a compact neutral-particle analyzer for plasma diagnostics

0020-4412/04/4702- © 2004

MAIK “Nauka

/Interperiodica”0214

Instruments and Experimental Techniques, Vol. 47, No. 2, 2004, pp. 214–220. Translated from Pribory i Tekhnika Eksperimenta, No. 2, 2004, pp. 87–93.Original Russian Text Copyright © 2004 by Chernyshev, Afanasyev, Dech, Kick, Kislyakov, Kozlovskii, Kreter, Mironov, M. Petrov, S. Petrov.

INTRODUCTION

The experimental conditions provided by modernthermonuclear facilities have considerably changed therequirements for the equipment for plasma diagnostics.For example, the diagnostics of charge-exchange atomsrequired that a compact neutral-particle analyzer bedeveloped [1–8]. The reasons for this were as follows:

First, smaller-size analyzers can be built more easilyat experimental facilities. A compact device can beinstalled virtually anywhere near a plasma machine. Itcan be easily moved somewhere else or replaced, ifnecessary, with another device. A set of several com-pact analyzers can be used for multichord plasma diag-nostics.

Second, diagnostic equipment intended for modernplasma facilities or fusion reactors of the future shouldbe able to operate in high neutron and

γ

-ray fields. Unfor-tunately, the detectors used in analyzers (channel multi-pliers and microchannel plates) are sensitive to neutronsand

γ

rays. Therefore, such detectors must be used withmassive shields, which reduce the intensity of irradia-tion. The weight of a shield required to protect a devicefrom the 2.5 MeV D–D neutrons produced by a plasmafacility with a neutron yield of

~5

×

10

14

neutron/s canbe roughly estimated as follows:

The analyzer and plasma are usually separated byseveral meters. This distance cannot be significantlyincreased without reducing the data statistics. The neu-tronflux density at a distance of 3 m from the facility is

~5

×

10

8

neutron/(s cm

–2

). Estimates show that a cham-

ber with double walls made of a 15-cm-thick layer ofborated polyethylene and a 5-cm-thick layer of lead isnecessary to reduce the neutron flux at the site of thedevice to an acceptable level (

~10

7

neutron/(s cm

–2

)).Such a chamber could accommodate an analyzer of amedium size (~1 m, including the vacuum pumpingsystem). The chamber weight is estimated at 4 t, whichis of limited feasibility technically. As for the futurefusion reactors that will use D–T plasma, they willrequire protection from neutrons with an energy of14 MeV. Protective shielding used for this purposewould be even more bulky.

Thus, downsizing the analyzer allows the weight ofthe protective shielding to be considerably reduced. Forexample, a threefold decrease in analyzer size makes itpossible to reduce the weight of the neutron shieldingby an order of magnitude. This shielding is suitableeven for the case of D–T plasma.

Third, the equipment of present-day plasma facili-ties and fusion reactors of the future must be able tooperate in strong magnetic fields. Usually, the scatteredmagnetic field at the site of analyzers is

~10

–2

–10

–1

T.Unfortunately, the design of medium-size analyzersmakes them sensitive to external magnetic fields. Theprinciple of operation of these devices is based on thespectrometry of secondary ions produced by stripping.In medium-size analyzers, the secondary ions travel aconsiderable distance from the point of stripping to thedetector, and their trajectories are affected by externalmagnetic fields. Therefore, the use of analyzers in facil-ities with magnetic plasma confinement usually

GENERAL EXPERIMENTALTECHNIQUES

A Compact Neutral-Particle Analyzer for Plasma Diagnostics

F. V. Chernyshev

1

, V. I. Afanasyev

1

, A. V. Dech

1

, M. Kick

2

, A. I. Kislyakov

1

, S. S. Kozlovskii

3

, A. Kreter

4

, M. I. Mironov

1

, M. P. Petrov

1

, and S. Ya. Petrov

1

1

Ioffe Physicotechnical Institute, Russian Academy of Sciences, ul. Politekhnicheskaya 26, St. Petersburg, 194021 Russia

2

Max-Planck-Institut für Plasmaphysik, Association Euratom-IPP, Garching, D-85748 Germany

3

St. Petersburg State Technical University, ul. Politekhnicheskaya 29, St. Petersburg, 195251 Russia

4

Institut für Plasmaphysik, Forschungszentrum Jülich GmbH, Trilateral Euregio Cluster, Association Euratom, Jülich, D-52425 Germany

Received June 24, 2003; in final form, August 25, 2003

Abstract

—A new compact neutral-particle analyzer (CNPA) developed at the Ioffe Physicotechnical Institute,Russian Academy of Sciences, is described. The device is used as a mass and energy spectrometer for the simul-taneous analysis of the hydrogen (0.8–80 keV) and deuterium (0.66–36 keV) charge-exchange fluxes emittedby a plasma. A thin (100 Å) diamond-like foil is used for stripping instead of the conventional method of strip-ping in gas. The analyzing magnetic field is produced by two powerful (1 T) permanent NdFeB magnets insteadof conventional electromagnets. These two innovations have made it possible to decrease considerably the size(

169

×

302

×

326

mm) and weight (42.5 kg) of the analyzer. To increase detection efficiency, the device usesadditional electrostatic acceleration of ions scattered by the stripping foil and provides a magnetic field config-uration with two-coordinate focusing. The analyzer has been used in experiments on the Wendelstein 7-AS stel-larator at the Max Planck Institute of Plasma Physics (Garching, Germany). The results of the first measure-ments performed using this analyzer are described.

INSTRUMENTS AND EXPERIMENTAL TECHNIQUES

Vol. 47

No. 2

2004

A COMPACT NEUTRAL-PARTICLE ANALYZER FOR PLASMA DIAGNOSTICS 215

requires massive magnetic shields. As a result, the sizeand weight of the equipment increase considerably. Thechamber walls in an analyzer are conventionally madeof soft magnetic iron and have a thickness of severalcentimeters. This provides a reduction of the externalmagnetic field by a factor of 20–100, but leads to a con-siderable increase in the weight of the device. Forexample, the magnetic shields for an analyzer with acharacteristic size of ~1 m may weigh ~100–500 kg. Adecrease in the analyzer size makes it possible to reduceconsiderably the weight of the magnetic shields.

In this work, the design and principle of operation ofa compact neutral-particle analyzer are described andthe calibration data are given. We also describe theresults of the first measurements performed using thisanalyzer on the Wendelstein 7-AS stellarator (IPP,Garching, Germany).

DESIGN AND PRINCIPLE OF OPERATION

The compact neutral-particle analyzer (CNPA) is amass and energy spectrometer intended for the simulta-neous analysis of the hydrogen (0.8–80 keV) and deu-terium (0.8–40 keV) charge-exchange fluxes emittedby a plasma. The device’s principle of operation isbased on the stripping of incoming neutral particles andthe subsequent analysis of secondary ions in parallelmagnetic and electric fields. A diagram of the device isshown in Fig. 1; the main parameters are given in Table 1.

Stripping and Accelerating System

The incoming flux of neutral atoms is ionized as itpasses through a thin (100 Å) diamond-like foil [9]. Themain advantage of this technique over the conventionalmethod of stripping in gas is that it has no need of anintrinsic vacuum pumping system. The vacuum pump-ing system of the experimental setup is sufficient forstripping. In addition, for particles with energies higherthan 10 keV, stripping in a foil is far more effective thanstripping in gas. This may be important for studyinghigh-energy ion tails, the intensity of which usuallydecreases with increasing energy.

