measurement of the passive fast-ion d-alpha (fida) emission on … documents/ep 17th... · the 15th...

1The 15th IAEA Technical Meeting on the Energetic Particles, G.Z.Hao, et.al, 2017

G.Z.Hao, W.W. Heidbrink, D.Liu, L.Stagner (UCI)

M.Podesta, A.Bortolon (PPPL)

Measurement of the passive fast-ion D-alpha (FIDA) emission on the NSTX-U tokamak

The 15th IAEA Technical Meeting on the Energetic Particles

Sep. 5-8, 2017

Princeton, USA


Outline

� Introduction

� FIDA diagnostic and experiment conditions

� FIDASIM modeling of active and passive FIDA signal

� Passive FIDA measurement

� summary


FIDA diagnostic measures fast ion density profile through charge-exchange(CX) spectroscopy

W.Heidbrink,RSI(2010),

PPCF(2004)

� Fast ion exchanges an electron with the neutrals� active-FIDA: emission by CX between fast ions and commonly injected neutrals

� passive-FIDA: FIDA emission related to background neutrals

� FIDA: collect the Doppler-shifted Balmer-alpha light

� Doppler shift of the emitted photon depends on the fast ion velocity

Neutral

(injected or background)


Basic parameters of NSTX-U

Beamline 2 (outboard) Beamline 1(inboard)

(2C)(2B) (2A) (1C) (1B) (1A)


Active and passive FIDA emissions can be monitored on NSTX-U

� FIDA diagnostics:

� Space resolution: ~5 cm, with16 channels cover 85~150 cm

� Temporal resolution: 10 ms for v-FIDA, 10 or 5ms for t-FIDA

� Energy resolution : ~10KeV

� The v/t-FIDA are most sensitive to the trapped and passing particles, respectively

� Every line of sight (LOS) intersects the last closed flux surface and can measure p-FIDA light

from the edge. While, only the LOS that intersects the injected beam can measure a-FIDA

signal.

-150 -100 -50 0 50 100 150X(cm)

-150

-100

-50

0

50

100

150

Y(c

m)

(b)

2A1C

2C

1B

0 50 100 150 200R (cm)

-200

-100

0

100

200

Z(cm

) NSTX-U #205080,610ms

(a)v-FIDA

t-FIDA

Plan view of FIDA geometryElevation view of FIDA geometry

Red: active reviews

Blue: ref. reviews

Active beamline


A brief schematic of FIDA system

� The light collected by the paired views

passes through a bandpass filter, grating,

filter strip and lenses before it is acquired by

CCD camera in 2X16 spectral images.

� Typically, the cold D_alpha line is about a

thousand times larger than the FIDA

emission.

� Scattering of the D_alpha light by optical

instruments can cause the contamination of

the FIDA spectral

cold D_alpha line


Evidence of the existence of scattered light in measured raw signal

� The finite CCD counts are obtained at the

boundary pixels of CCD, which is expected

to be zero due to the presence of the

bandpass filter in the optical system.

� The CCD counts at boundary pixels depend

almost linearly on the cold D_alpha intensity

summed over several dominant fibers.

� The scattering level is in the order of ~0.1%

for the test case carried out on the table. In

NSTX-U experiment, the scattering level

may be different from this value.

� Singular value decomposition (SVD) method

is used to make the scattering correction on

the measured raw data.

(Hao, et.al. submitted to Rev. Sci. Inst.,2017)


� After scattering correction, both FIDA spectra and spatial profiles get much better

agreement with predictions from FIDASIM. This scattering method works

reasonably well for most fibers, but fails in a few innermost fibers in some cases

because those active and passive views intercept the divertor at different radii.

� In the below presented signals, the scattering correction is carried out.

