Diagnostic performance of the beam emission spectroscopy system on theNational Spherical Torus ExperimentD. R. Smith, R. J. Fonck, G. R. McKee, and D. S. Thompson Citation: Rev. Sci. Instrum. 83, 10D502 (2012); doi: 10.1063/1.4728094 View online: http://dx.doi.org/10.1063/1.4728094 View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v83/i10 Published by the American Institute of Physics. Related ArticlesTuning the electron energy by controlling the density perturbation position in laser plasma accelerators Phys. Plasmas 19, 063104 (2012) Recent progress of the HCN interferometer on J-TEXT tokamak Rev. Sci. Instrum. 83, 10E303 (2012) Extreme ultraviolet spectroscopy and modeling of Cu on the SSPX Spheromak and laser plasma “Sparky” Rev. Sci. Instrum. 83, 10E101 (2012) A two photon absorption laser induced fluorescence diagnostic for fusion plasmas Rev. Sci. Instrum. 83, 10D701 (2012) A new interferometry-based electron density fluctuation diagnostic on Alcator C-Mod Rev. Sci. Instrum. 83, 10E301 (2012) Additional information on Rev. Sci. Instrum.Journal Homepage: http://rsi.aip.org Journal Information: http://rsi.aip.org/about/about_the_journal Top downloads: http://rsi.aip.org/features/most_downloaded Information for Authors: http://rsi.aip.org/authors

Downloaded 13 Jun 2012 to 198.35.0.239. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

REVIEW OF SCIENTIFIC INSTRUMENTS 83, 10D502 (2012)

Diagnostic performance of the beam emission spectroscopy systemon the National Spherical Torus Experimenta)

D. R. Smith,b) R. J. Fonck, G. R. McKee, and D. S. ThompsonDepartment of Engineering Physics, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA

(Presented 8 May 2012; received 8 May 2012; accepted 21 May 2012; published online 11 June2012)

The beam emission spectroscopy system on the National Spherical Torus Experiment measures lo-calized density fluctuations on the ion gyroscale. Optical sightlines provide core to edge radial cov-erage, and the sightlines are aligned to typical pitch angles to maximize cross-field spatial resolution.Sightline images are 2–3 cm, and point spread function calculations indicate image distortion frompitch angle misalignment and atomic state finite lifetimes is minor with a 15% increase in the imagesize. New generation photodetectors achieve photon noise limited measurements at frequencies up to400 kHz with refrigerant cooling at −20 ◦C. Measurements near the pedestal show broadband tur-bulence up to 100 kHz, and poloidal correlation lengths are about 10 cm. Plasma turbulence signalscan be 2–3 orders of magnitude above photon noise and amplifier thermal noise. © 2012 AmericanInstitute of Physics. [http://dx.doi.org/10.1063/1.4728094]

The beam emission spectroscopy (BES) diagnostic sys-tem on NSTX (Ref. 1) measures Dα emission (n = 3 → 2, λ0

= 656.1 nm) from deuterium heating beams.2 BES measure-ments are sensitive to ion gyroscale fluctuations associatedwith low wavenumber (low-k) turbulence and Alfvén modes.The beam velocity induces a Doppler shift in beam emission,and optical filters isolate beam Dα emission from thermal Dα

emission. The intersection of optical sightlines with the neu-tral beam volume provides spatial localization, but a rigorousassessment of spatial and k-space measurement propertiesrequire point spread function calculations.3 BES measure-ments are sensitive to plasma density fluctuations accordingto δIDα/IDα = C (δn/n) where C = C(ENB, n, Te, Zeff) ≈ 1/2.2

The NSTX BES system includes two optical assembliescentered at r/a ≈ 0.45 and 0.85.1 The optical views are pitchedupward at 26◦ and 37◦ for alignment with typical NSTX pitchangles to optimize cross-field spatial resolution. As shown inFigure 1, 56 sightlines provide core-to-SOL radial coverageand four discrete poloidal arrays, and sightline images are2–3 cm at the neutral beam. The measurements are sensi-tive to fluctuations with k⊥ρs � 1.5 where ρs ≈ 0.5–1.5 cm(ρs ≡ √

2miTe/qB is the ion sound gyroradius). Highthroughput optics collect light at f/1.5 with 2.3 mm2-ster én-tendue and image the light onto 9-fiber optical bundles fortransmission to photodetectors. Interference filters with 4 nmpassbands at 661 nm pass redshifted Dα emission from deu-terium heating beams and block thermal Dα emission. At nor-mal incidence, the interference filters pass full energy and 1/2energy beam components, but 1/3 energy components nearCarbon-II lines are attenuated. Tuning the filters to lowerwavelength can pass 1/3 energy components and Carbon-IIemission. Note that emission spectra from motional Stark ef-

a)Contributed paper, published as part of the Proceedings of the 19thTopical Conference on High-Temperature Plasma Diagnostics, Monterey,California, May 2012.

b)Electronic mail: drsmith@engr.wisc.edu.

fect and fast ion Dα diagnostics indicate the C-II emission iscomparable to or less than the neutral beam Dα manifold.1

New generation PIN photodiode detectors and frequency-compensating, wideband preamplifiers provide photon-noiselimited measurements at frequencies up to 400 kHz with re-frigerant (non-cyrogenic) cooling at −20 ◦C.1, 4 The data ac-quisition system samples at 2 MHz (output) with an onboardfield programmable gate array (FPGA) post-processor circuit.The FPGA functions as a finite impulse response filter to sup-press high frequency thermal noise. The wideband, low noisedetectors enable measurements of high frequency Alfvén andenergetic particle modes up to about 800 kHz in addition tolow-k turbulence at low frequency.

