Polychromator S Scott PPPL R Mumgaard CFS J Ko NFRI Paper CP1100101 60 th APSDPP Portland OR Nov 59 Work supported by US DoE contracts DESC0016409 and DEAC0209CH11466 and grant DESC0016614 ID: 935877
Download Presentation The PPT/PDF document "2018 beam-into-gas calibration of the KS..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
2018 beam-into-gas calibration of the KSTAR MSE Background Polychromator*
S. Scott (PPPL), R. Mumgaard (CFS), J. Ko (NFRI)Paper CP11.0010160th APS-DPPPortland, OR Nov. 5-9
* Work
supported by US DoE contracts DE-SC0016409 and DE-AC02-09CH11466 and grant DE-SC0016614
.
Slide2The MSE background polychromator first constructed for use on
Alcator C-Mod has been moved to KSTAR, and the system was recently expanded to include all 25 sightlines.A recent beam-into-gas calibration of the polychromator revealed several anomalies, including an extreme sensitivity of the measured angle to the placement of the filter passband to the MSE spectral lines.These anomalies are now understood to be caused by secondary beam emission.Calculations of the secondary beam emission effect reproduce semi-quantitatively the sensitively to filter placement and the observation of polarized light several nm away from the MSE spectrum.
Curiously, measurements and both numerical and analytic calculations show that secondary beam emission has a minimal effect on measured polarization angles when the magnetic field pitch-angle is zero.
The secondary beam emission effect can be made negligibly small by operating at factor ~100 lower torus pressure, but this is not viable on KSTAR due to concerns of wall heating by the beam.A standard approach to dealing with secondary beam emission is to perform the b2g calibration at multiple pressures, and extrapolate back to zero pressure. This assumes that the effect is linear in pressure, which may not necessarily be satisfied in KSTAR due to attenuation of the secondary beam ions by charge exchange.
Overview
KSTAR_MSE_MEMO_015r.pdf Page / 36
2
Slide3ObjectivesDetermine optimum placement of filter passband relative to MSE spectrum (
JJ)Obtain accurate angular calibration (~0.2o) for all BT, all polarization angles.
30 beam-into-gas shots taken on Sept. 3, 2018
Each shot: five 100 ms beam pulses. Takes ~40 ms for 1st beam to reach equilibrium voltage. Use 40-90 ms
period for data analysis.Excellent beam performance. Lost only 1 of 150 beam pulses.
TF constant in a shot. 5 values of PF 5 known pitch angles.
Acquired data at B
T = 1.2, 1.9, 2.5, and 2.7 Tesla. Got extra data at 2.7 T with alternate set of pitch angles.At each BT, changed filter temperature on successive shots to move filter passband relative to MSE spectrum.‘Feature’: torus pressure varies factor ~2 among the five beam pulses, but was reproducible shot-to-shot.This was exploited on three shots (1.2, 1.9., 2.5 T) with same PF on all beam pulses to quantify effect of torus pressure on the calibration.This data was crucial to understand large anomalies in the data at low BT.
Beam-into-gas calibration procedure
KSTAR_MSE_MEMO_015r.pdf Page / 36
3
Slide4MSE 101KSTAR_MSE_MEMO_015r.pdf Page / 36
4
An electric field splits the H
a line into a multiplet
.
In a tokamak, the electric field is the Lorentz field due to a beam neutral’s motion across the magnetic field (E = V x B). The motion also provides a Doppler shift.
The
multiplet includes p sigma lines which are linearly polarized perpendicular to E, and three p lines to the red of the s lines that are polarized parallel to E.We use narrow passband filters to preferentially pass p light at the expense of s light.If the filter passband is positioned to allow some s
light to sneak in:
The polarized signal intensity is reduced.
The polarized angle can also change – due to 3D collection effects + optical aberrations
Dl
Doppler
Slide5The MSE spectrum is compressed at low BT
KSTAR_MSE_MEMO_015r.pdf Page / 36 5
Slide6Probably OK at 2.5 TKSTAR_MSE_MEMO_015r.pdf Page / 36
6
Slide7At sufficiently large filter offset, the filters are passing mostly secondary emission
KSTAR_MSE_MEMO_015r.pdf Page / 36 7
The range of wavelengths of secondary emission that passes thru
the filters, i.e. the width of the beige regions in the cartoon,
is a strong function of sightline and beam geometry.