After making scattering correction, FIDA signal is in better agreement with modelling


Experimental setup for FIDA emission measurement

#205080

0.5

0.6

0.7

0.8

(a) Ip(MA)

0.0

0.5

1.0

1.5

2.0(b) PNBI (MW) 1C 2A

0.6

0.8

1.0

1.2

1.4

1.6

1.8

(c) ne (1019 m-3)@117cm

@140cm

0.6 0.8 1.0 1.2Time (s)

0.0

0.2

0.4

0.6

0.8

1.0(d) Te (keV)@117cm

@140cm

� In this study, the typical discharge

is a deuterium plasma in a low-

confinement mode in a limiter

configuration.

� In the flat-top phase, the central

density n_e~1.4E19, and

temperature T_e ~0.9 keV.

� Most results shown below are

averaged over 5 cycles.

current

Density

Temperature


Model the FIDA signal with FIDASIM

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6R (m)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Z (

m)

NSTX-U#205080,929ms

(a)

-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5X (m)

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Y (

m)

(b)

Orbit of an 83 keV fast ion born from the 2A source.

� The a-FIDA simulation is standard: fast-ion

distribution from NUBEAM output is sampled

and the light produced by CX between fast ions

and the injected and halo neutrals is computed

� The p-FIDA simulation:

� Axisymmetric populations of fast ions are

employed;

� Toroidally asymmetrical cluster of fast

ions is ignored, since the born orbit of fast

ion is far less concentrated toroidally than

a typical DIII-D case [Bolte NF, 2016];

� Fast ions expelled to the edge region are

ignored, since the typical discharge is

MHD-quiescent.

� Uncertainty in the edge (i.e. background)

neutral density is the dominant

uncertainty in p-FIDA modelling


� Owing to the large gyro-radius

of fast ions in spherical

tokamak, fast-ion orbits do

traverse the edge, which

produces the appreciable p-

FIDA signal in theory.

� The background neutral

density profile is computed by

TRANSP code.

The used fast ion and background neutral density profiles in FIDASIM simulation

#205080

0.1

0.2

0.3

0.4

0.5

0.6

(10

12 c

m-3)

(a) fast ion density

gyroradius

110 120 130 140 150R (cm)

108

109

(cm

-3)

(b) edge neutral density


The experimental signals agree well with FIDASIM modelling

0.0

0.5

1.0

1.5

2.0

PN

BI(M

W)

(a)

NSTX-U#205080

2A1C

0.0

0.1

0.2

0.3

0.4

0.5

(b) t-FIDA @117cm

0 20 40 60 80 100 120Relative time (ms)

0.0

0.2

0.4

0.6

0.8

(

10

16 p

h/s

-sr-

m2)

(c) v-FIDA @117cm

Cycle-averaged FIDA signal for (b) t-FIDA and (c) v-FIDA.

Red : a-FIDA plus p-FIDA signals; Blue : p-FIDA signal

Curves with and w/o error bars denote experimental and

simulated results. Amplitude of simulated results is

rescaled to match the measured total photons.

integrated in 651-654 nm

� The trend of the time evolution of the

measured FIDA signals agrees well with

the simulated results for both v- and t-

FIDA for the chosen fiber (not for all

fibers).

� The p-FIDA signal depends on the beam

injection geometry:

� For v-FIDA, the p-FIDA light is

largest when more perpendicular

source (1C) is injected; while, for t-

FIDA, the p-FIDA signal is equally

large when more tangential source

(2A) injects

� For t-FIDA, the p-FIDA light is comparable

to a-FIDA signal; p-FIDA contribution to

the total signal is larger than that for the v-

FIDA when active beam is injected.


Passive signals are often comparable to active signals

� The p-FIDA signal of t-FIDA is generally

comparable to a-FIDA. The passive

contribution to the total signal for t-FIDA

(when active beam is on) is larger than

that for v-FIDA

� The p-FIDA light of t-FIDA for more

perpendicular sources is comparable to

that for more tangential beam sources.

� During 2C blips, sawteeth occur, which

is likely the reason for the poor

agreement between the active and

reference view signals of t-FIDA during

2C blips.

Database results from the blip shots.

Each sample represents one time

slice in the database. Red: active view

data; Blue: ref. view data.


FIDASIM predictions successfully describe the measured bumps in the spectra for outboard beam-source (2A)

Cycle-averaged spectra of p-FIDA signal measured on the t-

FIDA active views during beam-2A on (red) and 2A off (blue)

at (a) R=140 cm and (b) 117 cm. (c,d) Difference between the

beam-on and the beam-off spectra (red) and the FIDASIM

simulation (black).

� The bumps in the spectra

related to the non-fully

slowed-down distribution

(shown in next slide) of fast

ions is measured. This

features are less prominent

on outer fiber (R=117 cm)

than on the inner channel

(140 cm).

� FIDASIM predictions

successfully describe the

measured bumps in the

spectra.


Non-fully slowed-down distribution of fast ions has local maxima in phase space

20 40 60 80Energy(keV)

-0.5

0.0

0.5

1.0

Pitch

0.0 0.2

0.3

0.5

0.7

0.8

1.0

(b)

#205080, @R=117cm

-1.0

-0.5

0.0

0.5

1.0

Pitch

0.0 0.2

0.3

0.5

0.7

0.8 1.0

(a)

20 40 60 80Energy(keV)

-0.5

0.0

0.5

1.0

Pitch

0.0 0.2

0.3

0.5

0.7

0.8

1.0

(b)

#205080, @R=140cm

-1.0

-0.5

0.0

0.5

1.0

Pitch

0.0 0.2

0.3

0.5

0.7

0.8 1.0

(a)

Fast-ion dist.

Weight-function (651-654nm)


For monitoring passing fast-ion dynamics in time using active beam method, p-FIDA contribution should be subtracted on NSTX-U

� The experimental spectra agree well with simulated results for the inboard beam-source(1C) injection.

� At the R=140 cm chord, the a-FIDA emission (black curve in (c)) is almost zero. For both the active

and reference views, the measured signals are mainly produced by the passive FIDA emission.

� At the R =117cm, the p-FIDA is comparable to a–FIDA as shown in (f)

� For monitoring passing fast-ion dynamics in time using the active beam method, P-FIDA contribution

should be subtracted for inboard beam injection.

Black: a-FIDA signal

Blue: p-FIDA signal


� The p-FIDA signal is larger at

the edge

� P-FIDA increases with major

radius in the region R > 127 cm

Experimental radial profiles of integrated p-FIDA signal agree well with FIDASIM prediction

Integration of the cycle-averaged p-

FIDA signal of t-FIDA. Amplitude of

simulated results is rescaled to match

the measured total photons.


Time-slice method yields larger a-FIDA signal than reference view method due to time-evolving passive contribution

� Time-slice method ignores the evolution of the p-FIDA emission in the neighbor time slices during

beam modulation.

� For the reference view method, the p-FIDA emission is monitored in time and p-FIDA contribution

is successfully removed.

� When using time-slice subtraction to determine a-FIDA emission in the presence of time-evolution

of p-FIDA signal, care must be taken to correct for the temporal variation of p-FIDA signal.


Summary

� Passive FIDA light makes a relatively large contribution to the total FIDA signal in

NSTX-U. Four features of the signals are consistent with the modeling of the p-

FIDA light.

(i) The time evolution of the signals agree well with theory

(ii) The p-FIDA signals show the expected dependence on beam injection angle

(iii) The spectra agree with the theoretical prediction for both inboard and

outboard beam injection

(iv) The radial profile agrees with the theory prediction

� The p-FIDA signal is detectable for fast ions in the edge region. The p-FIDA

technique may be employed to monitor the fast-ion dynamics in the edge region,

as long as the bremsstrahlung emission is not too strong to drown out FIDA

emission.

� When p-FIDA light is appreciable and time evolving, a reference view is needed

to measure the active FIDA emission

� For more precise modeling of the p-FIDA emission, a two- or three-dimensional

calculation of the edge neutral density is required. This, however, is left for future

work.

measurement of the passive fast-ion d-alpha (fida) emission on … documents/ep 17th... · the 15th...

Documents