The BES poloidal array at R = 140 cm provides multi-point measurements for calculating the spatial and tempo-ral properties of turbulence. Figure 2 shows example BESmeasurements at two radial locations. Note that BES signalschange sharply at NBI steps, and BES auto-power spectracan exceed photodetector dark noise power spectra by 2–3orders of magnitude. Poloidally separated channels in the R= 140 cm poloidal array (r/a ≈ 0.80–0.95) (see Figure 1)provide measurements of the poloidal correlation length Lc,decorrelation time τ d, and poloidal wavenumber kθ associ-ated with low-k turbulence. Coherence spectra in Figure 3show reduced coherence with greater poloidal separation, asexpected, and cross-phase spectra show greater cross-phasewith greater poloidal separation. Autopower spectra (Figure3(c)) show plasma turbulence signals 2–3 orders of magnitudegreater than dark noise and photon noise levels, and filtereddata (Figure 3(d)) show turbulent eddies propagating alongthe poloidal array. The turbulence quantities are derived fromtime-lag auto- and cross-correlation functions in Figure 3(e).Data windows of 15–40 ms are segmented into 30–150 binswith 512–2048 time points per bin. The time-lag correlationfunctions for each bin are averaged to obtain the average time-lag correlation function and variance for the data window.The cross-correlation envelope (Hilbert transform) at zero

0034-6748/2012/83(10)/10D502/3/$30.00 © 2012 American Institute of Physics83, 10D502-1

Downloaded 13 Jun 2012 to 198.35.0.239. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

10D502-2 Smith et al. Rev. Sci. Instrum. 83, 10D502 (2012)

110 130 150120 140

5

10

15

0Z (cm)

R (cm)

FIG. 1. BES sightlines provide radial coverage from core to scrape-off-layerregions and four discrete poloidal arrays.

time-lag decreases with channel separation, and the poloidalcorrelation length Lc is the corresponding 1/e length as shownin Figure 3. The time-lag of the peak cross-correlation enve-lope increases with channel separation, and the linear relation-ship gives the eddy velocity vg . The peak cross-correlation en-velope decreases with time-lag, and the decorrelation time τ d

is the corresponding 1/e time-lag. The poloidal wavenumberkθ is derived from the eddy size inferred from the eddy ve-locity and the time-lab between auto-correlation anti-nodes.Finally, note that coherence spectra are above statistical un-certainties at frequencies below about 35 kHz in Figure 3(a),

−10

0

10

Low

−f (

G)

−20−10

010

D−

alf [

dB]

0

5

Neu

tron

s

0 0.5 1Time (s)

Fre

quen

cy (

kHz)

0

20

40

60

80

100

Fre

quen

cy (

kHz)

Aut

opow

er (

V2 /H

z; d

B)

0

20

40

60

80

100

Fre

quen

cy (

kHz)

0

20

40

60

80

100

−40

−20

0

20R=129 cm R=142 cm

0 0.5 10

5

Time (s)

Sig

nal (

V)

0 50 100

−40

−20

0

Frequency (kHz)

534 ms

R=129 cmR=142 cmNBI (au)

R=129 cmR=142 cmDark noise

AP

S (

V2 /H

z; d

B)

(a) (b)

(c)

(d)

(e)

(f)

(g)

FIG. 2. BES auto-power spectrograms for measurements at (a) R = 129 cmand (b) 142 cm; (c) low frequency odd-n magnetic fluctuations; (d) thermalDα emission measurements; (e) neutron measurements; (f) BES time-seriesdata with NBI power (green line); and (g) BES auto-power spectra at 534 mswith photodetector dark noise spectrum.

−200 −100 0 100 200−0.5

0

0.5

1

Time lag (µs)

Cor

rela

tion

kp = 0.16 cm-1

(e) (f)0 5 10

0

0.2

0.4

0.6

0.8

1

Channel separation (cm)

Cor

rela

tion

Lcp = 9.9 cm

512.7 512.8 512.9 513

−2

0

2

4

Time (ms)

Filt

ered

dat

a (a

u)

−4(d)(c)

1 100−50

−40

−30

−20

−10

0

Frequency (kHz)

Aut

o-po

wer

spe

ctra

(dB

[V2 /H

z])

10dark noise

photonnoise (ref)

Auto-powerspectra

0 50 1000

0.2

0.4

0.6

0.8

1

Frequency (kHz)