Slide8KSTAR MSE optics experience strong aberrations (factor ~30 bigger than on C-Mod)
KSTAR_MSE_MEMO_015r.pdf Page / 36 8
As on C-Mod, the
invessel calibration data is well-fit by a periodic function of angle:
qmea = c
o + q +
c
2 cos(2q+f2) + c4 cos(4q + 4f) + c8 cos(8q + f
8)
The coefficients c2
, c
4
… are caused by optical aberrations in the optics. c
2
is caused by
diattenuation in mirrors.
In principle, this behavior is entirely ‘captured’ by the invessel calibration … that’s why we do it.
But, unlike
invessel
calibration, in plasmas MSE will see both
p
and some (orthogonal)
s
light.
s
light experiences a different correction due to cos(2
q
) term.
The measured angle will then be a function of the ratio of
s
to
p
light intensity: influenced by both atomic physics and filter
cwl
location.
Slide9This was one motivation for exploring operation at large filter offset: it decreases the ratio of ‘polluting’
s light to the desired p light.But even in absence of optical aberrations, ‘3D collection effects’ can cause the measured angle to be a function of the filter offset (Geelen[2013], Mumgaard[2015]).Unfortunately, the data at large filter offset are confounded by a second effect: secondary emission from beam neutrals. This led to worries about diagnostic/data integrity & a rather wild goose chase.
KSTAR MSE optics experience strong aberrations (factor ~30 bigger than on C-Mod)
KSTAR_MSE_MEMO_015r.pdf Page / 36
9
Slide10Measured polarization angle is
very sensitive to filter offset at low TF. This effect is stronger in the MSE-BP system than in the original KSTAR-MSE diagnostic .Linear polarization fraction (LPF) generally decreases with increasing filter offset.Strength of polarized signal and LPF vary significantly with polarization angle. Correspondingly, the statistical error bars vary significantly with polarization angle also.
Measured angle changes by ~1 degree as torus pressure changes factor 2.
Polarized light intensity and polarization fraction decrease with increasing pressure.
The measured delta (= measured polarization angle – true polarization angle) a non-monotonic function of true polarization angle.
Optimum
filter offsets are pretty small, probably in the range 0.1 – 0.25
nm.Puzzles and challengesKSTAR_MSE_MEMO_015r.pdf Page / 3610
Slide11The linear polarization fraction increases with BT
.Generally, the behavior of the diagnostic + calibration improves with increasing BT.Trends as a function of offset are pretty smooth -- for a given sightline, B
T
, and PF. So it is possible to interpolate the existing data onto a set of common offsets for all sightlines. The smallness of the optimum filter offset means that our existing ensemble of filters is well matched to our needs.
When the deltas are interpolated onto a common offset for all sightlines & plotted versus polarization angle, the data look reasonably OK at 1.9, 2.5, and 2.7 T.
Good news & expected trends that
were
observedKSTAR_MSE_MEMO_015r.pdf Page / 3611
Slide12Measured polarization angle is very sensitive to filter cwl at low B
T (1.2 T)KSTAR_MSE_MEMO_015r.pdf Page / 36 12
‘Spectral sensitivity’ =
d
q
mea
/d
l up to 80o /nm.The spectral sensitivity is also very sensitive to actual polarization angle – note change in sign as qpol increases from 76o to 104o.The spectral sensitivity also varies with sightline - much smaller at edge sightlines.
This behavior suggests that gyro-orbit effects (Yuh[2008],
or 3D collection, realistic-beam, & spectral effects (Geelen
[2013], Mumgaard[2015]) are very strong.
MSE spectrum compressed at 1.2T:
l
(
s
0
) –
l
(
p
3
) = 0.17 nm.
cf
filter FWHM ~ 0.65 nm and aperture broadening ~0.23 nm (
ko
[2015]).
