ITER plasma rotation and Ti profiles from high-resolution crystal spectroscopy
R Barnsley, L-C Ingesson, A Malaquias & M O’Mullane
ADAS/SANCO (Atomic data and impurity transport codes)
- Evaluation of suitable impurities and ionization stages.
- Simulations of line and continuum emission.
- Impurity contributions to Prad and Zeff.
Integration into ITER
- Vertical coverage with 2-D curved crystal optics and 2-D
- Two or more graphite reflectors for the region
inaccessible by direct views.
- Optimization of sensitivity.
- Simulation of signal-to-noise ratios.
- Study of quasi-tomographic derivation of rotation and Ti.
R Barnsley, Moscow, Nov 2003. 1
ITER-98 impurity profiles
R Barnsley, Moscow, Nov 2003. 2
ITER profiles used for SANCO ADAS / SANCO modelled line/continuum ratios
and signal modelling for H- and He-like Kr:
- Chord-integrated ratios.
- Reference case: f-Kr = 10-5 . Ne, Prad ~ 700 kW.
R Barnsley, Moscow, Nov 2003. 3
ADAS / SANCO results for f-Kr = 10-5 . ne:
- (Left) Ionization balance. (Right) Radiated power components and total.
- Prad ~ 700 kW (integrated over plasma volume).
- Zeff ~ 0.01
- Kr ionization stages down to ~ Kr 26+ have x-ray lines suitable for crystal Doppler spectroscopy.
- Most of the radiated power is not in the H- and He-like stages.
R Barnsley, Moscow, Nov 2003. 4
ADAS / SANCO results for f-Kr = 10-5 . ne:
- (Left) He-like Kr 34+, 1s 2-1s2p, 0.945 Å. (Right) H-like Kr 35+, 1s-2p, 0.923 Å.
- Line radiation: photon/cm 3.s.
- Continuum: photon/cm3.s.Å.
- For signal calculations, Deuterium continuum was multiplied by Zeff 2 (~2.22).
R Barnsley, Moscow, Nov 2003. 5
R Barnsley, Moscow, Nov 2003. 6
R Barnsley, Moscow, Nov 2003. 7
ITER-98 x-ray spectrometer array (XCS-A)
5 lines of sight
• Provides good neutron shielding
• Access to plasma remote areas
- Signal attenuation (10% transmission)
- Reflection from graphite implies
narrow bandwidth (~1%)
R Barnsley, Moscow, Nov 2003. 8
X-ray discrete multi-chord option
The new system is integrated
at eport9 (16 LOS)
and uport3 (5 LOS)
Direct viewing lines without
Two spectral arms are used
for each viewing line:
•One for He like Ar (edge)
•One for He like Kr (core)
R Barnsley, Moscow, Nov 2003. 9
Multi-chord X-ray spectrometer option
ISO views of eport9
R Barnsley, Moscow, Nov 2003. 10
Core views with continuous coverage on equatorial port 9
- Upper and lower systems give
continous coverage of the plasma
core r/a <~ 0.7
- Compatible with the option of
discrete lines of sight, by
- Reduced number of crystals and
- Spatial resolution ~10 mm.
- Plasma vertical position control
with soft x-ray array.
- Plasma rotation measurements
can still be performed by two
R Barnsley, Moscow, Nov 2003. 11
Two or more graphite reflector based lines of sight will complete plasma coverage
R Barnsley, Moscow, Nov 2003. 12
Option for equatorial port
- Allows continuous imaging
- Minimises blanket aperture
R Barnsley, Moscow, Nov 2003. 13
X-ray Views Referred to Mid-plane Profiles
R Barnsley, Moscow, Nov 2003. 14
Spherically Bent Crystal
+ Allows plasma imaging
+ Improves S/N ratio with smaller entrance aperture and smaller detector
f s/f m = -1/cos(2B)
- No real focus for B < 45°
f s: Sagittal focus f m: Meridional focus B: Bragg angle
R Barnsley, Moscow, Nov 2003. 15
Toroidally Bent Crystal
A Hauer, J D Kilkenny & O L Landen. Rev Sci Instrum 56(5), 1985.
When combined with asymmetric crystal cut, gives considerable freedom in location of foci.
R Barnsley, Moscow, Nov 2003. 16
2-D bent crystal
The source is deep and optically thin.
(not to scale)
A toroidally-bent crystal is required, to place the
spatial focus in the plasma.
