines elements were aligned with magnetic field to prevent E parallel to B. In addition, the electrode spacing was increased from 1 cm to 1.5 cm. S-parameter measurements indicated that the antenna was not significantly modified with respect to electrical characteristics. These measurements confirm our attempt to maintain a 50 Ohm transmission line while eliminating E parallel to B-field arc paths. During the last campaign, a strong RF plasma edge interaction limited the injected power to ~2.5 MW into H-mode and the interaction appeared to follow field lines. The BN tiles were aligned and all metal surfaces except the Faraday screen were covered or removed. Measurement of the antenna tile position on all antennas confirmed that the J-port antenna is at the same location as D and E-port antennas. In order to interrupt long field lines across the antenna, an insulating septum was installed. The back plane feedthrus were also modified to reduce the E field parallel to the B-field. In addition, four optical arc monitors, six B-dot probes, and an MKS pressure gauge were installed. These modifications have contributed to the successful operation and understanding of this antenna.

After commissioning the fault control system and repairing phase shifter #4 (see below), the J-port antenna has been tested to 3.0 MW without significant RF-plasma edge interaction at the antenna corners in H-mode plasmas. It heated well and reached ~26 kV briefly. The 2-2.5 MW is routinely injected and the voltage reaches ~22-24 kV during H-mode which is ~50% higher than during the last campaign. The optical arc monitor has been demonstrated to have signal correlated with arcs during vacuum conditioning. In plasmas operation, most faults do not have a corresponding signal in the optical arc monitor. This suggests that most faults are occurring elsewhere. During a particular experimental day, the antenna had numerous faults when the voltage reached ~20 kV and degraded over a series of discharges. The fast data indicated that some arcs survived for 30-80 msec. Up to 100 J could be available to dissipate in the arc; therefore, we reduced the reflected-to-forward power ratio necessary to generate a trip by 25% for all future experiments. This successfully limited the arcs to ~15 msec or ~15 J per MW injected. We suspect arcing in the antenna strap is responsible for these arcs, because no signal corresponding to the arcs were observed in the optical monitor signals, and the antenna "paper clip" geometry could allow an arc on one antenna strap half without resulting in a phase balance or full reflected power fault in the transmission line. We often observe an interaction at the antenna midplane near the septum and side limiters at about the location of the short. This interaction was more severe when running inner wall limited L-mode discharges, and was weakly dependent on plasma current. A phase scan showed that the nominal [0,pi,0,pi] was the most effective heating phase and had little or no negative edge interaction. An outer gap scan was also completed and suggested a gap of 1-1.5 cm was better than larger outer gaps. Antenna performance was insensitive to toroidal field from 5-5.6T. The new optical monitoring diagnostics of the J-port strip lines detected signal when the D and E-Port antennas were active. Comparisons with the voltage data indicate the induced voltage from D and E-Port antennas are low when the light signal is detected and high when no light is observed. This result suggests that the low induced voltage is sufficient to initiate multipactoring. The induced voltage from D and E-Port antennas decreased as the H/D ratio decreased and absorption improved.

Several equipment failures required repair before significant progress on the J-port antenna plasma operation could begin. FMIT#4 had one of the driver by-pass capacitors arced. The capacitor was replaced and the transmitter was brought back into operation with an additional spark gap installed across the capacitor to prevent a future occurrence. Similar spark gaps will be installed on the other transmitters in the near future. During conditioning, the FMIT#4 phase shifter began breaking down after it reached 25 kV. During the inspection, we found that the push/pull rod and a teflon insulator had become tracked. The push/pull rod material was replaced with G7, which has a lower index of refraction and dissipation factor. We have had very good success with replacement G7 rods in our other stub tuners and phase shifters. The tracked teflon insulator was of more concern because it is the second time an insulator at this location has shown this behavior. We have discussed this with the vendor, but have not identified a strong candidate explanation. The focus has been upon the center conductor and the teflon insulator itself. The center conductor was carefully inspected to ensure proper contact was made when fully assembled. Sharp edges were rounded and mating surfaces were blended (these could have enhanced the local electric field). Additional vent holes were added to ensure proper evacuation and purging of assembly. New teflon insulators were installed, but these too failed. Therefore, we simply removed the teflon insulators from this location. This system has operated without problem since. The D-port antenna was out of service for a time due to arcing in the phase shifter. There were arcs both across the teflon surface and across the gap. Additional arc tracks were found near the push/pull rod (but not on it), suggesting that arcing was occurring across the conductor gap. We had been operating up to 40 kV with 20 psig of N₂, which is close to the design maximum. The arc location was also an area with enhanced E-fields due to the geometry. Thus, the breakdown was probably a result of exceeding the phase shifter operating limits in N₂. Future operation will be with 20 psig SF₆, which will raise the power handling by a factor >10 and result in a safety factor of about 2 in voltage. Once the phase shifter was repaired, the D-port antenna experienced numerous false phase balance faults due to a bad demodulator. Another demodulator was installed and the antenna fault protection behaved properly.

Operations and Diagnostics

Operations

During the first week of operation after the extended vacuum break, the H/D ratio was in excess of 2, and after the second week dropped to between 0.3 and 0.6, The first boronization of the campaign was carried out Monday night-Tuesday morning, June 11-12, with an average deposition of 2050 A. Approximately 3.5 hours of ECDC in helium at 5.0x10^-5 Torr followed boronization. The following run was devoted to boronization recovery and conditioning of the J-port four-strap antenna into plasma. The machine ran well, with no startup or other difficulties which have been associated with some previous post-boronization runs. The H/D ratio was in the 5% range, which is satisfactory for the

hydrogen minority ICRF heating scenario.

Diagnostics

During this quarter we acquired an ultra-fast-framing, visible camera. The use of this camera is part of a recently-funded SBIR Phase II among Princeton Scientific Instruments, PPPL, and C-Mod. Our PPPL collaborator, Stewart Zweben, has been instrumental in this project. The camera is capable of taking 12 frames at a frame-rate of up to 2 million frames per second. We have used it successfully to make two important, but quite different measurements (see above).

Initial data were collected from the in-vessel Penning ionization gauges during

the May 25 startup discharges. These cold cathode ionization gauges, mounted to the vacuum vessel wall at four locations, use the ambient magnetic field inside the tokamak for operation. Three gauges are positioned in the main chamber to record toroidal and poloidal variation in the neutral pressure. One gauge is located in the lower divertor to record fast changes in the divertor neutral pressure when the divertor bypass flaps are opened. The main-chamber gauges appear to be operational. No-plasma calibration shots

indicate that their signal is proportional to the torus neutral pressure. Initial data during plasma operation indicate weak toroidal variation in the neutral pressure (less than a factor of 2) and strong poloidal variation (up to a factor of 5). In order to avoid a saturated signal, the divertor gauge must be operated in a reduced bias mode where a

non-linear pressure response might be expected. Nevertheless, this gauge appears to provide information on the rate of change of pressures in the divertor when the bypass flapper is opened. Initial results suggest that the transient neutral exhaust rate through the C-divertor flapper is small (~1 torr-liter/s). Further experiments are planned.

A half-day was dedicated to a cross-calibration of the Thomson scattering

density response based on cut-off of the ECE signal in high density, reduced

field discharges. Corrected calibration factors were obtained, and are being

applied.

Diagnostic Neutral Beam

The DNB is operating at full voltage and current, that is at voltages near 49 kV and currents slightly above 5 A. Two bakes of the DNB have significantly improved the base pressure but did not completely remove the water from the beam. We have not yet repeated the best component mix that we observed in February, but we are continuing to

move in that direction and we expect to repeat those results with additional arc conditioning.

We are well into the reionization studies which will help us understand the operation of a DNB on a machine with a high midplane pressure. Beam into gas measurements have been made over a pressure range that is consistent with that produced in C-Mod. Of course, we have done these measurements without a magnetic field so that the

reionized component cannot contribute to the runaway reionization that previously resulted in duct damage. After some initial experiments in He to observed CXRS spectra, we have turned once again to B after C-Mod was boronized. In helium, the background from the edge appears to obscure the CXRS emission. There were some indications of CXRS spectra, and we may return to He after completing the current work in boron. For B, we have some promising results with regard to observation of the n= 6 to 5 transition at approximately 2980 A. A line appears at the right time and in approximately the right spectral region. We will continue observations with improved temporal resolution and improved spectral calibration to verify this observation.

Repairs to the MSE optics in Feb. were successful, and the system is operating as expected. With marginal improvements to the full energy component of the DNB from the last run campaign, MSE has made some measurements of pitch angles during plasmas. However, a gas into torus calibration still remains to be done to convert measured anglesto pitch angles. Still suffering from low signal to noise due to low full energy components in the beam, current MSE pitch angle measurements still have errors too large to be used in deriving an accurate q-profile.

The BES diagnostic is making fluctuation measurements. Following recalibration of the filters and the rotation control mechanism, the BES system was used for some beam width measurements. These will continue. The BES system was tested by attempting measurement of ambient D-alpha emission and of bremsstrahlung. The former appears to

resolve plasma flucutations. The latter are at the noise level of the system as expected for that type of measurement. Fluctuation measurements with the beam are imminent.

Lower Hybrid Project

(see separate report)

Collaborations/Participation in the Fusion Science Community

ASDEX Upgrade

Time has been devoted during two runs for ASDEX shape development. This experiment will produce equivalently shaped plasmas on C-Mod as on ASDEX Upgrade and attempt to match dimensionless parameters n*, r* and b near the edge of the plasma (the H-mode pedestal region), both at the L-H transition and in the fully developed H-mode. We can now generate plasmas needed for the ASDEX/C-Mod scaling studies.

National Fusion Collaboratory

Alcator personnel will be involved with a SciDAC funded "National Fusion Collaboratory". The goal of the collaboratory is to advance fusion sciences research through more effective integration of experiments, theory and modeling. This integration will be enabled via the adoption of a common set of distributed computing and visualization tools. MIT's role in the collaboratory will primarily involve implementation of modern certificate based authentication and authorization mechanisms into the MDSplus and SQL server data systems. MIT will also help develop distributed computing applications based on these data management systems and the GLOBUS

toolkit.

Domestic Travel

Ron Parker and Jim Irby visited PPPL to serve on the FIRE Pre-Conceptual

Design Review Committee from July 5-7. Very good progress has been made

on magnet, divertor, PFCs, vessel, and fueling systems.

Howard Yuh attended the PPPS-ICOPS meeting in Las Vegas from June 20-22 to present an overview oral on the C-Mod EDA mode focusing on the modeling efforts and progress made with the new C-Mod Beowulf parallel Linux computer cluster.

International Travel

J. Snipes, I. Hutchinson, D. Mossessian, C. S. Pitcher, and J. Rice attended the EPS meeting in Madeira, Portugal from 18 - 22 June 2001. Four papers on C-Mod results were presented.

Near Term Plans

The immediate goal of this run period is to evaluate the performance of the J-port antenna, and the DNB. Among the scientific goals for this run period are: H-mode operation with ICRF power in excess of 5 MW; further documentation/understanding of the QC mode and double transport barrier plasmas.

The J-port antenna is now at 70 MHz to allow for 2-frequency heating ITB experiments.

Several unanswered questions regarding the double transport barrier plasmas will be addressed: Can the ITB be sustained in steady state? Why is the ITB foot location near r/a = 0.5, regardless of ICRF resonance location, or whether it is formed in Ohmic plasmas or by pellet injection? Why is there an abrupt threshold (in lowering B_T) for the ITB formation by ICRF heating? Does the ITB appear only with the ICRF resonance on the high field side, or is there a power threshold? How is impurity transport affected by the ITB?

Schedule

Plasma operations will continue until the end of July. After this time, there will be an extended maintenance period. During this interval the main activities will be a full inspection of the TF magnet and an upgrade of the inner wall/divertor. This upgrade will allow for operation with I_p above 2 MA and with increased flexibility in plasma shaping.

Modifications in preparation for the Lower Hybrid experiments will also be performed at this time.