Coh

eren

ce

Δz = 2 cmΔz = 4 cmΔz = 6 cm

(a)0 50 100

0

0.5

1

1.5

Frequency (kHz)

Pha

se (

θ/π)

(b)

FIG. 3. (a) Coherence and (b) cross-phase spectra among four channels inthe BES R = 140 cm poloidal array during an ELM-free, MHD quiescent pe-riod; (c) auto-power spectra and photon and dark noise for reference channel;(d) filtered data (8–50 kHz) used for calculations in (e) and (f); (e) time lagauto- and cross-correlation functions and envelopes; and (f) correlation en-velope at zero time lag vs. channel separation. Black line in (c)–(e) denotesreference channel.

consistent with autopower spectra exceeding noise levels atfrequencies below about 35 kHz in Figure 3(c).

Backlight fiber images are 2–3 cm at the center of theneutral beam, but accurate spatial and k-space characteri-zation of BES measurements require point spread function(PSF) calculations.3 PSF calculations convolve optical systemproperties with neutral beam profiles, plasma geometry, andatomic state lifetimes5 that distort the optical image. For toka-mak geometry, PSF calculations can ignore small pitch anglecorrections and reduce the calculation to 2D without compro-mising accuracy, but steep ST pitch angles require full 3D PSFcalculations. Figure 4 shows fiber bundle images at R = 133,140, and 143 cm. At R = 140 and 143 cm, excited state life-times in the low density edge region dominate image distor-tion. At R = 133 cm, magnetic field misalignment dominatesimage distortion. Without excited state lifetime and magneticfield geometry effects, fiber images are 3.75 cm across at full-width half-max (FWHM, 50% of peak intensity). With theeffects, the FWHM images are in the range 3.6–3.9 cm, soimage distortion from atomic excited states and field line ge-ometry is mild.

In summary, the BES diagnostic system on NSTX mea-sures density fluctuations on the ion gyroscale to inves-tigate low-k turbulence and Alfvén modes. Sightline im-ages are 2–3 cm at the neutral beam, and point spreadfunction calculations indicate image distortion from fieldline misalignment and atomic state lifetimes are minor.

Downloaded 13 Jun 2012 to 198.35.0.239. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

10D502-3 Smith et al. Rev. Sci. Instrum. 83, 10D502 (2012)

−20 0 20 40 60 8080

100

120

140

160

Machine X Axis (cm)

Mac

hine

Y A

xis

(cm

) R=143 cmR=140 cmR=133 cm

NB sources

(b)

−10 0 100

2

4

6

8

10

12

Distance from NB center (cm)

Mis

alig

nmen

t with

B (

deg)

Y axis (cm)

Z a

xis

(cm

)

−5 0 5−5

0

5R=133 cm R=140 cm R=143 cm

(c) (d)

R=133 cm

R=140 cm

R=143 cm

110 120 130 140 1500

2

4

6

8

10D

ensi

ty (

1013

/cm

3 )

R (cm)

0

10

20

30

40

Pitc

h an

gle

(deg

rees

)optimal pitchangle

FIG. 4. (a) NSTX density and pitch angle profiles with BES measurements atR = 133, 140, and 143 cm; (b) BES sightlines crossing neutral beam sources,and short cross-lines show imaging planes; (c) sightline misalignment withmagnetic field along sightlines; and (d) PSF image contours at 50% (solid)and 10% (dashed) of peak intensity with magnetic field puncture points (dia-monds) on successive image planes.

New generation photodetectors with refrigerant cooling at−20 ◦C achieve photon noise limited measurements at fre-quencies up to 400 kHz with plasma turbulence signals2–3 orders of magnitude above noise signals. Initial measure-ments show broadband turbulence near the pedestal top withfrequencies up to 100 kHz and poloidal correlation lengthsaround 10 cm. Two-megahertz sampling and anti-alias fil-ters enable Alfvèn mode measurements at frequencies up to800 kHz.

This work was supported by (U.S.) Department of Energy(DOE) Grant Nos. DE-FG02-89ER53296, DE-SC0001288,and DE-AC02-09CH11466.

1D. R. Smith, H. Feder, R. Feder, R. J. Fonck, G. Labik, G. R. McKee,N. Schoenbeck, B. C. Stratton, I. Uzun-Kaymak, and G. Winz, Rev. Sci.Instrum. 81, 10D717 (2010).

2R. J. Fonck, P. A. Duperrex, and S. F. Paul, Rev. Sci. Instrum. 61, 3487(1990).

3M. W. Shafer, R. J. Fonck, G. R. McKee, and D. J. Schlossberg, Rev. Sci.Instrum. 77, 10F110 (2006).

4N. Schoenbeck, S. D. Ellington, R. J. Fonck, K. Jaehnig, G. R.McKee, D. R. Smith, I. Uzun-Kaymak, and G. Winz, Rev. Sci. Instrum.81, 10D718 (2010).

5I. H. Hutchinson, Plasma Phys. Controlled Fusion 44, 71 (2002).

Downloaded 13 Jun 2012 to 198.35.0.239. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions