PPPL—2395 NEUTRON ION TEMPERATURE MEASUREMENT __„_ „ DE87 005629
J.D. STRACHAN, H.W. HENDEL, J. LOVBERG,
and E.B. NIESCHMIDT
Plasma Physics Laboratory
Princeton University
Princeton, New Jersey 08544 USA
ABSTRACT
One important use of fusion product diagnostics is in the determination of the deuterium ion temperature from the magnitude of the 2.5 HeV d(d,n) He neutron emission. The detectors, calibration methods, and limitations of this technique are reviewed here with emphasis on procedures used at PPPL. In most tokam&ks, the ion temperature deduced from neutrons is in reasonable agreement with the ion temperature deduced by other techniques.
To be published in Proceedings of the Workshop on Diagnostics for Fusion Plasmas, (Varenna, Italy, 3-13 September 1986). MASTER
fllf.THIBUTION uF TitIS bOGUMEN'f IS UNLIMITED 1.0 INTRODUCTION
Measurements of the fusion reaction products have bet.i useful in
describing plasmas- from many tokamak devices. The original 111 and a
subsequently common use of these measurements has been to determine the
central ion temperature from the magnitude of the 2.5 MeV d(d,n) He neutron
emission. In this paper, we will review the techniques and limitations
associated with this measurement and will illustrate the procedures primarily with results from PPPL devices.
The 2.5 MeV neutron measurement presents challenges familiar to many
plasma physics diagnostics. Generally, one tries to improve spatial resolution, energy resolution, and frequency response while maintaining an accurate absolute calibration. For the specific purpose of deducing an ion temperature from the neutron emission magnitude, the aspects of frequency response and absolute calibratiin of the neutron source strength are most important and will be discussed in this paper. This topic was extensively reviewed at the 1982 Varenna diagnostic Conference /2-4/ and although much of the material in this paper must overlap material presented then, a new comparison of published neutron-deduced ion temperatures with the ion temperature from other diagnostics has been made for data published after
1£82, In this regard, a change in the role of the neutron measurements hss occurred. Before 1982, the neutrons were quoted primarily to substantiate charge exchange ion temperature III. Presently, with tokamak devices becoming much larger than the neutral particle mean-free path, there is a trend towards quoting neutron ion temperatures supported by impurity Doppler broadening measurements without reporting charge exchange data (e.g., the Doublet III results at the London IAEA Conference /5/). Further support for the neutron- flux-deduced ion temperature often comes from neutron spectral measurements.
2 2.0 NEUTRON DETECTOR SYSTEMS
On most tokamaks, the 2.5 MeV neutron emission can change by many orders
of magnitude a) between different experiments (Fig. 1) or b) in time through
the duration of a single plasma. The actual neutron flux at the detector will
be composed partly of virgin neutrons and partly of scattered neutrons. Since
the scattered neutrons are degraded in energy, it is often an advantage to measure the neutron intensity using moderated detectors which are insensitive to neutron energy so that virgin and scattered neutrons are counted with equal weighting. Thus a detector which is well shielded from the plasma by machine components will have an efficiency comparable to a detector which has a good view of the plasma.
Moderated detectors usually consist of an array of proportional counters
(Fig. 2) surrounded by an hydrogenous material (e.g., polyethylene).
Energetic neutrons are thermaliaed collisionally with the hydrogen in the moderator and are then transported to the proportional counters which usually have reasonably high cross sections at epithermal neutron energies. Typical
3 10 235 detectors /4,8,9/ employ He(ntp)t, B(nfa)Li, or U(n,f) epithermal reactions so that a neutron event in the proportional counter produces a specific energy pulse height corresponding to the Q of the detector reaction. The neutron count rate is determined by monitoring the counts in a pulse height window around this energy. For any detector, there are lower energy wall counts whose number is in proportion to the count rate in the neutron peak as well as noise counts which peak at low energy and are due to hard X rays or electromagnetic pick up. Typical operation of the detectors
3 includes spectral monitoring of the pulse heights to ensure that the detectors
are primarily counting neutrons.
In order to handle the large dynamic range (Fig. 1), several proportional counters with diffe-ent efficiencies are operated simultaneously. The PLT detector package in Fig. 2 consists of three BF,, detectors: a) the most efficient detector is enriched in B; b) the next most sensitive is a similar detector wrapped in a Cadmium shield which reduces the flux of epithernial neutrons from the moderator to the proportional counter by roughLy a factor of twenty; and c) the least efficient detector which is depleted in B has a low
BFT pressure, is wrapped in Cadmium and has a lower efficiency by another factor af twenty.