Surprising (to me): spectral sensitivity is smallest at small filter offset, where we might have expected to have highest pollution of
p
light by
s
light.
Slide13Similar trends are observed by the original KSTAR MSE system
KSTAR_MSE_MEMO_015r.pdf Page / 36 13
Both systems experience strong spectral sensitivity
Spectral sensitivity is strongest at low BT and almost negligible at 2.5 T
Spectral sensitivity has strong dependence on polarization angle.
Spectral sensitivity also has strong dependence on sightline
Generally, the KSTAR MSE system experiences smaller spectral sensitivity than MSE background
polychromator.System differences:Original KSTAR MSE: FWHM = 0.45 nm and tilt tuning (distorts filter passband).MSE background polychromator: FWHM = 0.45 nm and temperature tuning
Slide14The spectral sensitivity decreases rapidly with BT
KSTAR_MSE_MEMO_015r.pdf Page / 36 14
Separation between MSE
p
line and polluting s line increases linearly with B
T.
Prelude of things to come: in many ways, calibration data is ‘better behaved’ at higher B
T.We plan future operation at filter offset of 0.15 – 0.25, where spectral sensitivity is much smaller.
Slide15As expected, the polarized light intensity decreases with filter offset
KSTAR_MSE_MEMO_015r.pdf Page / 36 15
With increasing offset, the filter passband is moving away from the MSE
p
2, p3, and
p4 lines.
Less overlap with MSE lines
less measured intensity.
Slide16Surprisingly, the linear polarization fraction also often decreases with filter offset
KSTAR_MSE_MEMO_015r.pdf Page / 36 16
Would have expected the LPF to increase secularly with offset, as the filter passband moves further from the MSE
s line.
Trend not true for all polarization angles … LPF actually increases with filter offset for some sightlines & polarization angles.
Detailed modeling of MSE spectrum, 3D collection effects, etc. by MSESIM is required to sort out this physics.
Slide17Measured polarization angles are a function of torus pressure
KSTAR_MSE_MEMO_015r.pdf Page / 36 17
This behavior is expected from secondary beam emission
.
Limited measurements: 5 beam pulses in single shots all with PF=0 & factor ~2 variation in torus pressure.
Can extrapolate to P=0 for ‘official’ calibration, if the secondary beam emission effect is linear in pressure..
Next calibration: get data at 2+ pressures for all BT and all PF.
Modeling with e.g. MSESIM will be performed to examine 3D effects when both s and p light passes thru filters.
Slide18Large (10’s of degrees) anomalies observed in C-Mod MSE calibrations with perpendicular beam injection.
Deviations from true polarization angle scale linearly with torus pressure.Deviations reduced, but not eliminated, by tilting beam 7o off perpendicular.
Numerical simulations (
Yuh[2008]) show that secondary beam emission is a problem even for tangential injection on NSTX.On C-Mod, effect was strongly dependent on viewing sightline and was observed to vary with filter offset as well.
Effect would be most apparent at low B
T because at large filter offset (
cwl
Dl) we are effectively looking ONLY at emission from secondary beam neutrals – little or none from primary emission.Effect would be stronger for the MSE-BP system than for the original KSTAR-MSE system due to larger filter passband (0.65 vs 0.45 nm).Anomalous behavior at large Dl, low BT is caused by secondary beam emission KSTAR_MSE_MEMO_015r.pdf Page / 36
18
Slide19Physics of secondary beam emission
KSTAR_MSE_MEMO_015r.pdf Page / 36 19
A beam neutral ionizes off-
midplane.
Gyrates about field line as it follows field line to
midplane.
Charge-exchanges at
midplane, emitting an MSE photon.But its polarization direction E = V x B differs from primary emission due to gyro motion.
Slide20Elevation view of mapping ion-birth positions to MSE sightline
KSTAR_MSE_MEMO_015r.pdf Page / 36 20
Slide21Consideration of secondary beam emission motivated by large anomalies in MSE b2g calibrations on C-Mod
KSTAR_MSE_MEMO_015r.pdf Page / 36 21
Measured angle deviates from true polarization angle by 10’s of degrees with perpendicular injection.