The main disadvantage of the foil is its strong scat-tering of particles with energies lower than 10 keV. Thisleads to a decrease in detection efficiency at low ener-gies. In order to reduce the particle loss caused by scat-tering during stripping and to increase the device’s sen-sitivity in the low-energy range, the CNPA uses anadditional electrostatic acceleration of ions. For thispurpose, a voltage of +5 kV is applied to the strippingfoil. To prevent undesirable deflection of particlescaused by external scattered magnetic fields and theedge fields of the analyzing magnet, the stripping foiland the acceleration gap are protected with a magneticshield.

Dispersion System

A conventional spectrometer circuit with parallelanalyzing magnetic and electric fields is used in theCNPA. As the secondary ions pass through the mag-netic field, they are deflected in the horizontal planethrough an angle of

~90°

and enter the analyzing elec-tric field produced by an electrostatic condenser. Thisfield deflects the particles in the vertical direction. Suchdouble magnetic and electric analysis provides massand energy separation of particles at the site of thedetectors (Fig. 1a).

The magnetic field is produced by two strong (1 T)permanent rare-earth NdFeB magnets manufactured byOAO NPO Magneton (Russia). The permanent mag-

4 5

A

+

6

7

3

2

1

A

0

3 5 6

HD

Mas

s

Energy

(b)

(a)

Fig. 1.

Diagram of the analyzer: (

1

) stripping and accelera-tion system; (

2

) stripping foil; (

3

) analyzing magnet;(

4

) Hall probe; (

5

) analyzing electrostatic condenser;(

6

) shielding mask at the entrance to the detectors;(

7

) detectors; (

A

0

) atomic flux emitted by plasma; (

A

+

) sec-ondary ions; (H) hydrogen detector array; and (D) deute-rium detector array.

216


Vol. 47

No. 2

2004

CHERNYSHEV

et al

.

nets are placed on each side of the gap of the magnetthat produces the analyzing field (Fig. 2). It should benoted that the arm of the magnet yoke, in which the per-manent magnets and the magnet gap are located, hasconsiderably higher magnetic resistance than the sec-

ond arm of the yoke. Thus, the second arm protects thesensitive areas of the analyzer (in particular, the magnetgap) from external magnetic fields by closing themthrough itself. Calculations show that the analyzer canoperate in external magnetic fields of up to 5 mT with-out additional shielding.

Simulation of Particle Trajectories in Magneticand Electric Fields

To reduce particle loss and provide the focusing oftheir trajectories at the site of the detectors, the param-eters of the dispersion and acceleration systems of theanalyzer were optimized using computer simulationmethods. The magnetic field was specified by simulat-ing magnetic-field sources and conductors with a non-linear dependence

B

(

H

)

. The spatial distribution of themagnetic field was determined using the finite spatialelement method. The electric field was specified usingconductors with given potentials and dielectrics. Thespatial distribution of the electric field was determinedusing the finite superficial element method. A modifiedRunge–Kutta method was used to calculate the ion tra-jectories; the initial parameters of the particles wererandomly selected using the Monte Carlo method.

The simulated and measured distributions of themagnetic field in the magnet gap are given in Fig. 3.The magnetic field increases in the direction in which

Table 1.

Main parameters of the CNPA

Parameters Parameter value

Mechanical

Dimensions (with ports and connectors), mm 169

×

387

×

428

Dimensions of the analyzer housing, mm 169

×

302

×

326

Weight, kg 42.5

Solid angle of observation (maximum), m

2

sr

~2

×

10

–8

Electrical and magnetic

Voltage for the acceleration of secondary ions, kV + 5

Magnetic field in the analyzing-magnet gap, T 0.23–0.7

Voltage across the analyzing

condenser

, kV

+5

Permanent magnets used to produce the analyzing magnetic field

Magnet type (composition) NdFeB

Dimensions, mm 10

×

45

×

76

Residual magnetization

B

r

, T ~1

Detectors

Detector type

Channel electron multiplier

Number of detector arrays 2

Distance between detector arrays, mm 10

Number of detectors in the array

14 + 13 (up to 40 per detector array)

Permissible values of neutron flux and magnetic field at the site of the analyzer

D–D neutron flux (without additional shielding), neutron/(s cm

2

)

10

7

External magnetic field (without additional magnetic shield), mT

5

A

+

A

+

M

P

G

Y

Fig. 2.

Isometric plot of the analyzing magnet: (

M

) perma-nent magnet; (

P

) pole tip; (

Y

) second arm of magnet yoke;(

G

) gap of the analyzing magnet; and (

A

+

) trajectories ofsecondary ions.


Vol. 47

No. 2

2004


the ion trajectories emerge from the magnet. This pro-vides ion focusing in the vertical plane, whereas it isattained in the horizontal plane by selecting the optimalshape of the magnet poles.

Detectors

CEM KBL channel electron multipliers manufac-tured by Dr. Sjuts Optotechnik GmbH (Germany) wereused to detect ions passing through the analyzer. A pho-tograph of the detector unit is shown in Fig. 4. The maincharacteristics of the detectors are given in Table 2. Thetwo arrays for the detection of hydrogen and deuteriumcontain 14 and 13 detectors, respectively.

This analyzer modification uses detectors with 3-,5-, and 10-mm-wide entrance windows; this makes itpossible to optimize the energy characteristics of thespectrometer. In the array for the detection of hydrogen,a place is reserved for a 15th detector. It can be installedif the detected particles have an energy of 80 keV. Theanalyzer design makes it possible to install up to40 detectors with a 3-mm-wide entrance window in thedetection array.

Energy Range

Permanent magnets provide continuous energy tun-ing of the analyzer, i.e., constant energy distributionbetween its channels. In the CNPA, the energy rangecannot be changed by modifying the dispersion system(as is done in conventional analyzers with electromag-nets); this is why it detects particles over a very wideenergy range. The ratio between the simultaneouslymeasured maximum and minimum energies is 100 forhydrogen and 55 for deuterium.

RESULTS OF CALIBRATION

The analyzer was calibrated in an energy range of0.66 to 20 keV using monokinetic hydrogen and deute-rium atomic beams. The calibration was performed onan experimental setup similar to that described in [10].The following energy characteristics of the device weredetermined during calibration: the energy ratio betweenchannels, the energy resolution, and the detection effi-ciency. For energies higher than 20 keV, the analyzercharacteristics were determined by computer simula-tion based on calculation of trajectories using theMonte Carlo method. The results of calibration areshown in Figs. 5 and 6 in graphical form; the mainenergy characteristics of the CNPA are given in Table 3.

The energy ratio between channels and the energyresolution of the CNPA channels for the hydrogen anddeuterium detector arrays are shown in Fig. 5. Theexperimental values of the energy widths of the chan-nels are represented as vertical strokes; the shaded areaindicates changes in the energy width along the detec-tor array as determined using the model calculation.

The energy width of the channels at half-height var-ies from 60% for low energies to 10% for high energies.In Fig. 5, vertical lines indicate the regions correspond-ing to three groups of detectors with different widths ofthe entrance windows. The measured values of thechannel energies and the energy resolution agree wellwith the results of calculation, except for a minor dis-crepancy at energies lower than 4 keV (Fig. 5a). Thiscould be the result of an inaccurate theoretical estima-tion of the scattering that accompanies low-energyatom stripping.

The dependence of the absolute efficiency of thedetection of hydrogen and deuterium atoms on the

B

X

Cross sectionof the magnet

0.8

0.6

0.4

0.2

0

B

,

T

0 20 40 60 80 100

X

,

mm

Fig. 3.

Magnetic field in the analyzing-magnet gap. Pointsare the experimental data; a solid line presents the results ofcalculations.

Table 2.

Characteristics of CEM KBL detectors (accordingto catalog of Dr. Sjuts Optotechnik GmbH, Germany)

Parameter Parameter value

Entrance window, mm

(3/5/10)

×

10

Amplification factor

1

×

10

8

Pulse-height distribution width (FWHM), %

<40

Pulse duration (FWHM), ns

8

Maximum counting rate in the pulse count-ing mode, pulses/s

106

Dark current at a threshold of 2 × 106,pulses/s

<0.02

218

INSTRUMENTS AND EXPERIMENTAL TECHNIQUES Vol. 47 No. 2 2004

CHERNYSHEV et al.

energy per atomic mass is shown in Fig. 6. The detec-tion efficiency was determined as the ratio between thedetector counting rate and the level of the incidentatomic flux. At high energies, the efficiencies of thedetection of hydrogen and deuterium are nearly thesame. This can be explained by the equal probability ofthe stripping of isotope atoms with equal velocities. Inthe energy range below 10 keV, the efficiencies of thedetection of hydrogen and deuterium differ consider-ably. The cause of this difference is not completelyunderstood. It is probably caused by the better focusingof deuterium ions at the exit from the analyzing magnet.

It should be noted that the CNPA has a wide maxi-mum solid angle of observation (~2 × 10–8 m2 sr). Incombination with its high detection efficiency, itallows, if necessary, the device sensitivity to beincreased 10–100 times relative to conventional analyzers.

The calibration procedure also included estimationof the mass resolution of the CNPA. It was found thatthe neighbor-mass suppression factor (H/D) was better

than 1 : 1000. This value is consistent with the resultsof the model calculations.

TESTING ON THE WENDELSTEIN 7-AS STELLARATOR

The analyzer described in this work was tested onthe Wendelstein 7-AS stellarator at the Max PlanckInstitute of Plasma Physics (Garching, Germany).Mixed hydrogen–deuterium plasma was used in theexperiments. It had the following parameters: effectiveradius of plasma reff ~ 15.3 cm, toroidal magnetic fieldBtor = 2.5 T, electron density ne(0) = 5 × 1019 m–3, andelectron temperature Te(0) = 1.8 keV. Neutral injection(1.4 MW) and radio-frequency heating at electroncyclotron frequency (0.8 MW) were used for the addi-tional heating of the plasma. The analyzer was installedat an angle of ~45° to the plasma axis. The device’s lineof sight passed through the central region of the plasma.

Energy

Mas

s

Fig. 4. Photograph of the detector unit.

Table 3. Energy characteristics of the CNPA

Parameter H array D array

Minimum energy Emin, keV 0.8 0.66

Maximum energy Emax, keV 80 36

Dynamic energy range, Emin/Emax 100 55

Energy resolution (FWHM), % 60–10 for 0.8–80 keV

Neighbor-mass suppression factor (H/D) 1 : 1000

Detection efficiency 5 × 10–4–0.4 3 × 10–4–0.1

for 0.8–80 keV for 0.66–36 keV



The energy spectra of hydrogen and deuteriumatoms measured by the so-called passive method areshown in Fig. 7. These spectra constitute an integralcharacteristic along the analyzer’s line of sight; i.e.,they represent a superposition of spectra from differentregions of plasma with considerably differing parame-ters.

Reconstruction of the ion distribution in plasmafrom the passive spectra of charge-exchange atoms is arather difficult procedure. It includes model calcula-tions of atomic fluxes using the spatial distributions ofthe density and temperature of ions and electrons, aswell as the distribution of the density of neutral parti-cles. Solving this problem is beyond the scope of thiswork. We have restricted ourselves to a simplified inter-pretation of the experimental results. This interpreta-tion is based on the fact that high-energy atomic fluxes(in the case under consideration, fluxes with energieshigher than 3–4 keV) incident onto the device are pro-duced mainly in the hot inner region of the plasma. Thetemperature determined from the slope of the high-energy spectral region corresponds to the ion tempera-ture characteristic of this area of plasma. It should benoted that this method helps estimate the lower limit ofthe ion temperature in the central region, which can dif-fer considerably from the actual temperature, espe-cially in the case of a dense plasma. Such a qualitativeapproach yields an ion temperature of ~0.8 keV forboth hydrogen TH and deuterium TD. The ratio betweenthe hydrogen and deuterium atomic fluxes for the isoto-pic composition of the plasma is nH : nD ~ 1 : 2.

CONCLUSIONS

A new compact neutral-particle analyzer (CNPA)has been developed at the Ioffe Physicotechnical Insti-tute, Russian Academy of Sciences. The device has thefollowing advantages over conventional analyzers:

2

Γcx(E)/E0.5, rel. units

E, keV

109

108

107

106

4 8

H

0 6 10

1010

1011

1012

D

Fig. 7. Energy spectra of hydrogen and deuterium chargeexchange atoms, measured on the Wendelstein 7-AS stellarator.

100

Efficiency

E, keV/au

10–1

10–2

10–3

10–4

101 102

CalibrationModel

HD

Fig. 6. Efficiencies of the detection of hydrogen and deute-rium.

1 2 3 4 5 6 7 8 9 10 11 12 13

Calibration

Model

(a)102

101

100

Entrance windowof detector: 3 mm 5 mm 10 mm

EH, keV

1 2 3 4 5 6 7 8 9 10 11 12 13 Channel no.

Calibration

Model

(b)102

101

100

Entrance windowof detector: 3 mm 5 mm 10 mm

ED, keV

14 15

Fig. 5. Energy ratio between channels and the energyresolutions of the CNPA channels for (a) hydrogen and(b) deuterium detector arrays.

220


CHERNYSHEV et al.

• It has small dimensions (169 × 302 × 326 mm) anda low weight (42.5 kg), requires little free space, andcan be easily protected from radiation and scatteredfields; several analyzers can be easily combined into asystem for multichord plasma diagnostics.

• The analyzer has high sensitivity (10–100 timeshigher than that of conventional analyzers) due to itswide solid angle of observation and high efficiency ofdetection.

• Vacuum pumping is not required, since there is nogas puffing; power sources for the analyzing magnet arenot needed because permanent magnets are used.

The CNPA has been tested on the Wendelstein 7-ASstellarator.

REFERENCES1. Afrosimov, V.V., Berezovskii, E.L., Gladkovskii, I.P.,

et al., Sov. Phys. Tech. Phys., 1975, vol. 20, p. 33.

2. Gladkovsky, I.P., Izvozchikov, A.B., and Petrov, M.P.,Nucl. Instrum. Methods Phys. Res., Sect. A, 1980,vol. 175, p. 441.

3. Hayashi, K., Hashimoto, K., and Yamato, H., Rev. Sci.Instrum., 1985, vol. 56, p. 359.

4. Davis, S.L., Müller D., and Keane, C.J, Rev. Sci.Instrum., 1983, vol. 54, p. 315.

5. Izvozchikov, A.B., Petrov, M.P., Petrov, S.Ya., et al., Sov.Phys. Tech. Phys., 1992, vol. 37, p. 201.

6. Kislyakov, A.I., Khudoleev, A.V., Kozlovskij, S.S., andPetrov, M.P., Fusion Eng. Des., 1997, vol. 107, p. 34.

7. Medely, S.S. and Roquemore, A.L., Rev. Sci. Instrum.,1998, vol. 69, p. 2651.

8. Afanasyev, V.I., Gondhalekar, A., Babenko, P.Yu., et al.,Rev. Sci. Instrum., 2003, vol. 74, p. 2338.

9. Gott, Yu.V. and Motlich, A.G., Nucl. Instrum. MethodsPhys. Res., Sect. A, 1978, vol. 155, p. 443.

10. Afrosimov, V.V., Gladkovskii, I.P., Gordeev, Yu.S., et al.,Sov. Phys. Tech. Phys., 1961, vol. 5, p. 1378.

a compact neutral-particle analyzer for plasma diagnostics

Documents