On TFTR (Fig. 3), the dynamic range is handled by an array of fission fragment detectors /10/ where a large spread in detector efficiency is created by the amount and isotope of uranium used. At high count rates, pulse counting cannot be used and the detectors are operated in current mode. In this situation it is difficult to check the pulse height spectrum to ensure that the counters are counting neutrons. Instead, it is usually sufficient to monitor whether the ratio of count rates of the various detectors is fixed at the value expected for neutrons since X-ray and electromagnetic noise tend to influence all detectors about equally.
The frequency response of moderated detectors is limited to about 0.2 msec which is the transit time of the thermal neutron from the moderator into the proportional counter. The purpose of the Cd shield external to he moderator is to eliminate a slower signal component due to neutrons thermalized by tokamak components. In practice, the actual frequency response is usually limited by counting statistics. 1.0 msec counting bins for TFTR and 10 msec counting bins for PLT give useful time evolution of the ion temperature.
4 3.0 CALIBRATION PROCESS
Since the moderated neutron counters are insensitive to incident neutron
energy and since a potentially large fraction of the flux at the detector is
due to scattered neutrons, an absolute calibration is needed to relate the
detector count rate to the neutron source strength of the plasma. The most
usual technique is to place a neutron-emitting radioactive source into the
vacuum vessel during a machine opening. The radioactive source is moved to
locations inside the vessel to approximate the distributed plasma source. If
the source has a known emission level and the detector count rate is measured,
then an absolute calibration is obtained.
An example of PLT calibration data tl\t is shown in Fig. 4 where a PuBe
source was moved toroidally around the device on the minor axis. The detector v>as a moderated He proportional counter located about 3 m from the vacuum
vessel. There is a sizeable reduction in the neutron count rate compared to an unobstructed source (solid line in Fig, 4) due primarily to the toroidal
field coils at 70" and 110" and also due to the PLT center column at 270°.
The accuracy quoted in the literature for this type of calibration differs widely between different experiments /7,12/. The uncertainty in this calibration is principally that the neutron energy spectrum of the radioactive source is different from the fusion neutron spectrum. This implies that the neutron scattering and thus the fraction of scattered flux will differ between the calibration and the actual plasma operation. TFTR work /13/ with different neutron sources indicates that this effect may be small (~ 10%) for detectors located behind a considerable amount of material (so that the
5 detector flux is largely composed of multiply scattered neutrons). The
overall accuracy is probably in the range of 10 - 20% with some further
(approximately 32) uncertainty due to the low flux emitted by radioactive
sources, which means that only the most efficient detector is calibrated. For
this reason, each less efficient detector is cross calibrated with the most efficient detector during plasma operation (Fig. 3).
An independent absolute calibration can be obtained using neutron activation /1A,15/. The TFTR system (Fig. 5) pneumatically transfers activation foils from the tokamak vacuum vessel to a counting room some 200 m away. The basic principle is to determine the virgin flux of neutrons at the foil location from the measured activation level, and then the virgin flux can be related geometrically to the total emission level of the plasma. The activation sample is located right on the vacuum vessel, which is a favorable location for reducing the relative fluence of scattered neutrons. Further, threshold activation reactions are used to reduce and quantify the scattered component. For the case of 14 MeV d(t,n)ot neutrons, a large number of threshold reactions are available and a good determination of the virgin flux is possible /15/. For the case of 2.45 MeV d(d,n) He neutrons, only a small number of threshold reactions are available and principally the In(n.n') reaction has been used (Fig. 5). For PLT, the total uncertainty is in the range of 20 * 40% /14/, where the scattered flux is determined by measuring the activating neutron reflection coefficients for the PLT machine components and integrating over the whole volume. PLT construction was favourable for such experiments since the coils were encased in epoxy (i.e. hydrogen) so that backscatterd 2.5 MeV neutrons are unable to activate the Indium.
