From: WOLFE@PSFC.MIT.EDU
Reply-To: WOLFE@PSFC.MIT.EDU
Subject: ALcator C-Mod Weekly Highlights
Newsgroups: sci.physics.plasma
Organization: MIT


                  Alcator C-Mod Weekly Highlihgts
                        May 29, 2000


Operations
----------

Plasma operations continued at Alcator C-Mod last week with four run days
scheduled and completed. A total of 69 plasma shots were produced, with a
startup reliability of 80%. Experiments in support of four MiniProposals were
carried out; we also produced a series of shots aimed at comparing ICRF-heated
H-mode quality under present conditions with those produced in previous
campaigns.

Plasma physics operations will continue this week. A fresh boronization is
scheduled for Tuesday night.

Physics and Analysis
--------------------

This week, experiments were carried out to continue the investigations into
the scaling of the EDA/ELMfree boundary with the ultimate goal of
understanding the physics of EDA. It has previously been determined that the
EDA phase usually occurs within a specific range of the safety factor q95. In
this run, we varied plasma elongation at fixed q95.  There were no obvious
effects of elongation on the EDA/ELMfree boundary.  The EDA or ELMfree
character did not change markedly throughout the scans.  There were
differences due to impurity injections, RF power changes, radiated power
build-up, etc, but nothing obviously systematic with kappa. The scan covered a
limited range of elongation, but if the dependence were as strong as that for
q, the effect would have been apparent.

Part of one run day was devoted to a mini-proposal (MP#269) to test the
theoretical prediction of Perkins and White that ICRF-induced plasma rotation
in C-Mod should change direction from co- to counter-current as the ICRF
resonance is moved from the low to high field side of the tokamak. The results
so far are interesting but not conclusive.  The toroidal magnetic field was
first scanned down from 5.3T to 4.5T and then back up to 6.0T. The ICRF power
was kept constant at approximately 2 MW. H-mode plasmas were obtained at all
fields between 4.5 and 5.3T. Baseline rotation measurements at 5.3 T (ICRF
resonance at 0.48 cm to the LFS) indicated V_tor = 50 km /s (from High
Resolution X-ray Spectroscopy) with EDA H-modes and stored energies of about
95 kJ. The plasma rotation was still observed to be in the co-current
direction at 4.9 T (ICRF resonance at 4.6 cm to the HFS).  The mean V_tor was
slightly lower than at 5.3 T, although the stored energy was also lower (80
kJ). At 4.5 T the rotation velocity indeed reversed sign to about -20 km/s.
However, these plasmas were quite different from the plasmas at 4.9T and 5.3T
because the ICRF heating was far off-axis (9.4 cm to the HFS). Evidence of a
hollow T_e profile was seen and sawteeth were suppressed during the H-mode
phase. Evidence of an internal transport barrier (ITB) was also seen in that
the central density was rising during the H-mode phase to about 4 X
10^20. Difficulty coupling RF to the high-field plasma prevented us from
getting comparable results when the toroidal field was increased to 6.0 T,
moving the ICRF resonance 9.4 cm to the LFS.  These experiments will be
continued.

One run day was devoted to the start of a systematic study of the scaling of
H-mode pedestal parameters (MP#271). RF power was varied in a series of discrete
steps for target plasmas at different currents. Data were obtained at 800kA and
600kA. An attempt to go to high current (1.4MA, corresponding to q~2.7)
resulted in a disruption, which unfortunately caused sufficient degradation of
machine conditions that further experiments were not possible on this day;
overnight discharge cleaning was required before good plasmas were again
obtained. These pedestal scaling experiments will be continued.

Last Friday's run was devoted to a test of predictions by Y. Nakamura (JAERI)
concerning the existence and location of a "neutral point" which exhibits
enhanced stability for vertical displacement events after a thermal quench. On
JT-60U, and other tokamaks, it was found that when the plasma is operated with
the magnetic axis height at this neutral point, the initial vertical drift
after a thermal quench is significantly slower than usual, sometimes allowing
vertical control to be re-established, or for ameliorative action to be
taken. In comparison with JT60-U, C-Mod typically operates at higher
elongations, closer to the ideal axisymmetric stability limit, and therefore
provides a more demanding test of this effect.  We ran a series of standard
plasmas at a range of different heights (keeping the elongation nearly
constant by moving the x-point along with the center), and disrupting them
using double-barrel silver-doped lithium pellet injection.  For plasmas with Z
< +3 cm (measured with the core x-ray arrays), the thermal quench led to a
prompt loss of vertical stability, with the plasma moving downwards.  For
plasmas with Z > +3.5 cm, the thermal quench was followed by a prompt upward
movement.  However, with Z centroid set for +3.0 and +3.5 cm, we observed
clear signs of enhanced vertical stability.  In one case the plasma remained
near the midplane for 8-9 ms before finally heading downward. The last shot of
the day exhibited vertical stability for more than 20 ms after the pellet
injection, but for this case we need to confirm whether the pellets did indeed
cause a full thermal quench, not just a minor disruption.

