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 Highlights
Sep 5, 2000

Operations
----------

Plasma operations continued at Alcator C-Mod last week. Four run days were
scheduled and completed. The first boronization since the July vent was
carried out on Monday night. Approx. 140 psi of diborane gas was used
corresponding to about 2000 Angstroms of boron on the walls. The boronization
was followed up with 2.5 hours of ECDC in helium at 5.e-5 Torr.

C-Mod operation was run on the usual Tuesday through Friday schedule. A total
of 94 plasma shots were produced, with a startup reliability of 90%. Two run
days were devoted to experiments by the pedestal/H-mode physics research
group, and two to ICRF conditioning and development.

Plasma operation will continue this week.


Physics
--------

With its good energy confinement, moderate particle confinement and lack of
large ELMs, the EDA H-mode could be a highly favorable regime for a tokamak
reactor - if we can understand the underlying physics sufficiently to
extrapolate into new regimes. The key to understanding EDA seems to be in the
quasi-coherent fluctuations which have been observed by the reflectometer,
PCI, and probes. These fluctuations seem always to accompany EDA and have
been shown to drive significant particle flux. An experiment carried out this
week employed the A-side fast scanning probe with a magnetic coil embedded in
the probe head along with two Langmuir probes, and the fast framing tangential
camera, which was set up with an Halpha filter and a hydrogen puff to observe
the quasi-coherent (Q-C) fluctuations in ohmic H-mode EDA plasmas. Ohmic EDA's
are chosen for these experiments to limit the incident heat flux on the
scanning probe, permitting deeper penetration. A principal goal was to look
for magnetic components of the quasi-coherent mode with the scanning probe.

Q-C fluctuations were clearly seen on the reflectometer, electrostatic probes
and on the magnetic probes, with frequency between 100 and 150kHz. Although
the magnetic sensor itself never got closer than 1 cm to the separatrix, the
amplitude of the signal reached 400 T/sec peak to peak. Accounting for the
mode frequency, this corresponds to about .4mT or about 20 A/cm^2, which is of
the same order as the total calculated edge current density. The
electrostatic probes, which are part of the same probe head, saw the modes
quite clearly - with similar time histories. The main distinction is that
they only see the mode when they are immersed in it (deep penetration) while
the magnetic probe needs only to be close to the separatrix. The conclusion is
that the Q-C mode has a very significant magnetic component.

On one shot the probe also observed some ELM precursors from type III ELMs.
These had a longer decay length as seen on the probe scan, and were also seen
with the wall magnetics, suggesting that they have a lower m number. These
fluctuations were not seen on the floating potential, strengthening this
conclusion. Thus we would conclude that the Q-C mode is distinct from ELM
precursors.

We also obtained D-alpha images with the gated tangentially-viewing camera for
most shots. Clear differences among L-mode, ELMfree H-mode, and EDA were
seen. Emission from L-modes is quite broad, extending 5-6 cm outside the
separatrix. In ELMfree H-mode, the region of emission is much narrower and
hugs the separatrix. During EDA, clear fluctuations are seen well outside the
separatrix. These images suggest additional transport in the far scrape-off
(consistent with profile measurements). They do not appear to be a visualization
of the Q-C fluctuation itself.

Thursday's run was devoted to documentation of pedestal parameters in
ELM-free, i.e. non-EDA, H-modes with ICRF heating. Data were obtained at 1.0 and
1.2 MA, at a toroidal field of 5.4 tesla. RF power was <2MW, and the H/D ratio
increased to >10% over the course of the day, resulting in less than optimum
heating efficiency. On Friday's run it was found that between-shot ECDC in D2
resulted in stable H/D fractions of ~5%, which is preferred for efficient
minority heating.

A set of eight BP-loops mounted on the outboard limiter located between G- and
H-ports were incorporated into the EFIT flux reconstructions, in an effort to
improve the reliability of the inferred location of the last closed flux
surface. Discrepancies between the EFIT reconstructions and other diagnostics
of order 5mm, particularly at the outboard midplane have been previously
noted. Inclusion of the limiter BP-loops, which are much closer to the plasma
than the original magnetics diagnostics mounted on the outboard wall, did
result in reconstructions which placed the separatrix further in by about 5mm,
in better agreement with independent diagnostics. However, the results, and
subsequent testing, indicated a discrepancy between the new and old loops at
the level of ~2%; this discrepancy, which is well outside the nominal accuracy
of 0.5%, was later determined to be due to a systematic difference in the
calibration techniques used for the two coilsets. An independent check of the
absolute calibrations is now in preparation.

Dynamic outer gap scans carried out during the previous week give some
independent indication of the correct location of the last closed flux surface
by comparing magnetic fluctuations with the EFIT calculations of the outer
gap. The amplitude of the outboard midplane magnetic fluctuations increases
as the gap is reduced, then remains constant. We speculate that the
fluctuations remain constant when the gap is in fact zero and the mode cannot
get any closer to the coil. The time history of the RMS of the magnetic
fluctuations was compared with the EFIT calculations using either just the
original poloidal field pick-up coil set alone or including the integrated
outboard limiter coils, as discussed above. The comparison indicates that the
fluctuation amplitude is constant when the outer gap is zero for the EFIT
calculation including the outboard limiter coils, suggesting that these
reconstructions give a better estimate of the outer gap. However, questions
about the coil calibrations must be resolved before a definitive conclusions
can be drawn.


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

ICRF antenna conditioning resumed this week following overnight boronization
on Monday. Hydrogen/deuterium ratios in the plasma were now reduced to the 5%
level, allowing ICRF hydrogen minority heating experiments to be initiated on
Wednesday. Inner-wall limited discharges were run in order to avoid L-H mode
transitions, which complicate antenna matching and power balance analysis. D-
and E-Port antennas conditioned up to 1.35 MW each relatively easily. The
J-Port antenna began to fault above 0.8 MW, with accompanying impurity
injections. A series of low power ramps at different plasma outer gap
settings produced loading resistance measurements from the J-Port antenna that
indicate similar values and power behavior to those obtained before the July
modification, and different from the other two antennas. A further series of
discharges with different outer gap settings and a constant power level of 0.8
MW from each antenna in sequence allowed a heating efficiency comparison of
the antennas to be made. The J-Port heating efficiency, as indicated by
central electron temperature increase, stored energy increase, and neutron
rate, was lower than the other two antennas (roughly half). A phase change of
the J-Port antenna from (0,pi,0,pi) to (0,pi,pi,0) resulted in severe faulting
with ejections from many different points on the antenna.

Further conditioning and a repeat of the antenna comparisons were carried out
on Friday. The J-port antenna conditioned better compared to the Wednesday
run, in the sense that the number of faults were greatly reduced. However,
there was still quite a bit of glowing on the antenna and many impurity
injections. Power ramps were used to try and get around this condition and
these were successful in getting to higher launched powers (up to 1.2MW) but
resulted in disruptions due to the injections. During this conditioning at a
plasma current of 1 MA the disruptions started to occur at lower and lower
powers. Eventually it was necessary to lower the power to 400kW (200kW each
pair) to get the antenna to run cleanly. The current was then changed to 0.8MA
and conditioning was repeated. The total achievable power was raised to ~1.2MW
during this set of conditioning shots just as at 1MA.

On Friday, performance of the J-port antenna was significantly better than
that on Wednesday. The central temperature increase, stored energy increase,
and the neutron rate were only slightly lower (~15%) for the J-port antenna
compared to the other two antennas.


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

The Diagnostic Neutral Beam was brought up during the last week. The modreg
problems that caused DNB operation to cease a few days before the end of C-Mod
operation in June were traced to relatively minor problems with power for the
control system and with fuses for the main beamline power. An intermittent
problem early in the week with one of the beam cryo pumps was associated with
high ambient cell temperature. This did not limit operation and has not
recurred. Apparent modreg cooling problems were traced to a faulty interlock
and resolved. The beam plasma source was conditioned early in the week. The
accelerator was then vacuum conditioned to approximately 47 kV. In the course
of the run on Friday, the beam was operated into the plasma.

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

Stew Zweben (PPPL) came to C-Mod 8/31-9/1 to continue to work with the C-Mod
physicists and computer engineers on optimizing and automating data
acquisition for the new 2-D edge turbulence imaging diagnostic.