6 4.0 ION TEMPERATURE DETERMINATION
The computation of the ion temperature from the total neutron yield requires integration of the Maxwellian deuteron distribution with the energy- dependent d(d,n) He fusion cross section to obtain the fusion emissivity of
the plasma. Most of the reactions occur between those ions in the Maxwallian
travelling in opposite directions and having a relative energy near the Gamow peak (E = 6.3 T- ' ). The d(d,n) He fusion cross section has been measured extensively as reported in the literature lib I and is known to an accuracy of about 2Z down to a deuteron energy of 20 keV and not measured accurately below
20 keV. The accuracy of the fusion emissivity as a function of temperature
{Fig. 6) for temperatures above about 3 keV is better than 5% where there is additional uncertainty introduced by use of analytic forms for the Maxwellian reactivity /17/. For T- < 3 keV, the emissivity is based upon extrapolation of the cross section below 20 keV. The literature quotes several results with stated accuracies much less than the differences between these different results (Fig. 6). Probably the curve by Peres /18/ is the most accurate, and if we take the uncertainty to be the difference between that work and the recent work of Hively /17/ (- 50%), we see that the deduced ion temperature is uncertain by only - 52 due to the steep temperature dependence. If the absolute neutron calibration is in the range of 20 to 302, as discussed above, then the calibration causes only a few percent uncertainty in T- at 1 keV
(Fig. 6).
The largest uncertainty in the neutron ion temperature arises from the uncertainty in the profile shapes and tho deuterium density. Usually the electron density is measured (or assumed to be nearly parabolic) and the deuteron density is calculated as a fraction of the electron density based
7 upon Z r£ and impurity measurements. The uncertainties in these values can
lead to 102 uncertainties in che deduced T^. The ion temperature profile is
also uncertain and is often assumed or calculated on the basis of neoclassical
transport. A PLT neutron emission profile is calculated in Fig. 7 Ibl on the
basis of the Thomson scattering n (r) and the passive charge-exchange ion
temperature profile. Clearly, the ion temperature deduced from the neutron
emission is roughly the volume-weighted ion temperature of the plasma region
inside r/a < \. The uncertainties in the profiles, deuterium depletion, cross
section, and absolute calibration cause the neutron-deduced ion temperature to
be uncertain by 10-202. An important problem arises during hydrogen neutral
beam heating of deuterium plasmas uhere hydrogen from the beams can further
increase the deuterium depletion. On TFTR, a spectroscopic estimate was made
of the hydrogen/deuterium ratio so that the neutron ion temperature remained
in good agreement with other temperature measurements (Fig. 8) /10/. However,
the total uncertainty is about s 20% for these cases.
A serious source of error in the deduction of an ion temperature from the neutron emission can be the assumption of a. Maxwellian deuterium population.
If any deuterons exist in a high energy tail, they can significantly increase the neutron emission without increasing the average ion energy. Although such a situation happened unexpectedly on £eta /19/, this has not happened on tokamaks except when an energetic deuteron tail was purposely created by auxiliary heating (such as during deuterium neutral beam injection). On tokamaks, the ion temperature is usually measured by the charge exchange diagnostic, and in all published cases where a Maxwellian deuteron population is expected, there has been good agreement between the neutron and charge- exchange ion temperature (Fig. 9). Figure S is an updated version of Fig. 6 from Ref. 2 which had included data up to mid-1982. In Ref. 2 it was
8 concluded that "The good agreement is really an indication of the fact that
neutron-deduced ion temperatures are often not quoted unless they are
substantiated by charge-exchange measurements. The value of che neutron
measurement as an ion temperature diagnostic Is primarily, then, as an
independent check on the charge exchange diagnostic which can often have its
own diagnostic complications."
The situation since 1982 has been different and often charge-exchange ion
temperatures have not been quoted. On machines such as TFTR and JET, as well as D-III at high densities, the charge-exchange neutrals reionize before escaping the plasma. In these cases, the neutron emission is used routinely to determine the central ion temperature and on TFTR, energy balance calculations often model the neutron emission directly. In this case, the neutron ion temperature is checked by comparison with impurity ion temperatures determined by spectroscopic Doppler broadening (Fig. 8). Figure
10 shows a comparison of the impurity ion temperature with the neutron ion temperature for TFTR ohmically heated plasmas /39/ and in all cases, these temperatures agree reasonably. In several experiments, the central ion temperature has also been measured by neutron spectroscopy of the Doppler broadening of the 2.5 HeV neutron emission /25,39,42-45/. In most published cases, the two neutron temperature determinations (Fig. 11) are in reasonable agreement. An exception to this situation is the ASDEX work /45/ which claimed a nonthermal neutron spectrum during H° + D neutral beam beating. It is now thought that the instrumental broadening /46/ was larger than initially expected and Doppler broadening of the charged fusion products as well as charge exchange measurements have not substantiated the existence of any non thermonuclear component 747/.