Raw spatial resolution depends on:
- Crystal height
- Chord length in plasma
- Chord-weighted emission
- Optical aberrations and crystal bending
Requires / ~ 10-3 (cf. / ~ 10-4 for -focus)
For a crystal of height h:
- r(Uport) ~ h/6 ~ 1 cm
- r(Eport) ~ h/3 ~ 2 cm
- r/r ~ 100 (optically)
R Barnsley, Moscow, Nov 2003. 17
R Barnsley, Moscow, Nov 2003. 18
Factors leading to choice of Bragg angle
Low Bragg angle (~30°) :
+ Reduced dispersion: = /tan.
a) Smaller first-wall penetration for a given bandwidth.
b) Smaller detector movement for tuneable spectrometer.
+ Larger crystal radius for a given crystal-detector arm - helpful with long sight-line.
+ Greater choice of crystals for short wavelengths.
+ Detector more remote from port plug.
+ Reduced effect of conical ray geometry for imaging optics.
- Shallower input angle to detector - parallax problems with gas-chamber detector.
~ Requires a toroidal crystal for imaging at B < 45°
R Barnsley, Moscow, Nov 2003. 19
Effect of input geometry on Johann sensitivity
Johann optics allow us to trade S/N with band-pass, while
maintaining peak sensitivity at the central wavelength
Crystal Crystal filling factor
Shield “c” 1 2 3
R Barnsley, Moscow, Nov 2003. 20
Parameters of the upper port imaging crystal spectrometers
The upper port system consists of two spectrometers, able to observe both H- and He-like lines of
Ar and Kr.
Toroidally bent, asymmetrically cut, crystals give enough free parameters to:
1) Place the meridional (imaging) focus in the plasma ~6m
2) Place the sagittal (dispersion) focus in the port plug ~3m
3) Keep a compact crystal-detector arm ~1.3m
Crystal toroidal radii: Sagittal ~ 4m Meridional ~ 1m
Crystal aperture: ~25 x 25 mm2 Spatial resolution > 25mm
Ion species B range Crystal 2d (nm) range (nm)
Ar XVII / XVIII 26° -28° SiO2(1010) 0.851 0.375 - 0.400
Kr XXXV / XXXVI 26.5° - 28.5 ° Ge(440) 0.200 0.090 - 0.096
Detector: Aperture ~ 25mm x 100mm 2-D spatial resolution < 0.1mm
Candidate detectors: Advanced solid state e.g. CCD, or advanced gas detector e.g. GEM.
R Barnsley, Moscow, Nov 2003. 21
R Barnsley, Moscow, Nov 2003. 22
Outline detector specification
Total detector height (~800 mm) = observed plasma height (~4 m) x demagnification (~0.2)
Individual detector height: ~160 mm for 5 detectors
Detector width in direction: ~50 mm
Vertical resolution: ~5 mm, for >100 resolvable lines of sight
Horizontal resolution: ~0.1 mm
QDE / Energy range: > 0.7, 6 – 13 keV (Uport also 3 – 6 keV)
Average count rate density: ~106 count/cm2.s
Peak count rate density: ~107 count/cm2.s
n- background count density: ~104 count/cm2.s
(flux of 10 6 n-/cm2.s, 10% sensitivity. 90% shielding)
This performance is typical of detectors in use or in development for high-flux sources such as
- Gas-microstructure proportional counters.
- Solid state arrays with individual pulse processing chain for each pixel.
R Barnsley, Moscow, Nov 2003. 23
Calculated signals for reference case:
- f-Kr = 10-5 . Ne Prad ~ 700 kW Zeff ~ 0.01
- Vertical image binned into 35 chords.
- Poisson noise added for 100 ms integration time.
R Barnsley, Moscow, Nov 2003. 24
Estimated Poisson signal-to-noise ratios based on counting statistics
- SNR ~ (Integral counts in line) / sqrt(line + continuum + n-background).
- Main noise source for data reduction is continuum, not n-background.
- A wide operational space is available between 10 -7 < f-Kr < 10-4.
- Uses a modest instrument sensitivity of 1.4 . 10 -7 cm2 per chord. (10x higher is possible).
R Barnsley, Moscow, Nov 2003. 25
Fits to the simulated noisy raw data
- Illustrative of the raw data quality – (obviously) not
the best method of analysis.
- Due to the narrower profile, chord-integral effects
are less for H-like Kr than for He-like.
- For r/a > 0.7, lower-ionized Kr ions or lower-Z
impurities are required.
- Under favourable conditions, a quasi-tomographic
deconvolution is possible (L-C Ingesson et al).
R Barnsley, Moscow, Nov 2003. 26