Strong dependence on both polarization angle and sightline.
Tilting the beam 7
o
reduced the anomaly to a few degrees, but it was still present.
Slide22KSTAR_MSE_MEMO_015r.pdf Page / 36
22
Numerical simulation reproduces measured effect of secondary neutrals on C-Mod
Yuh[2008] surprisingly, secondary emission confounds MSE calibration even with tangential injection.
To avoid effect entirely on NSTX – with tangential injection –Yuh concluded that torus pressure must be below 0.05
mTorr.
This may not be feasible on KSTAR – several
mTorr required to avoid excessive wall heating. Also possible photon-statistics limitations at low pressure.Will probably have to measure at high pressures and extrapolate to p=0.Calculations had to assume non-statistical state populations to match angular data.
Slide23Unmistakable evidence for existence of secondary emission: polarized light in ‘blue’ filters
KSTAR_MSE_MEMO_015r.pdf Page / 36 23
Slide24Calculations of secondary beam emission predict the observed polarized light at 651.5 nm
KSTAR_MSE_MEMO_015r.pdf Page / 36 24
Emission from gyro angles near zero has appropriate Doppler shift to pass thru blue filter with
cwl
= 651-652.
No emission is predicted for red filter at 665.5 nm and none is observed.
This is ‘smoking gun’ evidence for presence of secondary emission in the measured signals.
Slide25Only a fraction of the secondary beam emission gets thru the filter
KSTAR_MSE_MEMO_015r.pdf Page / 36 25
The nominal filter
cwl is chosen to pass the MSE p lines.
Due to secondary emission of ‘gyro-motioned’ neutrals, a range of angles other than polarization angle of primary neutrals passes thru the filter passband. What we measure is an average.
Other gyro-angles are rejected, because their Doppler shift puts them outside the filter passband.
As we shift the filter passband further and further to the red of the
p lines, we get less and less p light.But: this simply moves the filter passband to accept emission from other gyro-angles whose angle is even further from the polarization angle of the primary neutrals.Summary: with increasing filter offset: less and less p light + more-or-less constant light from secondary emission at weird angles.
Slide26A range of secondary gyro-angles has a Doppler shift that will pass thru the filter
KSTAR_MSE_MEMO_015r.pdf Page / 36 26
Each birth-position & gyro angle yields a unique polarization angle due to
E = V
x B.
Sum the polarization angles over all gyro-angles that yield a Doppler shift that will pass thru the filter.
This yields a mean secondary polarization angle for each ion birth position.
Slide27Calculated polarization angle of secondary beam emission
KSTAR_MSE_MEMO_015r.pdf Page / 36 27
Polarization angle from secondary emission differs ~20
o from primary beam emission.
This explains the large ‘spectral sensitivity’: as we increase the filter offset, we get less and less
p
light, but secondary emission intensity is roughly constant.
Measured polarization angle is an intensity-weighted average of the primary and secondary polarization angles.
Slide28Effect of secondary emission on calibration was linear in pressure on C-Mod
KSTAR_MSE_MEMO_015r.pdf Page / 36 28
C-Mod: ramp torus pressure in b2g shots with constant TF & PF.
Change in measured angle is linear in pressure (Ko[2007,2009]).
An analytic model of the secondary emission shows that expected change in angle is also linear in pressure (Scott 2007).
Slide29KSTAR is in the nonlinear regime of secondary beam emission
KSTAR_MSE_MEMO_015r.pdf Page / 36 29
The drift distance of some beam ions between their birth & drift across
midplane exceeds the mean-free-path against charge exchange.
This means that different birth positions will experience different amounts of attenuation between birth & crossing the
midplane.
This means that
different birth positions will contribute differently to secondary intensity and this effect will vary with gas pressure.Expected ratio of secondary- to primary beam emission may be a nonlinear function of gas pressure.This may compromise a linear extrapolation of measurements at multiple pressures to zero pressure … more work needed here.