9 5.0 SUMMARY
An important use of fusion product diagnostics is the measurement or the magnitude of the 2.45 MeV neutron emission from the d(d,n) He fusion reaction. For essentially all published data, the deduced ion temperature is in good agreement with the ion temperature deduced from other diagnostics.
The principal difficulty is knowledge of the ion temperature profile and the deuterium depletion by impurities and hydrogen.
10 ACKNOHLEDGMtb'TS
The authors thank K.M. 7oung snd tiia TFTR Diagnostic Croup for their support. This work was supported by U.S. DoE Contracr R<>. Df,-ACO7D-CH0-
3073. REFERENCES
III ARTSIMOVICH, L.A., et al., Sov. Phys. JETP 34, (1972) 306,
111 STRACHAN, J.D., Proceedings of the Course Diagnostics For Fusion Reactor
Conditions, Varenna, Italy, Vol. _1, (1982) 383.
Ill HENDEL, H.W., Proceedings of the Course Diagnostics For Fusion Reactor
Conditions, Varenna, Italy, Vol. 1, (1982) 327.
/4/ JARVIS, O.N., Proceedings of Che Course Diagnostics For Fusion Reactor
Conditions, Varenna, Italy, Vol. l_, (1982) 383.
15/ KITSUNEZAKI, A., et al., Conf. on PI. Phys. and Contr. Nucl. Fus. Res..
London, UK, Vol. I, (1984) 57.
/6/ STRACHAN, J.?., PPPL-TM-303 (1977).
Ill STRACHAN, J.D., ec al., Nucl. fus. 21, (1981) 67.
/8/ KNOLL, C.F., "Radiation Detection and Measurement," (Wiley, New York,
1979).
I9l MARION, J.B. and FOWLER, J.L. (ed.), "Fast Neutron Physics, Part X
Techniques," (Interscience, NY, I960).
/10/ HENDEL, H.W., IEEE Trans, on Nucl. Sci. 33, (1985) 670.
/Ill STRACHAN, J.D. and ZANKL, C, PPPL-TM-353 (1981).
/12/ PAPPAS, D., PARKER R.R., PFC/RR-78-5 (MIT, 1978).
/13/ NIESCHMIDT, E.B., et al., Rev. Sci. Instr. 56, (1985) 1084.
/14/ ZANKL, G., et al., Nucl. Inatr. and Meth. 185, (1981) 325.
/15/ NIESCHMIDT, E.B., Rev. Sci. Instr. 5_7, (1986) 1757.
/16/ JARMIE, N., and BROWN, R.E., Nucl. Instr. and Meth. BlQ/11, (1985) 405.
/17/ HIVELY, L.M., Nucl. Fus. 12, (1977) 873; Nucl. TechnjFuS. 3_, (1983) 199.
/18/ PERES, A., J. Appl. Phys. 50, (1979) 5569.
/19/ ROSE, B., TAYLOR, A.E., and WOOD, E., Nature jUU, (1958) 1630.
12 /2Ql STODIEK, U., et al., Plasma Phys. and Contr. Nucl. Fus. Research Vienna,
Austria, Vol. 1, (1981) 465.
/21/ Equipe TFR, Plasma Phys. and Contr. Nucl. Fus. Res., Vienna, Austria,
Vol. II, (1981) 75.
/22/ Equipe TFR, Plasma Phys. and Contr. Nucl. Fus. Res., Vienna, Austria, Vol
n, (1981) 547.
/23/ Equipe TFR, Plasma Phys. 24, (1982) 615.
/24/ HOSEA,J.C, et al.t Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol 1^, (1981) 95.
/25/ BREJSATI, M., et al., Nucl. Fus. 18, (1978) 1205.
/26/ EUBANK, H.P., et al., Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. J_, (1979) 167.
/27/ HOSEA, J.C., et al•, Heating in Toroidal Plasmas, Grenoble, France, Vol
I, (1982) 213.
/28/ MEADE, D.M., et al., Plasma Phys. ani Contr. Nucl. Fua. Res., Vienna,
Austria, Vol. I_, (1981) 665.
/29/ SWAIN, D.W., et al., Nucl. Fus. 21, (1981) 1409.
/30/ GONDHALEKAR, A., at al., Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. I, (1979) 199.
/31/ BAGDASAROR, A.A., et al., Plasma Phys. and Contr. Nucl. Fus. Res.,
Vienna, Austria, Vol. 1^, (1979) 35.
/32/ BESLIZOV, A., et al., Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. I, (1976) 3.
/33/ BERRY, L.A., et al., Plama Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. I, (1976) 49.
/34/ MURAKAMI, M., ^»t al., Phys. Rev. Lett. 42, (1979) 655.
13 /35/ MORI, M., et al.t Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. I, (1984) 443.
/36/ MAZZUCATO, E., et al.t Plasma Phys. and Contr. Nucl. Fus. Res,, Vienna,
Austria, Vol. I, (1984) 433.
/37/ PORKOLAB, M., et al., Plasma Phys. and Ccntr. Nucl. Fus. Res., Vienna,
Austria, Vol. £, (1984) 463.
/38/ WAGNER, F., et al., Plasma Phys. and Contr. Nucl. Fus. Res., Vienna,
Austria, Vol. I, (1984) 45.
/39/ CORTI, S., et al •, Proc. European Conf. on Plasma Phys. and Contr. Fus.,
(Budapest, Hungary, 1985) paper 170.
/40/ BITTER, M., and MEDLEY, S.S., compiled the Doppler broadening
temperatures for TFTR.
Ml/ FISHER, W.A., CHEN, S.H., GWINN, D., and PARKER, R.R., Nucl. Instr. and
Meth. in Phys. Res., 219, (1984) 179.
/42/ FOLEY, J., M. Sc. Thesis Princeton University (1985).
/43/ JARVIS, O.N., et al •, 12th European Conf. on Contr-. Fus. and Plasma
Phys., Pt. 1 (Budapest, 1985) 223.
/44/ DRESSING, G., et al., Lecture at the Royal Society Discussion Meeting
(London, UK, 1986).
/45/ STEINMETZ, K., et al., 12th European Conf. on Contr. Fus. and Plasma
Phys., Pt. 1 (Aachen, 1983) 51.
746/ HUBNER, K., Private Communication.
/47/ BOSCH, S., in Proceedings of Workshop on Diagnostics for Fusion Plasmas,
(Varenna, Italy, 1986) to be published.
14 FIGURE CAPTIONS
Figure 1 d(d,n) He fusion reaction rate from TFTR experiments in deuterium
(solid circles) and projected d(t,n)a fusion reaction rates
(dashed circles) as a function of the power used to heat the
plasma.
Figure 2 Schematic diagrams of the PLT epitheriral neutron detector system
/6,7/. The detectors consist of a lead shield to avoid hard X
ray counts, an external Cd shield to avoid counts from neutrons
thermalized in machine components, a polyethylene moderator; an
array of epithermal neutron counts, and associated pulse counting
electronics.
Figure 3 Counting ability of the TFTR array of fission fragment o detectors. The local flux range at the detectors of 10 * 1 ? —? —1 • 9
10 cm s corresponds to the TFTR emission levels of lCr •*
1019s-1 (Fig. 1) for a detector set about 5 m from the TFTR
vessel.
Figure 4 PLT calibration obtained by moving a PuBe source toroidally
around the vessel on the minor axis and monitoring the count rate
of a fixed moderated neutron detector III!.
Figure 5 Schematic diagram of the TFTR pneumatic "rabbit" system for
neutron activation of foils. The 2.5 MeV neutron flux
measurement will be based on Indium foils. The insets illustrate
15 the neutron activation of Indium and the subsequent 4.5 hr In
decay with PLT data /14/.
Figure 6 Fusion reactivity of the d(d,n) He reaction for Maxwellian
deutevon populations of temperature T-. The crosses are the fit
from Ref. 17, the squares are from Ref. 18.
Figure 7 Te(r), T^r), ne(r) fron a PLT ohmic plasma /6/ with the
calculated radial dependence of the neutron emission including
the uncertainty.
Figure S Time evolution of the TFTR neutron emission and deduced ion
temperature for a 4.2 MW H° •• D+ neutral-beara-heated plasma /10/
and compared to the titanium k Doppler broadening and charge
exchange measurements.
Figure 9 Comparison of the neutron-deduced ion temperature with the charge
exchange ion temperature for all tokamak experiments where both
measurements were published (1,2,10,20-40/ and where the
deuterium ions are expected to be Maxwellian (such as ohmic
heating, H° + D+ neutral beam heating, and 3He or H minority ICRF
heating).
Figure 10 Comparison of the impurity ion temperature with the neutron ion
temperature for TFTR ohmically heated plasmas /41/.
16 Figure 11 Comparison of the ion temperature as determined by the magnitude
of the neutron emission with the ion temperature as determined by
neutron spectroscopy for all published data /25,39,42-45/.
17 in20 #86X1238 IV I 1
I0'9
/ I0'8 - /D°-T+/ / NBI / I0!7 / / / / ,' * / / * I016 — / * V " Jm 1015 _ MNBI^A-Df ^
I011
/ / I0'3 _ / Ohmfc /f\ —
jnl2 u i // 1 ~ ^yI
0'° - / / -
in9 1 1 0.1 1.0 10.0 100.0 INPUT POWER (MW)
Fig. 1
13 BF, PulM Htight Sptetrum 18% fl", 10 cm Hg
HT 1 + 600 N PulM BFj Caunltr Htight (187. B10,10cm Hg) / 109 PC 452 Amp 441SCA Prtomp T, = 2JIJ
109 PC 452 Amp 441 SCA Prtamp
109 PC 452 Amp 441 SCA Prtamp
Polyethylene' % B^O em Hg) (96 \^ BFj PulJt Hilght Spectrum HT ^ 96 V. B^Ocm Hg + I300N
PulM Htight'
Pig. 2
19 z NEUTRON FLUX In(n/cm s)
ha to o
NEUTRON SOURCE STRENGTH , n/s # 81X0472 I03
Expected Variation For An Unobstructed Source
O O O cr LJ
J_ l 90° 180° 270° 360° TOROIDAL ANGLE OF PuBe SOURCE
Fig. 4
21 #86X1239
Sample
Irradiation End 115 115m 3 In + n — In + n' s •Pneumatic Transfer Tube (200 m in 20sec)
Ge (Li) Gamma-Ray NEUTRON ENERGY (MeV) Counting Station
15m. y ||5r In —- In
200 400 600 GAMMA ENERGY (keV)
Fig. 5
22 (T_D3S £.W3) A1IAU0V3H NO I Slid
23 #66X1300 THOMSON SCATT CHARGE EXCHANGE Zeff*4-6
I 1.0 'IIH .5-
10 20 30 40 0T 10 20 30 40 r(cm) r(cm) -jo5 I til £
^ 2 ,n I0Z LU s
10 10 20 30 40 10 20 30 40 r(cm) r(cm)
Pig. 7
24 #95X1587
SHOT * 12475
O
x 15
ar. h-
LU o cc o
fTini AChorge Exchange SHOT #12475
NSK4.2MW) J L TIME (sec) Fig. 8 25 #86X1267 - 1 i i i i i i i 1 i i i i i i r _ - - 0 Ohmic - 1 I x ICRF y i • NBi TFTRfW "-LT " pa iir - JETtXj>^ -r ^r" ASOEX PLTi/ A. TFTR(>XTFR A^J^TFR TFR X^ORMAK — PLTjXp- "T-IO _ jQ •~'iSX-B _ ALC-Ag/^TFR PDXQ^C^T-ro mm _ ST3T - - i i i i i i i 11 i i i 10 CHARGE EXCHANGE ION TEMPERATURE (keV) Pig. 9 26 #86X1240 NEUTRON ION TEMPERATURE (keV) Fig. 10 27 #86X1242 oASOEX h* JET • PLT oTFTR A ALCATOR ^ 3.0- Ti(0) NEUTRON FLUX (keV) Fig. 11 28 EXTERNAL DISTRIBUTION IN ADDITION TO UC-ZO Plasma Has Lab, Austra NaT') Unlv, AUSTRALIA aibllotheek, Fom-lnst Voor Plasma, NETHERLANDS Or. Frank J. Psolonl, Unlv of Wollongong, AUSTRALIA Prof. B.S, Llley, University of Walkoto, NEW ZEALANO Prof. I.R. Jonas, Flinders Univ., AUSTRALIA Prof. J.A.C. Cabral, Inst Superior Teen, PORTUGAL Prof. M.H. Brennan, Unk Sydney, AUSTRALIA Or. Oetavlan Petrus, ALI CUZA University, ROMANIA Prof. F. Cap, InsT Theo Phys, AUSTRIA Prof, M.A. Hellberg, University of Natal, SO AFRICA H. Goossens, Astronomlsch Instltuut, BELGIUM Dr. Johan de VII Hers, Plasma Physics, Nucor, SO AFRICA Prof. R. Bouclqua, Laboratorlum voor Natuurkunde, BELGIUM Fusion Dlv. Library, JEN, SPAIN Dr. D. Palumbo, Og XII Fusion Prog, BELGIUM Prof, Hans Wllhelmson, Chalmers Unlv Teen, SWEDEN Ecole Royale Mllltalrs, Lab ds Phys Plasmas, BELGIUM Dr. Lennart Stenflo, University of UMEA, SWEDEN Dr. P.H. Sakanaka, Unlv Estadual, BRAZIL Library, Royal Inst Tech, SWEDEN Lib. I Doc. Olv., Institute de Pasqulsas Espaclala, BRAZIL Centre de Recherchesen, Ecole Polytech Fed, SWITZERLAND Dr. C.R. James, Unlv of Alberta, CANADA Dr. V.T. Tolok, Kharkov Phys Tecn Ins, USSR Prot. J. Telchmann, Unlv of Montreal, CANADA Dr. 0.0. Ryutov, Siberian Acad Sel, USSR Or. H.M. Skarsgard, Unlv of Saskatchewan, CANADA Or. G.A. Ellseev, Kurchatov Institute, USSR Prof. S.R. Sreenlvasan, University of Calgary, CANADA Dr. V.A. Glukhtkh, Inst Electro-Physical, USSR Prof, T.jdor W. Johnston, INRS-Energle, CANADA Institute Gen. Physics, USSR Dr. Hannes Barnard, Unlv British Columbia, CANADA Prof. T.J.M. Boyd, Unlv College N Wales, WALES Dr. M.P, Bachynskl, MPB Tachnologlas, Inc., CANADA Dr. K. Schlndler, Ruhr Unlversltat, W. GERMANY Chalk River, Nuel Lab, CANADA ASDEX Reading Rm, IPP/Max-Planck-lnstltut fur Zhengwu LI, SW Inst Physics, CHINA Plasmaphyslk, F.R.G, Library, Tslng Hua University, CHINA Nuclear Res Estab, Jitllch Ltd, W, GERMANY Librarian, Institute of Physics, CHINA Librarian, Max-Planck Instltut, W. GERMANY Inst Plasma Phys, Acidemia Slnlca, CHINA Blbllothek, Inst Plasmaiorschung, W. GERMANY Dr. Peter Lukac, Komenskeho Unlv, CZECHOSLOVAKIA Prof. R.K. Janev, Inst Phys, YUGOSLAVIA The Librarian, Culham Laboratory, ENGLAND Prof. Schatzman, Observatolre de Nice, FRANCE J. Radet, CEN-BPS, FRANCE JET Reading Room, JET Joint Undertaking, ENGLAND AM Dupas Library, AM Oupas Library, FRANCE Dr. Tom Mual, Academy Bibliographic, HONG KONG Preprint Library, Cent Res Inst Phys, HUNGARY Or. R.K. Chhajlanl, Vlkram Unlv. INDIA Or. B. Oasgupta, Saha Inst, INDIA Dr. P. Kaw, Physical Research Lab, INDIA Dr. Phi I lip Rosenau, Israel Inst Tech, ISRAEL Prof. S. Cuperman, Tel Aviv University, ISRAEL Prof. G. Rostagnl, Univ 01 Padova, ITALY Librarian, Int'l Ctr Theo Phys, ITALY Miss Clella De Palo, Assoc CURATOM-ENEA, ITALY Blbllotece, del CNR EURATCH, ITALY Or. H. Yamato, Toshiba Res & Oev, JAPAN Dlree, Dapt. Lg, Tokamak Dav, JAERI, JAPAN Prof. Nobuyukl Inoue, University of Tokyo, JAPAN Research Info Center, Nagoya University, JAPAN Prof. KyoJI Nlshlkana, Univ of Hiroshima, JAPAN Prot. Slgeru Mor!, JAERI, JAPAN Prot. S. Tanaka, Kyoto University, JAPAN Library, Kyoto University, JAPAN Pro*. Ichiro Kawakaml, Ninon Unlv, JAPAN Prof. Sotosh I Itoh, Kyushu University, JAPAN Or. D.I. Choi, Adv. Inst Scl 4 Teeh, KOREA Tech Info division, KAERI, KOREA