Peaked density profiles, similar to those seen during last year's campaign,
were observed again in the Ohmic H-modes of the May 18 run.  Peaking factors
(n(0)/n_ped) > 3 were found at the end of the H-mode.  The peaking begins
early in the H-mode and increases throughout.  The lower plasma current
discharges had less peaking than the higher current ones, and the highest
peaking was observed in the discharge that returned to 5 T rather than
remaining at lower field.  These results indicate that the peaking is affected
by either the absolute field and current or their rates of ramping, and that
higher field and current increase the peaking at least up to the values
attempted so far.  It is not yet clear whether there is an optimum edge q
value or optimum edge dq/dt rate.  A miniproposal to further study this
spontaneous peaking phenomenon in H-mode (MP254) has been approved and is
awaiting scheduling

ICRF Systems
---------------

Both D and E-port antennas were operated in support of physics experiments
this past week.  Into EDA H-modes, 2.8 MW was coupled to the plasma, and we
are investigating the absorption efficiency.  If we assume 90% absorption
efficiency, the H-factors were up to ~1.8, consistent with the C-Mod database,
so the efficiency is expected to be reasonably high.  We are still
experiencing faulting during plasma operation that we are continuing to
investigate.  The faulting may be a result of having high voltage points in
new locations compared to previous campaigns, e.g. gas barriers at high
voltage locations.  We may need to make further adjustments to the resonant
loops and power feeds, which are located external to the vacuum vessel, to
improve reliability.  In discussions with R. Wilson (PPPL), we have resolved
inconsistencies in the reported antenna voltages and resistances.  We are
beginning to cross check the various power and voltage measurements to
complete the diagnostic set and improve our understanding of the performance
of the antenna systems.

Randy Wilson and Joel Hosea visited MIT to participate in physics experiments
utilizing D and E-port antennas and to discuss plans to modify the J-port
antenna.  Before modifying the J-port antenna, we agreed that we will
investigate its ability to operate at high target densities and to reconfigure
the external plumbing to have straps 2 and 3 powered by one transmitter and
straps 1 and 4 to be shorted at what is normally their high voltage
point. During the last campaign, J-port would not operate at high target
densities and modifications were made to the antenna during the winter in an
attempt to address this problem.  Reconfiguring the external loops will allow
us to test the idea that parasitic loading may be resulting in lower heating
efficiency compared to D-port.  The modifications are aimed at improving the
shielding for the radial feeds.  We propose to add shields to protect the
strap leads and to prevent their fields from interacting with the plasma.  We
would also replace the side tile mounting plate with a solid one that also
covers the radial portion of the current strap, and mount BN tiles instead of
Mo.  The top and bottom protection tiles will be replaced with BN tiles as
well.

DNB Systems
-----------

The new booster transformer for the filaments was installed and tested.  Arc
current scans versus filament voltage were completed.  The suppressor and
gradient grid voltage scans were continued for 2 days until a suppressor
supply problem prevented further operation of the beam last week. Tests will
continue in this week.

Travel and Visitors
-------------------

Ian Hutchinson attended the FFCC meeting at PPPL on Friday.

Brian LaBombard, Spencer Pitcher, and Bruce Lipschultz, Rejean Boivin, Daren
Stottler (PPPL) and Sanjay Gangadhara presented papers at the Plasma Surface
Interactions Conference in Rosenheim, Germany. Posters on C-Mod research
authored by Chris Boswell and Robert Nachtreib were also presented.

As noted above, Joel Hosea and Randy Wilson visited from Princeton.

Y. Nakamura arrived from Japan for a two week visit to continue the
JAERI/MIT collaboration on disruption research.  He served as
co-session-leader for the neutral point experiment described above.