Slide30Unfortunately, birth positions yield different mean polarization angles
KSTAR_MSE_MEMO_015r.pdf Page / 36 30
As we change the torus pressure, we will change the relative weighting of the different birth positions to the total secondary beam neutral emission.
The torus pressure may have a complicated effect on the measured polarization angle.
Effect shown (on right) is the worst case identified to date.
Effect is often smaller for other sightlines and other magnetic-field pitch angles.
More work needed to determine when a linear extrapolation to zero pressure is accurate to ~0.2
o.
Slide31Behavior of emission from gFL
=0 is uniqueKSTAR_MSE_MEMO_015r.pdf Page / 36 31
Calculation says that emission from
all
gyro angles misses the filter passband for the inner sightlines (nominally 1-8).
These inner sightlines should not experience any emission from secondary beam emission and should therefore exhibit none of the peculiar behavior observed for the other field line pitch angles.
But what about the
outer sightlines, 9-25?
Slide32For the outer sightlines at
gFL=0, mean polarization angle of secondary emission is identically zeroKSTAR_MSE_MEMO_015r.pdf Page / 36
32
Probably due to some geometric identity.
Consistent with very weak spectral sensitivity that is observed at
g
FL=0.Calculation predicts absolutely no effect of torus pressure on measured angle at gFL=0. But a little is observed.Can we use this to get an angular offset, i.e. a true measurement of the Faraday effect, and then just apply it to all sightlines?That would be predicated on various ‘3D’ effects being small.
Slide33Extrapolating to zero pressure does not quite yield g=0.
KSTAR_MSE_MEMO_015r.pdf Page / 36 33
Differences are of order 1.5
o
.
This could be due to a combination of Faraday rotation and 3D viewing effects.
Or it could be a manifestation of nonlinear behavior, i.e. significant beam-ion attenuation between birth and crossing the
midplane.Analysis is on-going.
Slide34Secondary emission from beam neutrals is a likely explanation of the anomalous behavior at large filter offset and low BT
.Observed previously on C-Mod with near-perpendicular injection.Shown computationally to be a problem even for tangential injection.Consistent with the strong dependence of the anomaly with polarization angle.E
xplains why the anomaly is strongest at low B
T: smaller Stark shift, so a given filter ‘offset’ puts the filter passband further from the MSE p4
line: effectively, at large filter offset we’re observing only secondary emission.E
xplains why the sensitivity of angle to filter offset is larger in the MSE-BP system than in the existing KSTAR-MSE system: the MSE-BP filter fwhm
is 065 nm vs 0.45 nm for the KSTAR-MSE system.
Spectroscopic evidence: polarized emission observed in light passed by ‘blue’ filters.Summary: secondary emission from beam neutralsKSTAR_MSE_MEMO_015r.pdf Page / 36 34
Slide35Implication: the anomalous behavior at large filter offset doesn’t imply that anything is ‘wrong’ with the MSE diagnostic. Looking at large filter offset data is just looking at garbage.
Implication: to minimize the effect of filter offset, we should operate at the filter offset which maximizes the strength of the polarized light signal maximize ratio of MSE primary p emission to secondary emission. This implies operating at a filter offset of ~0.1 – 0.25 nm.
Summary: secondary emission from beam neutrals
KSTAR_MSE_MEMO_015r.pdf Page / 36
35
Slide36A problem: polarization angle differs between 1st and 2
nd beam pulsesKSTAR_MSE_MEMO_015r.pdf Page / 36 36
Three shots with PF=0 for all five beam pulses (one beam pulse failed – the only failure out of 150 beam pulses during the 30 shots).
Same torus pressure on 1
st
and 2
nd
beam pulses should measure same polarization angle.Differences thought to be due to differences in beam itself (next slide).Difference is larger at 1.2 T (0.5o – 1.0o) than at 2.5 T (~0.25o).
Slide37The first beam pulse differed from the others by a small amount
KSTAR_MSE_MEMO_015r.pdf Page / 36 37
Beam voltage is very similar (0.2
keV difference
Dl = xxx nm, not a problem).
Consistent difference of 3% in beam current.
Should affect beam perveance angular spread of beam polarization angle.Need to understand how much a 3% change in beam current might affect spread of beams.I think this demonstrates just how sensitive we are to beam spread, 3D collection effects, etc.
Slide38P
erform beam-into-gas calibration at g=0 only. Assume that the same ‘offset’ applies to all pitch angles.Advantage: requires minimal run time.
Perform standard beam-into-gas calibrations at multiple pitch angles, multiple angles
extrapolate each pitch angle data to zero pressure independently.
Advantage: does not require assumption that polarization angle offset is independent of polarization angle.
Disadvantage: requires more run time.
Could do approach #2 once, which would allow assessment of accuracy of approach #1.
Both approaches require assumption of linearity of extrapolation to zero pressure. Validity / accuracy of this assumption requires additional work.Possible approaches to an accurate beam-into-gas calibrationKSTAR_MSE_MEMO_015r.pdf Page / 3638
Slide39The beam-into-gas calibration successfully identified optimum filter ‘offset’ for entire range of BT (1.2 – 2.7 T).
Many anomalies in the calibration data now understood to be caused by secondary beam emission.First beam pulse seems to be a little different than the others.This does seem to affect the measurement.Either ‘fix’ the first beam pulse, or ignore it.Going forward, we need data at multiple torus pressures and more polarization angles to obtain an accurate angular calibration.
Use data at multiple torus pressures to extrapolate to p=0 to eliminate effect of secondary beam emission.
Validity of extrapolation procedure should be verified computationally.Residual worry: if we operate at small filter offset to minimize secondary beam emission, do we risk errors arising from changes in Is /
Ip between b2g and plasma measurements?
Conclusions
KSTAR_MSE_MEMO_015r.pdf Page / 36
39
Slide40Extra slidesKSTAR_MSE_MEMO_015r.pdf Page / 36
40
Slide41Ratio of atot (blue / pi) increases with torus pressure
KSTAR_MSE_MEMO_015r.pdf Page / 36 41
BT = 2.5 T, sightline 13.
Doesn’t quite match expectations: expected
that a linear fit would have ~zero intercept.
Slide42Similar behavior at 1.2 T
KSTAR_MSE_MEMO_015r.pdf Page / 36 42
BT = 1.2 T, sightline 12
Again, there is a non-zero intercept.
Slide43No surprises comparing ‘slope’ at 1.9 and 2.5 T
KSTAR_MSE_MEMO_015r.pdf Page / 36 43
Higher error bars for slopes at B
T = 1.2 T.
Possible technical problem with sightline 7.
Slide44Shot-to-shot reproducibility in measured angle is ~0.5 – 1.0
oKSTAR_MSE_MEMO_015r.pdf Page / 36
44 Pairs of successive shots at same filter temperature
same filter offset.
Filter temperature had been changed prior to first shot in the pair.
Probably due to simple blunder: not waiting sufficient time for filters to fully equilibrate.
We did think carefully about filter-heating issues and time-between shots.Simple transcription error. Once thermocouple reaches demand temp, 4 minutes needed for filter themselves to reach that temp. Only 1 minute allowed.
Slide45Expected behavior
True polarization angle
Measured polarization angle
equality
1.2 T
2.4 T
o
ffset of filter
cwl
relative to
p
3, nm
Ideal case
D
(measured – true angle) [
deg
]
0
1
‘Realistic’ case: spectral effects (mixing
s
and
p
light) cause additional deltas
b
eyond Faraday rotation.
Faraday rotation
Expect
D
to asymptote to a
c
onstant value at high offset
where measure
p
light only.
D
o
ffset of filter
cwl
relative to
p
3, nm
0
1
Linear polarization fraction
Expect LPF to asymptote to a
c
onstant value at high offset
where measure
p
light only.
KSTAR_MSE_MEMO_015r.pdf Page / 36
45
Slide46KSTAR is in nonlinear regime due to gas pressure & tokamak size
KSTAR_MSE_MEMO_015r.pdf Page / 36 46
Df
drift up to 32
o.
R Df
= 1.12 m
lmfp-cx = 1.03 m @2.0 mBar = 0.41 m @5.0 mBar
Slide47A spectroscopic feature to the blue of Ha was observed in C-Mod b2g
KSTAR_MSE_MEMO_015r.pdf Page / 36 47
Intensity [arb]
Intensity [arb]
Near-perpendicular beam injection on C-Mod, which is expected to increase magnitude of secondary emission.
Slide48Ko[2007]: “Effect of secondary beam neutrals on MSE: experiment”, paper NP9008?, APS-DPP, November 2007.Scott[2007]: “
Effect of secondary beam neutrals on MSE: theory”, paper NP90087, APS-DPP, November 2007.Yuh[2008]: Howard Yuh et al., “Simulation of the motional Stark effect diagnostic gas-filled torus calibration”, RSI 79 10F523 (2008).Ko[2009]: “Upgrade of the motional Stark effect diagnostic on
Alcator
C-Mod”, Ph.D. dissertation, pp. 83-98, June 2009.Geelen[2013]: Paul Geelen, “Simulation of Motional Stark Effect on C-Mod using MSESIM and PERF”, internship report, Eindhoven University of Technology, 2013.
Ko[2015] J. Ko et al, ”
Analysis of neutral hydrogenic emission spectra in a tokamak”
2015
JINST 10 P10009.Mumgaard[2015]: Robert T. Mumgaard, “Lower Hybrid Current Drive on Alcator C-Mod: Measurements with an Upgraded MSE Diagnostic and Comparisons to Simulation.” Ph.D. Dissertation, pp 60-66. Plasma Science and Fusion Center report PSFC/RR-15-16 and DOE/ET-54512-396, June 2015. ReferencesKSTAR_MSE_MEMO_015r.pdf Page / 36 48
Slide49‘Blue’ filter sometimes sees higher polarized intensity than p measurement
KSTAR_MSE_MEMO_015r.pdf Page / 36 49
… and polarized fraction can be larger than for the
p
line.
This is mostly an existence proof …
without a numerical simulation it is hard to relate the signal intensity at the blue filter to the level of pollution in the
p-filter by secondary emission.Secondary emission in spectrum passed by p filters probably has a completely different sightline dependence compare to spectrum passed by blue filters.
Slide50Variation of linear polarization fraction with BT
KSTAR_MSE_MEMO_015r.pdf Page / 36 50
LPF generally increases with BT.
Better separation of
s
and p lines at higher BT
– larger Stark shift.
At higher BT, the maximum LPF occurs at a higher filter offset.Again, interesting trends but not necessarily problematic for the b2g calibration.
Slide51The expected spectroscopic feature is clearly observed in the ‘blue’ filter
KSTAR_MSE_MEMO_015r.pdf Page / 36
51
This is ‘smoking gun’ evidence for the presence of secondary beam emission in the measurements.
Slide52Similar behavior is seen across entire range of BT
KSTAR_MSE_MEMO_015r.pdf Page / 36 52
Slide53Original KSTAR MSE system has good reproducibility
KSTAR_MSE_MEMO_015r.pdf Page / 36 53
Original KSTAR MSE diagnostic uses tilt-tuning to control filter
cwl.
These filters are mounted in magnetic bases to provide quick changes.
On successive shots, removed alternate filters and simply re-inserted them.
Reproducibility of all sightlines is typically within statistical error.
This suggests that the offset observed during the MSE background polychromator b2g calibration was due to insufficient time for filters to equilibrate in temperature.
Slide54Extrapolate calibration data to pressure=0 using only data at PF=0
KSTAR_MSE_MEMO_015r.pdf Page / 36 54
Plotted: difference between measured
polarization angle and true polarization
angle as a function of true polarization angle.
Data extrapolated to pressure=0 (red
) shows similar trends as original data.
Slide55KSTAR_MSE_MEMO_015r.pdf Page / 36
55 We have worried about the effect of secondary beam emission on MSE calibrations for > 10 years
Scott 2007 APS-DPP abstract: