Manual for Troubleshooting and Upgrading of Neutron Generators

IAEA-TECDOC-913

Manual troubleshootingfor and upgrading neutronof generators

INTERNATIONAL ATOMIC ENERGY AGENCY The IAEA does not normally maintain stocks of reports in this series. However, microfiche copie f thesso e reportobtainee b n sca d from

ClearinghousS I N I e International Atomic Energy Agency Wagramerstrasse 5 0 10 P.Ox Bo . A-1400 Vienna, Austria

Orders should be accompanied by prepayment of Austrian Schillings 100, fore for e chequa th f m th IAEf m o n i o n i r eAo microfiche service coupons which may be ordered separately from the I MIS Clearinghouse. The originating Section of this publication in the IAEA was Physics Section International Atomic Energy Agency Wagramerstrasse 5 PO Box 100 A-1400 Vienna, Austria

MANUAL FOR TROUBLESHOOTING AND UPGRADING OF NEUTRON GENERATORS IAEA, VIENNA, 1996 IAEA-TECDOC-913 ISSN 1011-4289 © IAEA, 1996 Printe IAEe th Austrin Ai y d b a November 1996 FOREWORD

During the past 20-25 years the IAEA has provided many new laboratories in the developing world with simpl voltagw elo e accelerator productioe th r sfo f neutronno e th a svi well known 2H(d,n)3He and 3H(d,n)4He reactions. (These neutron generators were originally supplied mainly for purposes of neutron activation analysis). However, the operation of these machines can often be halted or compromised by a lack of special or short lifetime components. Serious problems can also arise when the original well-trained technical staff leav e laboratorieth e operatorw ne d whe e an s th ns have less experienc n acceleratoi e r technology.

This manua s intendei l o assist d t operator troubleshootinn i s d upgradinan g f o g neutron generators directes i t I . d particularl operatoro yt techniciand san lesn si s experienced laboratorie thereford san descriptione eth principlee th f so techniqued san thesf so e machines e operatoar r oriented addition I . discussioa o n t e mai th nf n o characteristic f neutroo s n generators, detailed information is given on the function of particular commercial units, on common problems related to specific components of accelerators, and on methods of troubleshootin repaird gan . Detailed schemati circuid can t diagram providee ar s helo dt p operator developmene th n si improvemend an t generatorse th f o t .

probleme Th s treate Manuae th n di l have been collected during several IAEA missions in developing countries.

IAEe Th gratefuAs i . SztaricskaiT o t l performeo wh , majoe dth draftinre parth f o t g of the manuscript, and also to J. Csikai and S. Szegedi for their contribution to the drafting. The IAEA officer responsible for this publication was R.L. Walsh, Physics Section, Division of Physica Chemicad an l l Sciences. EDITORIAL NOTE

In preparing this publication for press, staff of the IAEA have made up the pages from the original manuscripts submittedas authors.the viewsby The expressed necessarilynot do reflect those of the governments of the nominating Member States or of the nominating organizations. Throughout the text names of Member States are retained as they -were when the text was compiled. The use of particular designations of countries or territories does not imply any judgement by the publisher, the IAEA, as to the legal status of such countries or territories, of their authorities and institutions delimitation ofthe or theirof boundaries. The mention of names of specific companies productsor (whether indicatednot or registered)as does not imply any intention to infringe proprietary rights, nor should it be construed as an endorsement or recommendation on the part of the IAEA. The authors are responsible for having obtained the necessary permission for the IAEA to reproduce, translate or use material from sources already protected by copyrights. CONTENTS

1. INTRODUCTION ...... 9 .

2. PRINCIPLES OF OPERATION ...... 11

3. DETERMINATION OF THE BEAM ENERGY ...... 23

4. TYPES OF NEUTRON GENERATORS ...... 26 4.1. Commerical neutron generators ...... 8 2 . 4.2. Sealed tube neutron generators ...... 34 4.3. Intense neutron generators ...... 8 3 .

SOURCESN IO . 5 : OPERATION PRINCIPLES, MAINTENANCE AND TROUBLESHOOTING ...... 43 5.1. High frequency ion sources ...... 43 5.2. Extraction of ions from ion sources ...... 48 5.3. Maintenanc dischargs ga f eo e pyrex bottle ...... 1 5 . 5.4. High frequency oscillators ...... 52 5.4.1. Troubleshooting of high frequency oscillators ...... 55 5.5. Pennin sourcen gio s ...... 8 5 . 5.5.1. Troubleshooting of Penning ion sources ...... 61

6. DEUTERIUM LEAKS ...... 63 6.1. The palladium leak ...... 63 thermomechanicae 6.2Th . needle lth lead ek an valv e ...... 5 6 . 6.2.1. The thermomechanical leak valve ...... 65 6.2.2. Maintenanc troubleshootind ean f thermomechanicago l leaks ...... 6 6 . 6.2.3. Needle valves ...... 67 6.2.4. Maintenanc f needleo e valves ...... 0 7 . 6.3. Calibratio f leano k valves s consumptioGa : n measurement sourcen io 3 7 f o s s 6.3.1. Measurement ...... 4 7 .

. DEUTERIU7 M ELECTROLYZERS ...... 6 7 . e floa7.1Th t . regulator electrolyzer ...... 0 8 .

8. REMOTE CONTROL OF THE HIGH VOLTAGE TERMINAL ...... 84 8.1. Mechanical control ...... 4 8 . 8.2. Electromechanical control ...... 5 8 . 8.3. Insulation transformer control ...... 6 8 . 8.4. Optical insulation control ...... 88 8.5. Computer control ...... 88

9. VACUUM SYSTEMS OF NEUTRON GENERATORS ...... 93 9.1. Important terms and units in vacuum technology ...... 93 9.1.1. Terms ...... 93 9.1.2. Units ...... 99 9.2. Vacuum pumps ...... 100 9.2.1. Vacuum system base combinatioa n do f no oil diffusion and rotary pumps ...... 102 (a) The rotary vane pump ...... 102 diffusioe Th ) (b n pump ...... 4 10 . (c) Combination of diffusion and rotary pump ...... 106 9.2.2. Vacuum system based on Ti-ion getter pump ...... Ill 9.2.3. Vacuum system base turbomoleculan do r pump ...... 3 11 . 9.3. Pressure (vacuum) measurements ...... 115 9.3.1. Thermal conductivity gauges ...... 5 11 . 9.3.2. lonization gauges ...... 116 (a) Thermionic ionization gauges ...... 116 (b) Cold cathode or Penning ionization gauge ...... 117 9.4. Pressure monitoring and leak detection ...... 118 9.4.1. Leak rate measurement ...... 8 11 . 9.4.2. Pumping speed measurement ...... 119 9.4.3. Leak detection ...... 3 12 .

10. BEAM ACCELERATION AND BEAM TRANSPORT SYSTEMS ...... 128 10.1. Electrostatic lens ...... 128 10.2. Unipotentia r Einzeo l l lens ...... 9 12 . 10.3. Troubleshooting of electrostatic focus lenses ...... 130 e acceleratio10.4Th . n tube ...... 1 13 . 10.5. Troubleshooting of acceleration tubes ...... 134

11. PRINCIPLE BEAF SO M FILTERS ...... 5 13 . 11.1. Electrostatic and magnetic beam deflection ...... 135 11.2. Troubleshootin f electrostatigo c deflectors ...... 7 13 . 11.3. Analyzing magnet f neutroo s n generators ...... 8 13 . 11.4. Vacuum chambers of deflecting magnets ...... 144 11.5. Problems with analyzing magnets ...... 5 14 .

12. QUADRUPOLE LENSES ...... 146 12.1e biaseTh . d quadrupole lens ...... 8 14 . 12.2biasee Th . d magnetic quadrupole doublet ...... 0 15 . 12.3. Troubleshootin magnetia f go c quadrupole lens ...... 4 15 .

13. HIGH VOLTAGE POWER SUPPLIES ...... 155 13.1. Electrostatic (Felici) high voltage generator ...... 5 15 . 13.2. AC-DC conversion high voltage power supplies ...... 9 15 . 13.2.1. The single phase half wave rectifier ...... 159 13.2.2. Cascade generators ...... 161 13.2.3. Improved cascade circuits ...... 4 16 . 13.3. Troublehsooting of high voltage power supplies ...... 168

14. BEAM LINE COMPONENTS ...... 3 17 . 14.1. Beam stops ...... 173 14.2. Beam scanners ...... 3 17 . 14.3. Wire electrode (matrix) beam scanners ...... 178 14.4. The Faraday cup ...... 179 14.4.1. Beam current integration ...... 2 18 . 14.5. Target assemblies ...... 182 14.5.1. Target replacement ...... 5 18 . 14.5.2. Air cooled target holder ...... 188 14.5.3. Replacement of the target at air cooled target holders ...... 188 14.5.4. Rotating and wobbling target holders ...... 191

15. CLOSED CIRCUIT COOLING SYSTEMS ...... 2 19 . 15.1. The Kaman cooling system ...... 192 15.1.1. Maintenance ...... 193 15.2. Closed circuit cooling system with soil heat exchanger ...... 5 19 .

16. PNEUMATIC SAMPLE TRANSFER SYSTEMS ...... 197

17. NANOSECOND PULSED NEUTRON GENERATORS ...... 0 20 . 17.1. Pre-acceleration nanosecond bunche bean dio m neutron generator ...... 0 20 . 17.2. Post-acceleration klystron bunching of a commercial neutron generator ... 204

18. THE ASSOCIATED PARTICLE METHOD ...... 206 18.1. Self-target formation by deuteron drive-in ...... 208

19. NEUTRON MONITORS ...... 1 21 . 19.1. Monitoring by long counter ...... 211 19.2. Fission chamber monitoring ...... 2 21 .

. SAFETY20 : HAZARDS RELATE NEUTROO DT N GENERATORS ...... 5 21 . 20.1. Radiation hazard ...... 5 21 . 20.2. Radioactive material storage and waste disposal hazard ...... 219 20.3. High voltage hazard ...... 219 20.4. Implosion hazard ...... 220 20.5. Pressure hazard ...... 0 22 . 20.6. Fire hazard ...... 220

21. CONSTRUCTION OF A NEUTRON GENERATOR LABORATORY ...... 221 21.1. Construction details ...... 1 22 . 21.2. Workshops ...... 223 21.3. Laboratory log book ...... 224

REFERENCES ...... 227

ANNEX A: LIST OF MANUFACTURERS AND COMPONENT DEALERS ...... 233

ANNEX B: TROUBLESHOOTING FLOW CHART FOR NEUTRON GENERATORS WITH RF ION SOURCE ...... 241 ANNEX C: TROUBLESHOOTING FLOW CHART FOR SEALED TUBE NEUTRON GENERATORS ...... 3 24 .

ANNEX D: TROUBLESHOOTING FLOW CHART FOR NEUTRON GENERATOR VACUUM SYSTEM ...... 5 24 .

CONTRIBUTORS TO DRAFTING AND REVIEW ...... 247 1. INTRODUCTION

Neutron generators are small accelerators consisting of vacuum, magnetic, electrica d mechanicaan l l components, radiation sources, cooling circuitd an s pneumatic transfer systems. Ther e variouar e sn sources io type f o s , beam acceler- atind transporan g t systems, targets, high voltag d othean e r power supplies, neutron and tritium monitors and shielding arrangements. e operationTh , maintenanc d troubleshootinan e a neutro f o g n generator require well trained technicians who can successfully undertake not only preventive maintenance of the machine but also its upgrading. Troubleshooting and the locating of faults in components can be just as difficul s a theit r prevention. Thoughtless component exchang n a accelerato n i e r may cause additional problems e correcth : t choic f componeno e a give r fo nt func- e machinth tio n i n e requires complete understandin e operatioth e genf th o g f -o n erator as well as the role of the component in the machine. n I troubleshootin r neutrofo g n generators s a wit, h other sophisticated equipment, it may be that the cause of malfunction stems from a single fault and than investigatioa t e wholth f o en syste s mboti h time consumin d unnecessaryan g . However n electricaa , l e higfailurth hn i evoltag e power suppla discharg r o y n i e the accelerating tube coul e causeb d y e troublhigb dth hn i evacuu m systemn A . experienced troubleshooter can avoid unnecessary investigation of a number of subunits. This Manual has been prepared not only for operators but also for those who are dealing experimentally with the upgrading of neutron generator laboratories. Principles of operation and output characteristics of neutron generators can be found in Refs [1-3]. This Manual is restricted to practical information required e operatoth by d servican r e personnel e - functione Th subunitde th . f e o ar ss scribed and proposals for troubleshooting in case of malfunction of a component are outlined. The Manual contains descriptions of some symptoms and proposals for their repair e proposeTh . d repai d maintenancan r e methods have been chosen bearing in mind that some laboratories may have poor infrastructure. The Manual contains a short description of the commercially available neu Iron generators, including the sealed tube and intense models. The description of the operation of a specific component is usually followed by a general guide to troubleshooting and repair. As this is NOT A SERVICE MANUAL, the guidelines for troubleshooting and maintenance are not given for all types of neutron generator, althoug e informatioth h voltagw lo y typ f nan o e e use b r hereiaccelera fo dn ca n - tor. The detailed description of some methods, such as pulsing or the associated particle technique, is intended to help in improving and widening the applica- e th origination f o s l machine e f lisTh o . tmanufacturer d othean s r sourcef o s component n Annei s si intende A x o assist d t laboratorie n developini s g countries o fint d supplier r sparfo s e part r improvedo s components e shorTh . t e revieth f o w hazards related to accelerators is particularly directed towards operators who have less experience in these matters and should be read with care by them. The detailed schemati d workshoan c p drawings shoul e usefub d n i upgradinl g neutron generator d acceleratoran s s wela experimentssw settin n i lne p gu .

10 2. PRINCIPLES OF OPERATION

) neutrokV 0 n10 generatorw voltagw fe lo a e ( e Th s produce neutrone th y b s following reactions:

2H(d,n)3He Q= 3.268 MeV (1)

3H(d,n)4He Q = 17.588 MeV (2)

The large cross section of the 3H(d,n ) 4He reaction permits high yields of fast neutrons to be obtained even at low deuteron energy (150-200 keV). The 0° differential 0 40 6 crosmb/s 2. d san e r ar section) (2 f d o sreaction an ) (1 s mb/sr, respectively e totaTh . l cross e sectioth H(d,nf o ne H )reactio a s ha n 3 4 broad resonance with a maximum value of 5 barns at E, = 107 keV. At this deuteron energy the total cross section of the H(d,n) He reaction is about 20 mb; therefore, the contribution of the D-D background neutrons to those emitted in T cros reactioT e D- neglectedthb s D- e n sectioa th peaca nt s s arouna kA ha n. d e 10th tritiu 7d keVan m , target e thicar s k metal titrides e neutroth , n yield show n increasina s bombardinV g ke o abou t tren0 p 40 u dt g energy (see Fig.l)e Th . D-D reaction is used for neutron production mainly by electrostatic accelera- torr cyclotronso s , wher e neutroth e n energ s i changey y changinb d e deuteroth g n

100 200 300 DEUTERON ENERGY (keV) Fig.l Neutron production cross section and the total yield of neutrons for a TijT, , thick target vs deuteron energy

11 i r 150- • Meos by Zr/ Nb '—£ Calc — Calc— .

- tt.5- 0.15

Ed(keV] 14.0

0.1

13.5-

30 60 90 120 150 180 EMISSION ANGLE Fig.2 Thick target neutron energy vs emission angle at different bombarding deuteron energies. The angular dependence of the neutron energy spread relates to E, = 150 keV. The energy broadening of D-T plasma neutrons at also[6],is keV shown10 T.=

energy. However, monoenergetic neutrons can be produced only at Ed < 4.45 MeV, below the deuteron break-up threshold. e e emitteenergth Th f o yd neutron ) depende (2 th n reactiond i sn o an s ) (1 s bombarding deuteron energy and the emission angle of neutrons to the direction of the deuteron beam.

e thiTh n target data recommende e angulath r fo r ] (Y [4 dd energ n)an y (En) distribution T reactione - welb D- ap ln d f ca an neutrono ss D sD- emittee th n i d proximated by the following expressions [1, 6] in laboratory system:

Yn = (3)

(4) reaT D- c r fo 2 d an 3 whilD = D- n e r fo 3 d an 5 d (4)= an , n ) (3 s eq n I tionse evaluateTh . r thicdfo dat] k [5 atarget ) n als(3 e ca describe b sos eq y b d

and (4). The values of the Y , Y^ EQ and Ej coefficients from a least-squares fit are given in Tables 1-5 for the 20-500 keV deuteron energy range [1,6]. In

12 ) hav(4 e d beean n T ) reaction checke(3 D- s e y eq th experimene b ,casd f th o e t using a point source in a scattering-free arrangement [1] and the Zr/Nb, Zr/Au,

Zr/Ta activity ratio method [7,8]. The E(E,0) functions at E = 50, 150 and d d

300 keV for a thick tritium target are shown in Fig.2.

Tabl. 1 e Recommended parameter value calculatior fo s f thino n target angular distributions of D-T source yields in laboratory system (equatio ) normalizen3 ° 90 o dt o (90°) Ed(keV) A A (mb/sr) o l *2 [Ref.4] 20 1 0.0220 0.00025 4.2942 30 1 0.0227 -0.0093 19.6126 40 1 0.0310 0.0007 52.8382 50 1 0.0344 0.0010 105.4180 60 1 0.0518 -0.0035 173.2630 70 1 0.0407 0.0011 249.3768 100 1 0.0482 0.0011 393.3834 150 1 0.0599 0.0009 316.3704 200 1 0.0678 0.0005 198.4180 250 1 0.0685 -0.0104 132.8133 300 1 0.0818 0.0005 95.0878 350 1 0.0904 0.0028 77.3864 400 1 0.1003 -0.0008 63.4112 450 1 0.1140 -0.0101 53.0290 500 1 0.1273 -0.0187 45.5970

Data obtained from the 89Zrr //92m Nb above activity ratio [6] produced in m (n,2n) reaction e analyticasth prov e f possiblo th e l e expressionsus e e calTh -. culated energy spread (6E ) [6] of neutrons (1/2 FWHM) as a function of emission angle for a thick TiT target is also shown in Fig.2 at E j = 150 keV. Typical e distributionshapeth f o s ] calculatee [9 sth Mont y b ed Carlo simulation code PROFIL [10] are shown in Fig.3.

e neutroTh n energy spread e th sfollowin refeo t gr circumstances: E, = (190±10)keV, point-like beam spot, D + analyzed beam, 11.5 cm sample-target distance, and 8 mm x 16 mm sample dimension. Recently, Kudo and Kinoshito [11] have develope a method de pulsth base en o dheigh t distributio e recoith f o ln edge for 4He nuclei in a 3He proportional counter for the determination of the energy

13 Table 2. Coefficient f Legendro s e polynominal calculatior fo s f thino n target, angular distribution sourcD D- f eo syield laboratorn i s y system (equatio ) normalizen3 ° 90 o dt

E a (90°) d A A A4 A5 (mb/sr) (keV) o l *2 *3 [Ref.4]

50 I 0.11787 0.58355 -0.11353 0.04222 0.16359 0.32016 100 1 0.01741 0.88746 0.22497 0.08183 0.37225 1.01828 200 1 -0.03149 1.11225 0.38659 0.26676 0.11518 1.95031 300 1 -0.10702 1.64553 0.63645 0.67655 0.35367 2.66479 400 1 0.02546 1.05439 0.21072 0.81789 0.59571 3.32222 500 1 -0.10272 1.09948 0.29820 1.09435 0.76159 3.63084

Tabl. 3 e Coefficient T e calculation equatioi th sD- r e fo th 3 nf no thick target neutron yiel s v demissio n angle

D-T fi.F j ————————————————————— (keV) Ao Al A2

1 0.030050 3 0.00035 100 1 0.04087 0.00062 1 150.04720 7 0.00083 1 200.05120 4 0.00096 250 1 0.05419 0.00110 300 1 0.05651 0.00119 1 320.05615 6 0.00119

e meaT th neutronssprea D- nd e e an denergdetecto th Th . f o ys calibratei r e th t a d reference energy poinf 14.0o obtaineV t 0Me t a d96°-98 ° emission anglese Th . energy V obtainespreake f 0 neutrono 1 t da thes d ^ E e <5 sangle o to s e seemb o t s small. e followinTh g variable n stronglca s y influenc e neutroth e n energy distribu- e typeth f targeo ) tions1 :e t rati th f atomsatomi o o) 2 o ,t cmolecula r ions,

14 Table 4. Value f parametero s e calculatioequation i sth r fo n4 f thi no n target neutron energy vs emission angle in laboratory system E d T - D D - D (keV) Eo El E2 Eo El E2 E3

50 14.04814 0.47679 0.00834 2.46073 0.24848 0.01282 0.00031 100 14.06732 0.67488 0.01719 2.47303 0.35237 0.02524 0.00062 200 14.10711 0.95596 0.03320 2.49771 0.50072 0.05044 0.00242 300 14.14704 1.17282 0.04923 2.52289 0.61581 0.07530 0.00589 400 14.18670 1.35640 0.06527 2.54798 0.71456 0.10013 0.00757 500 14.22569 1.51899 0.08249 2.57246 0.80285 0.12592 0.01024

Table 5. Values of parameters in equation 4 for the calculation of thick target neutron energy vs emission angl laboratorn i e y system E d D - T D - D (keV) Eo El E2 Eo El E2 E3

50 14.06520 0.42329 0.00682 _ . _ - 100 14.07883 0.57613 0.01222 2.46674 0.30083 0.01368 - 150 14.08942 0.66776 0.01600 - - - - 200 14.09680 0.72427 0.01908 2.47685 0.39111 0.04098 0.02749 250 14.10286 0.76661 0.02167 - - - - 300 14.10803 0.80001 0.02374 2.49712 0.47697 0.05124 0.02957 325 14.10723 0.79477 0.02347 - - - - 400 - - - 2.50981 0.55825 0.07125 0.02474 500 _ _ . 2.52140 0.62147 0.09816 0.03307

15 I I———I—I—I—

13.2 13.*. 13.6 13.8 14.0 14.2 14.4 14.6 14.8 15.0 15.2 NEUTRON ENERGY [ MeV] Fig.3 Calculated neutron energy distribution profiles for D-T reaction [9]

exp. IRK x exp. CBNM • eval Pavlik( . , 1989) fitted

90 89 lX Zr(n.2n) Zrx

ff(En)= VQ1En

100 - 52Cr(n.2n)51Cr

13 15

Fig.4 Parameters f energyo standard reactions [12]

16 3.0 V ke 0 20 = d E

O) z: 50

2.5

o or 30

2.0 -

10 I I I I 30 60 90 120 150 180 EMISSION ANGLE [9] Fig.5 Neutron energyV ke emissions v 0 20 reaction= D angle, D- E r t a fo

3) the energy spread of ions, 4) the slowing down and scattering of projectiles in the target. Therefore, a value of (14.00 ± 0.07) MeV at 96°-98° is recommended a referenc s a r absolutfo e e normalizatio e energth f o ny scale witneutronsT D- h . 90 8952 51 Shape d parameteran se cr(Eth f no ) s curver fo Zr(n,2n)s d Zan Cr(n,2n)r Cr energy standards give n Refi n . e [12show ar valuA ] n . f Figi n(464 o e 5)m.± 0 b is recommended for the cross section of the 93Nb(n,2n)92mNb fluence monitor aroun MeV4 1 d . The error of this method is determined by the shapes of the 90Zr(n,2n) or the Cr(n,2n) cross section curves and the source-sample geometry, but it does t exceeno keV0 5 n daverageo , , whilis e sensitivitth ,V e Me 5 1 y d betweean 3 1 n 50 %/MeV and 64 %/MeV for Zr and Cr, respectively. The activity ratio is measured with an accuracy of better than 1 % and so the error in the determination of the mean neutron s abouenergi keV0 V 2 t t14.a y .Me 1 This metho s alsi d o applie o t ddetermin e meath e n neutron n energa r fo y extended sample. In this case the Zr and Nb foils are placed back-to-back in different positions inside the sample.

e calculateTh d energy-angle function D reactioD- s e r showEar nnfo (0 n i )n Fig.5 for thin and thick targets at E, = 200 keV.

17 The measurements of the mean energy and its spread require appropriate en- ergd fluencan y e monitors e angula neutronD e Th measure.b D- n e r ca th syiel f do d either by the U(n,f) using a depleted U layer in a fission chamber or by the 238 115In(n,n')115mIn reaction. The a r(E ) curve is relatively well known [13] and its change between 2 and 3 MeV is within 2.0 % (see Fig.6). For 115In(n,n')115mIn reaction, the high cross section value in this energy range is advantageous; however, this inelastic process is sensitive for the scattered neutrons. Measurements carried out in Debrecen, Jiilich, Dhaka and Khartoum have indi- rate d originatV ) e datdiscrepancieE Me thath a( a3 t ebetwee d e an th 2 n i sn mainly from the presence of room scattered neutrons and self-target build-up in beam apertures e contaminatioTh . e neutroth f o n n spectru e th scattere y mb d neu- e decreaseb tron n a scattering-freca sf i d e arrangemen s i usede tcontributio Th . n of the scattered neutrons to the activity can be checked [14] via the 1/r2 relation between the apparent activity and the flux values.

EMISSION ANGL] [9 E 50 100 150 T

500

1.00 6 ( 90*) JO 200ke= , E Vt a e a

400

0.95

238 U (n,f

300 i —— i —— i —— I I i 1.0 0 3. 2.0 U.Q NEUTRON ] ENERG V Me [ Y

Fig.6 Absolute and relative cross section curves for the 238U(n,f) reaction between the threshold and 14 MeV neutron energy

18 2.4 D(d,ne H ) E.=200keV a 2.2

O measured 2.0 eye-guide —— calculated

>s V

J3 cQe> 1.4

1.2

i.o

0.8 LJ- 50 100 150 Emission angle

Fig.7 Measured and calculated angular yields of D-D neutrons

Table 6. Recommended cross section f energo s y monitorneutronD D- r fo ss

Neutron a(mb) a(mb) a(mb) e nergy M 58 Zn(n,p) Ni(n,p) 31P(n,p)

1.9 - 2.0 6.095 36.8 8.353 2.0 - 2.1 8.635 46.28 9.338 2.1 - 2.2 11.65 60.12 14.57 2.2 - 2.3 15.82 73.95 24.04 2.3 - 2.4 20.14 87.79 31.86 2.4 - 2.5 26.77 101.6 38.02 2.5 - 2.6 34.36 117.1 46.50 2.6 - 2.7 43.10 134.3 57.30 2.7 - 2.8 54.70 151.6 62.52 2.8 - 2.9 66.45 168.8 62.17 2.9 - 3.0 78.80 186.0 61.49

19 Table 7. Energies of residual particles emitted in T(d,n)4He, D(d,n) 3D(d,p)d Hean , 3H reactions

Residual Bombarding particles deuteron Energ f residuayo l particles (MeV) energy (MeV) 0° 90° 180°

n 0.050 14.554 14.068 13.599 froT mD- 0.100 14.783 14.088 13.432 reaction 0.150 14.960 14.108 13.304 0.200 15.117 14.128 13.203 0.250 15.259 14.148 13.117 0.300 15.390 14.167 13.042

n 0.050 2.723 2.462 2.225 from D-D 0.100 2.852 2.474 2.146 reaction 0.150 2.958 2.486 2.090 0.200 3.052 2.498 2.045 0.250 3.139 2.511 2.009 0.300 3.220 2.524 1.978

4He 0.050 4.041 3.531 3.086 from D-T 0.100 4.261 3.520 2.910 reaction 0.150 4.436 3.551 2.779 0.200 4.587 3.501 2.672 0.250 4.723 3.491 2.581 0.300 4.848 3.481 2.499

3 He 0.050 1.093 0.807 0.596 from D-D 0.100 1.223 0.795 0.516 0.150 1.329 0.782 0.460 0.200 1.423 0.769 0.416 0.250 1.510 0.757 0.380 0.300 1.591 0.745 0.349

20 Table 7. (cont.)

Residual Bombarding particles deuteron Energ f residuayo l particles (MeV) energy (MeV) 0° 90° 180°

P 0.050 3.324 3.036 2.772 froD mD- 0.100 3.465 3.048 2.657 reaction 0.150 3.579 3.061 2.617 0.200 3.681 3.073 2.566 0.250 3.773 3.085 2.523 0.300 3.860 3.098 2.486

3H 0.050 1.311 0.997 0.758 froD mD- 0.100 1.451 0.984 0.668 reaction 0.150 1.565 0.971 0.603 0.200 1.666 0.959 0.552 0.250 1.758 0.946 0.509 0.300 1.844 0.933 0.472

e Th 64Zn(n,p)64Cu, 58 Ni(n,p)e 31th P(n,p) d 58Can 31oS i - reactionre e ar s commended as energy monitors for D-D neutrons in the 2-3 MeV range. The latter is a pure beta emitter. Cross sections in the 2 and 3 MeV neutron energy range are summarized in Table 6. e angulaTh D neutronr D- yiel f s o i measured s e th 115 In(n,n')y b d 115mIn reac- V [15ke ]0 20 bombardintio t a n g deutero e seenb n Fig.7i energiesnn e ca th , s A . measure d an calculated d thick targe n gooi t yieldd e agreemenar s t with each other, provin e flath g t e crosth shap f so e section curv f o 115e In(n,n')115m- re In s recommendei b m 5 ± 5 e d2.1-2.actio 32 rangeth V betwee = n n i 9Me valu A a . f no e 2.1 and 2.9 MeV range for the determination of the D-D neutron fluence. For D-D reaction the neutron energy at & = 100° is almost constant in the

50 ^ Ed ^ 500 keV range. A value of En(100°) = (2.414 ± 0.010) MeV can be accept r normalizatioefo d e energth f o ny scale witneutronD D- h s using - 100-20in V 0ke cident deuteron energies. Considerin e possiblth g e e sourceenergth f o ys spread, rangdegrees0 V ke 10 th e5 e2 betwee d ,FWHd an an 0 0 Mn 10 valu e th e n musi e b t respectively. From the pulse height spectra measured at different emission angles by a 3He proportional counter, the mean energy and its spread can be derived [16].

21 The absolute source strength of the 3H(d,n4) He and 2H(d,n3) He reactions can be measured accurately (<0.5%) by means of the associated particle method (APM), i.e. by the registration of the He and He particles from the D-T and D-D reactions, respectively n alpha y ab , detecto a give n i rn solid angle. The differential cross sections of the 2H(d,n)3 He and 2H(d,p)3 H reactions are the same at 90° up to a few 100 keV deuteron energies [17]. Therefore, the source e strengtdetermineb n ca h y observinb d e recoith g l triton r protono s s with a silicon surface barrier detector n TablI . 7 energiee residua th f o s l particles reactionD D- d e summarizedar an s T produceD- e th . n i d In addition to the 93Nb(n,2n) , the 27Al(n,a) activation detector is also a good fluenc eT neutrons monitoD- r fo .r Fission h chamber"T d an s U with foils, long counters and liquid scintillators are used as prompt monitors for recording the source strength as a function of time. spectrThe of backgrouna d neutron determineare s d activatiothe eithe by r n threshold foils methoy prompb r o d t spectrometry base r exampl fo dn NE-21 a n o e 3 liquid scintillator. neutroV e Me idea Th 4 e n1 neutro lth fiel f o dn generator s i contaminates n i d practice by lower energy groups to some extent. The most important sources of these non-1 neutronV 4Me s are:

1) Elasti d inelastian c c scatterin V neutron e e originatargeth Me th f 4 n o i g1 ts l target assembly, sample holders, bulk samples, APM head and electronics, air, construction materials of the target room. 2) The D + D neutrons from the drive-in target, resulting in a 0.1-1% of the yielT D- d dependin e conditioth e tritiun th o g f o mn target. D neutron+ 3D ) s fro e beamth m apertures (i.e. beam acceleratin d transporan g t systems, diaphragms, etc.). n induceio , d D. reactione cas d th f 4nonanao ) ean n i Contributio ~ s D - e th f o n lyzed beam.

n I orde o t decreasr e influencth e f o these e parasitic neutrons e targeth , t should be placed in the center of the target room at equal distances from the walls, floor and ceiling. Thin wall, air cooled target holders [1] are ideal to w contributionlo t e backgrounge th f o s d neutrons e thickTh . , water cooled targets a inheavy-se t targe e tmixe th holde d d an bear m neutron generators (i.e. sealed tubes) may produce a contaminant of non-14 MeV neutrons of about 10%.

22 3. DETERMINATION OF THE BEAM ENERGY

According to eqs (3) and (4) the angular yield and energy of neutrons depend on the incident deuteron energy. e absolutTh e measure b n bea eio n menergca n y usin a b d f n o a ygelectrostat - ic, a 180° magnetic or a velocity analyzer. In addition to the absolute methods, the energy calibration of particle accelerators is usually performed by a 90° magnetic analyzer which must be calibrated with nuclear reactions having accurately known resonance and threshold energies. A magnetic analyzer containing entrance and exit slits is transparent only for ions with charge Q and energy E

) (5 Kf= 2E Q2/m

e analyzinth whers i K ge magnet calibratio R frequence NM th e nf i th yfactor s i f , strength of the magnetic field is measured by a nuclear-magnetic-resonance probe (usuall e a protononrelativistiy th s i n E probe d can ) kinetic energy [18]e ,i. the analyzer is a tool for Q/m separation and for measuring the energy E. In addition to the absolute methods, measurements of current through a bank of resistor s wideli s y use - dbecaus s simplicitit f o e - w hundreybelo fe a wV k d terminal voltage. However t a hig, h voltage, discharge e e surfacth th f n o o es resistor d theian s r unstable value n caus ca sn uncertainta e . f kV abouo y 5 ± t During the last decades, new methods based on nuclear reactions have been developed for precise beam energy measurement. For instance, by using a Ge(Li) detector to measure the energy of the direct capture gammas emitted in the 12C(p,y)13N non-resonant reaction, the absolute proton beam energy between 150 coulV e ke determineb d0 an35 d d wit a precisioh [19]V e absof ke abouo n Th . 4 -0. t lute energy calibration is possible even below 100 keV by using the D(p,y)3He nonresonant reaction [20]. This process was observed [21] as low as e D(p,y)valueQ th keV5 2 f e o s= 3Th .He , E 12C(p,y) d 16O(p,y)an 13Ne ar 17F 5.4936, 1.9435 and 0.60035 MeV, respectively. The nonresonant direct capture method has been developed [22] in an experimental arrangement as shown in Fig.8. The y-ray yield curves are measured at target location 1 with a Nal(Tl) detectore seconth t A d . location e y-rath , y transition e measureb n ca s d wita h high resolution Ge detector at 0 and 90 degrees. According to the kinematics of an X(p,y)Y reaction, the energy of a gamma photon emitted by a nucleus decaying t resa s i t Mx E° = ° + E

23 TARGET LOCATION 1 TARGET LOCATIO2 N

. 72cmJGe{Li) DETECTOR

8x8cmNaI(TU DETECTOR Fig.8 Experimental setup for measuring proton beam energy via the y-ray detection emitted in non-resonant (p,y) reactions

10" 1.0

11 -6 10 CO

-7 10

0.5

-9 10 C£.

,-10 10 459keV RES. 0.0 10' /1 I I ill I I 7 0. 6 0. 5 0. 4 0. 3 0. 2 0. 1 0. EplMeV]

Fig.9 yield curvesC f (p,y)o d an reactions B n o

24 Because of the Doppler and recoil effects, the energy of detected gammas differ e relatioss Th i fro. nmE E betweed an E n \J \J 3

2 2 n #2.1/2 E E (MeV) E ( 1+ = £E ii^LJ COS° = E0 —o———_ )—— — _ —) (7 0.5367 V 8ke ) (u M ° 2 c M 2 ° s

e anglth s ei e betwee0 y-rayth d d e an .directionth an n c / M v f wher= o s i f e By measuring E at 90°, the energy of protons can be determined from eqs (6) d d (7)an . Equatio ) show(7 n s tha t a 90°t , onle recoith y l effect mus e takeb t n into account e e detectoG calibratioTh .r e higfo th r f ho n energy gammaa s i s difficult task. If the measurements are based on the D(p,y)3He reaction, the 6.1291 V gamm7Me a line wit s singl it hd doublan e e escape peak t a s5.6181 d an 7

5.10717 MeV, respectively e useb r fo calibrationdn ca , . Suc a gammh a lins i e

16 emitted byN 16 produced in theO(n,p) reaction. The ^Ga isotope is also a good energy standare 833.th 6- n 4807.i drange V ke 0 . High energy gamma-ray standards for detector calibratio e summarizear n n i dRef . [23]. Procedur e energth f o ye calibration by non-resonant reactions has been described in detail in Refs [1, , 24]22 , .1920 , Typical yield curves for the (p,y) reactions on thick nB [1] and thin 12C [25] target e showar s Fig.9n i n . There are a number of resonance reactions recommended for energy calibration. However, below 200 keV proton energy only a single calibration point, the nB(p,y)12C resonanc = (163. E 0.21± t a e) keV s availablei , . Precise energy calibratio w voltaglo f o en neutron generator s i especialls y importan e th machin f i t a s i use echarges a d d particle acceleratoe th r fo r determinatio f atomio n d nucleaan c r data wit a high h e accuracth r fo s wel a ys a l improvement of various technological applications ( e.g. ion-beam imaging, ion microtomography, particle-induced X-ray emission, Rutherford backscattering, ion implantation, prompt radiation analysis). Therefore, improvement of the neutron generators with a beam energy analyzer a bea d man profile detecto s strongli r y recommended.

25 4. TYPES OF NEUTRON GENERATORS

The principle of the neutron generator is shown in Fig. 10. This schematic applies for all particle accelerator. A schematic drawing of a home-made generator workin n Debrecei g s i shown n Figi n . Fro.11 m this s figuri eas o t folt i ye - loe functionwth f eaco ss positionhit unid an t . The ion source, the extraction, the focusing, and the gas supply units placed on the HV terminal need electric power, cooling and insulated remote contro r Pennino l n sourcF systemsio R gs e i e typ r Th standar.fo e d (commerciar o l medium size) neutron generators while it is a duoplasmatron (duopigatron) if high current e requiredar s e deuteriuTh . s ga flo ms i w regulate n io d e intth o source by needle, thermomechanical or palladium leaks. As the ion beam analyzer requires relatively high power, the separation of the D beam is made usually after acceleration. n sourc io e pulsine Th eth f allowo g e productioth s a pulse f o n d beam, i.e. 14 MeV neutron burst, in the [is, range, while the nanosecond pulsing can be made before and after acceleration using special bunching units. The acceleration tube is generally a homogeneous field type allowing a high pumping speed for the vacuum pumps situated at the ground potential. The intense neutron generators (beam curren ) utilizt mA ove0 e1 r strong focusing single-gap d thre p an tube ga eo y otw b whicsre spacth he charge effecte takear sn into account. e Th high voltage power supplie e usuallar s y Cockcroft-Walton voltage multipliers, Felici-type electrostatic machines, medium frequency Allibone-type voltage multipliers, insulated core transformer HV supplies or parallel powered Dynamitron voltage multipliers. In a number of commercial neutron generators, separate insulating transformers are used to supply and control the HV terminal. These types are as follows: SAMES J-15 and J-25, TMC A-lll and KAMAN 1254 models. Most of the non-commercial neutron generators use a single insulating transformer or an insulated motor driven generator (see Fig. 11) o ensure t unit,e th powe th e r se th fo placedr n o HV terminal. The control of these power supplies and other components like needle valves can be performed by insulating electromechanically driven rods (MULTIVOLT, KFK d TOSHIBan I A neutron generators insulatey b r o ) d optical fibers (SAMES series T neutron generators, INGE-1 in Dresden, RTNS-II at LLL, OCTAVIAN and FNS in Ja- e pan)opticaTh . l cable se terminaV betweeth controH d e an llth n desk give th e computer control and regulation of the whole neutron generator. Most neutron generators havn beaio e m handling facilities: magneti r eleco c - trostatic quadrupole lenses, beam profile monitors, ion beam analyzers (simple electromagne r Wien-filter)o t , beam d stopan scannerss e ,th etce d sizTh an e.

26 terminaV H l Acceleration Beam handling Target sourcn Io e Lensesn io , ———— «=- Suppr.Volt. x + , + , + , , s supplGa y (d ,d2,d3) Beam analyzers Cooling Cooling Ion beam Pulsing.etc. 3H trap Power supplies !____ — 1 AccelV H . Control Insul.transf. console t Vacuum L______and _,„____-_J cooling systems

Fig 0 Block-diagram.1 f neutrono generator

HIGH VOLTAGE TERMINA_L_

LIQUI 2 TRAH D P 2QOkV

DIRECT TARGET -3QOV CURRENT

SUPPRESSO^ R -300V

300 kn

REGULATED MAINS

Fig 1 Diagram.1 workinga f o neutron generator

27 technical solutioe targeth f o tn holder e depende purposneutroth th n f o o se n generator. A neutron generator utilized mainly for activation analysis, neutron therap r materialo y s researc y havma h a bulke y water cooled target holder, espe- A targem tw fe current a e ciall intensr Th fo .y e neutron generators, having several kW target loads, can manage this load only with water cooled rotating targets. The thin wall, low mass target holders with air jet cooling [1] are recommended for "clean" neutron work, around 2 and 14 MeV. The vacuum system of a neutron generator consists of prepumps (mechanical, cryogenic d higan )h vacuum (diffusion n getterio , , turbomolecular- ) ad pumps e Th . vantages of the diffusion pumps are as follows: low cost, high pumping speed simple maintenance, long lifetime, no mobile components. Their disadvantages are: oil vapour contamination of the accelerator components and the target which can e decreaseb usiny b d f o gFLOMBI r SANTOVANo C d oilliquian s d nitrogen traps (see Fig. 11) not only at the inlet of the diffusion pump but also along the beam lines. e titaniuTh m getter pump n assurca s e clean vacuu d higman h pumping speer fo d hydrogen (deuterium) gas; however e higth , h absorption r rattritiufo e m released from the target makes it difficult to handle the used pump elements because of their high activity e ,cas f th moderato eve e n i n e commercial neutron generators. e cleanesTh t e achieveb vacuu n ca y usinmb d g turbomolecular pumps. Their disadvantages are the relatively high cost and the possibility of mechanical damage. In general, Pirani and thermocouple vacuum gauges are used for measuring the forevacuum. The high vacuum is measured mostly with Dushman type or Penning type e vacuuth f o m e gaugescontrollerus e Th . s makes possibl e th automatioe e th f o n neutron generator control. The rubber or Viton O-rings are changed these days for metal gaskets e contemporarth : y neutron generator e bakeablus s e vacuum components as well. With neutron generators used for research purposes, control is mostly manual, but the principle of minimal interlock and other control inputs is also followed with commercial generators. The multipurpose and intense neutron generators use microcomputer and computer control systems. The choice of neutron generator depends on its construction and purpose. This Manual outlines the commercial (medium), intense (high current) pulsed and sealed tube neutron generators, dealing with their operation, technical solutions, maintenanc d repaian es wel a s rwit a l h updating wit a vieh o t extendinw g their utilization in science and technology.

4.1 COMMERCIAL NEUTRON GENERATORS The commercial neutron generators are modest machines with production yields D reactionsD- d an , T respectivelyD- r fo s .n/ The 0 e 1 ar y d an s ofn/ abou 0 1 t 11 9

28 utilized in basic nuclear research, education and technology, for measurement of nuclea n i rlaborator datad an , y exercise o t studse differenth y t interactionf o s neutrons, detection of charged particles and neutrons. They are also used in accelerator technology for activation analysis, prompt radiation analysis, irradiation effect f o fass t neutrons, neutron dosimetry, etc. Practicall o n technology - ical development in the field of commercial neutron generators has been reported n i recen tw companie fe yeard onl an a e sy producinar s g these moderate ) machinesmA (150-302 1- , . 0kV The existing manufacturers are IRELEC (formerly AID and SAMES) in France, KFK Hungaryn i I e EFREMOth , V Institut Russian i e d MULTIVOLan , e th n Ti United Kingdom. The sealed tube replacement is still made by KAMAN NUCLEAR in the USA. These type f neutroo s n generato e stilar n operatioi r l n mani n y countries. As these machines are excellent devices for pure and applied research, service and education e IAE th s ,provide Aha de followinth the o mt g developing countries: Albania, Algeria, Bangladesh, Bolivia, Burma, Costa Rica, Cuba, Equador, Hungary, Indonesia e formeth , r Yugoslavia, Lebanon, Malaysia, Mongolia, Morocco, D.P.Rf o . Korea, Nigeria, Pakistan, Peru, Singapore, Sudan, Thailand, Turkey, Zambia. Unfortunately som f f ordertheso o e t eou , generatorandw n no additioi , e ar so t n repair of the machines, it is recommended that they be fitted with some upgrading components r exampleFo . , determinatio e nucleath f o rn level scheme d neutroan s n cross sections requires microsecon d nanoseconan d d pulsing units. These systems have been developed almost entirely in the laboratories where the neutron generators were constructed and are not available commercially. Therefore, strong co- operation is required between the developing and advanced laboratories for up- gradin e commerciath g l neutron generators with pulsing unitd othean s r components. A a numbes f manufacturero r s have closed e lasdowth t n i decaden , some "second hand neutron generator" companies (POTENTIAL in the USA or MULTIVOLT in the UK) have y usestartebu d o t machined d restoran s e the o mtheit r original conditions. POTENTIAL specializes in neutron generators manufactured by TEXAS NUCLEAR [29], while MULTIVOLT deals with those made by SAMES. The lack of spare parts and components causes many probleme userth - r sespeciall e fo sindustriallth n i y y less developed e fielcountrief th maintenanco dn i - s d repairan e . Dealing with "second hand neutron generators" is particularly important for the developing countries. Table 8 lists the most popular commercial neutron generators. Improve- ment of the original characteristics of a commercial neutron generator, i.e. analyzing magnet, quadrupole lenses, pulsing systems, associated target assemblies, etc., needs local or international co-operation with experienced laboratories. An excellent example of such international co-operation is the Fast Neutron Research

29 Tabl. 8 e Commercial pumped neutron generators

Manufacturer Type U/I Yield Pulsing [kV/mA] [n/s]

10 TMC A-lll 180/1.5 >10 yUS TEXAS NUCL. 280/7 >10U — n KAMAN A- 1254 190/2.2 >io ns-/^s TOSHIBA NT-200-5 200/1 >1010 — n KFKI NA-4B 120/1.3 >io — HIGH VOLTAGE LN-S 300/2 4xl010 — 10 ACCEL. 150/3.5 3xl0 fiS MULTIVOLT NA-150-2 150/1.5 >1010 ps NA- 150-4 150/3.5 SxlO11 fis 10 SAMES J-15 150/1.5 >10 JtS J-25 150/2.5 2xl010 A«s JB 150/2.5 2xlOU //s TB 300/8 ~1012 10 D 150/1.5 5xl0 /IS EFREMOV NG- 150-1 150/3 -1011 —— NGP-11 150/2 2xlOU — NGP-11M 175/5 SxlO11 —— NG-12-1 250/10 ~1012 —

Facility at Chiang Mai University, Thailand, where the commercially built SAMES J-25 neutron generato s beeha nr completed wit a h post-acceleration nanosecond pulsing system and an associated particle target head. This effort resulted in a good time-of-flight spectrometer laboratory based on a modest commercial neutron generator. These were originally small compact machines, manufactured mainlr fo y activation analysis. A comparisio e commerciath f o n l neutron generatorn sca (se) e10 Tabled an 9 s help user o t selec se appropriatth t e machina specifi r fo ec application, while knowledge of the technical solutions of these generators may help in the improve- men f thein machineo t n repaii ow r d troubleshootingd an ran s . Some componentn ca s e similab r enouge useb n n theii systemsdo t how re maiTh . n characteristicd an s

30 Table 9. Comparison of different commercial neutron generators

S A M E S TMC KAMAN T* KFKI MULTIV. EFREMOV 3 1 2 1 Typ. 1 1 No e 12345 0 1 9 8 7 6

Neutron yield [n/s] >1010 + + + + + + = ion + + + + + + + + <1012 - 1012 Beam energy [keV] <150 + = 150 [keV] + + + + + + <200 + + + >200 + + + Beam current [mA] >1 + + + + + + ++ + > 3 + + + + > 5 + n sourcIo e Penning + + + + RF + + + + + + + + Duoplasmatron + + Gas supply PD leak + + + + + + + + + Electrolyzer + + + + - -I Mechanical Sealed tube + HV supply Electrostatic + + + + + Mains frequency ++++ + + + Medium frequency + + Terminal controy b l + + + + Ins.transfor+ + . + + + + + Ins.rod Fiber optics + + +

* TOSHIBA

31 Table 9. (cont.)

C KAMATM NT S KFK E IM MULTIVSA . EFREMOV Type No. 123456 7 8 9 10 11 12 13

Accel.tube Homogeneou+ s Inhomogeneous Vacuum system Diff.pum+ p Titan getter + Turbomolecular Pneumatics HV terminal insulation Air + + Oil + SF-6 solid Cooling system Open + Closed circuits Power consumption + A <2.kV 5 A >2.kV 5 A kV 4 > >8 kVA Optional pulsing + s /< 0 10 < > 100 /«s + Beam+ stop Spare parts Easy + + Available Not easy + Maintenanc e machinth f o e e Comfortable Not comfort+ .

TOSHIBA

32 Table 10. Advantages and limitations of the commercial neutron generators

TYPE ADVANTAGES DISADVANTAGES SAMES-D Simple construction Mixed beam Simple insulation (oil) Not easy maintenance and Small room for installation repair of HV terminal Electrostati uniV H ct Horizontal use only SAMES-J Easy maintenance Mixed beam Reliable electropneumatic Unreliabl numerouo to d an e s safety system insulating transformers Easy beam line assembling Electrostati uniV H ct SAMES-T Easy maintenance and repair Mixed ion beam Reliable electropneumatic The maximum HV can be safety system achieved only in an environment Electrostatic HV unit that is not humid Insulator shaft motor generator terminaV H e poweC th A l t a r SAMES-JB Compactness Mixed ion beam Small roo r installatiomfo n Not easy maintenancd an e Simple installation and repair position change Special tools Safe operation in humid or low pressure atmosphere SAMES-TB Compactness, small roor mfo Mixe bean io d m installation Not easy dismantling, repair Simple position change and maintenance Safe operation in humidty Special tools TMC A-lll Compactness, easy achievement Mixed ion beam of horizontal or vertical beam Not easy dismantling, repair Simple closed circuit cooling and maintenance Unreliable needle valve KAMAN A-1254 Compactness, small room for Mixed ion beam installation Unreliable needle valve Simple position change 6 insulatioSF n FREO coolin3 N11 g

TOSHIBA Long 2 lifsupplD e y Mixed ion beam NT-200 Compactness Unavailable spare parts acceleratioV k 0 20 n voltage KFKI NA-4 Compactness Mixe n beaio d m Long life D~ supply Low acceleration voltage Small room for installation (120 kV) MULTIVOLT Compactness Mixed ion beam NA-150 Availability of spare Relatively low yield parts and components High D2consumption

33 technical solutions are given in Table 9, where the numbers correspond to the following manufacturer typesand s :

5 , 4 , 3 , 2 , 1 ) Typ . TB No e, JB , T SAME , J , S(D TMC (A-lll) 6 Typ. No e KAMAN (A-1254, A-7118 ) 7, Typ . No e TOSHIBA (NT-200) 9 Typ . No e KFKI (NA-4) Type No. 10 MULTIVOLT (NA 150-02 and NA 150-04) Type No. 11,12 EFREMOV (NG-150 I) Type No. 13

4.2 SEALED TUBE NEUTRON GENERATORS The sealed tube neutron generators or neutron tubes are 14 MeV neutron sources used in in-situ geological measurements (bore-hole logging), in hospitals ann i chemicald , biologica d industriaan l l laboratories. Most neutron tubes have a Penning ion source, a one- or two-gap acceleration section, a tritium target and deuterium-tritium gas mixture filling system. The pressure of the gas is controlled by a built-in ion getter pump and/or gas storage replenisher. The tritium displaced froe targeth m t durin e operatioth g s occlusioga s i absorbene th n y b d elements and is accelerated later back into the target, resulting in an extended target life [26]. The operation principle of a typical sealed tube generator is showe firs Th n tFig i n. seale.12 d tube neutron generato s developewa r y PHILIPb d S [27] and its originally low neutron yield could increase up to over 10 12 n/s. The advantage of the neutron tubes is their small size which makes them suitable for geological bore-hole logging or isocentric cancer therapy in hospitals. The disadvantages are their limited lifetime and the relatively high neutron production cost. Typical applications of the neutron tubes (with built-in high voltage power supplie d neutroan s n detectors s a follows e ar ): Bore-hole logging, deposit evaluation, uranium exploration and processing, mineral exploration, coal exploration d processingan l weloi l, logging s welga l , logging, under-sea exploratiod an n such utilization where the small diameter and the remote detection are important. For cancer therapy the isocentric irradiation is most important: the neutron source and the neutron collimator are rotated around the malignant tumour in n lineai th es a r sam y acceleratorwa e t radioactiva r o s e sources. The deuteron ions - bombarding the tritium target - are produced in a Penn- poweV H V r k suppln 0 sourc1 n insourceio io - g y e s powerei e0 consist Th .a y b d - n a insulatin f ino g g high voltage step-up transforme d singlan r e wave rectifier. The discharge current of the ion source can be observed by an !<, ion source amp meter. Owing to the negative resistance of gas discharges, a damping resistor

34 PENNING VACUUM

GUAGE REPLENISHER TiT TARGET PENNING ION SOURCE

ION SOURCE CURRENT

STEP-UP TRANSFORMERS PRESSURE CONTROL FEED BACK

VARIACS

ACCELERATION N SOURCIO E EXTRACTION

H V POWER SUPPLIES

Fig.12 Schematic diagram of sealed tube neutron generator

HV CONNECTION VHV CONNECTION RESISTOR TUBE TN26 HOUSING (ION SOURCE)

\ MECHANISF O M INSULATO L OI TIMR E COUNTE V CONNECTIOL R N EXPANSION (GAUGE AND RESERVOIR)

Fig.13 Schematic diagram e SODERNth f o sealed tube neutron generatorr fo bore-hole logging [28]

n sourcprotectio e eth s power supply. This power suppl ye voltag th float t f a o es U acceleratin f o m su ge th U voltagextractio d an en sourcio n - voltageeU e Th . + + • • e voltagregulateb e righn th ca ety b d han T dd mixe sidan ed D variacion e s Th . U e extractio extractee ar th y b nd voltage e intfirsth ot acceleratiod an p ga n c accelerateU acceleratio e th y b d n voltage withi e seconth nT p towardTi ga d e th s target. The operation of the sealed tube requires a definite gas pressure in the

35 s pressurga tubee s i regulateTh e. s replenishey heatinga b d e th g r (mainly zircon or titanium). The gas pressure is measured mainly by a Penning type vacuum meter. Some sealed tube neutron generators utilize feedback control between the neutron monitor and the replenisher to achieve a constant neutron yield. Sulphurhexafluoride (KAMAN 711), oil (several PHILIPS sealed tubes) or solid plastic materials (SODERN e tubesinsulatio th e e seale user th ar )fo df do n tube generator head [28].

e KAMATh 1 sealeN71 d tub s placei ea pressur n i d e vessel unde e insulatth r - ing compressed Sulphurhexafluorid n sourcio e s i e cooleegas Th y .electricallb d y insulating liquid FREON 113, while the target is cooled by circulated water outside the pressure tank. The arrangement makes it possible for the small so-called accelerator hea - d whic s onlo ha pairh f tw yo scoolan a couplt f d pipeo an e s e moveb cableo t d - se easilcircumstanced th adapan yo t te investigationsth f o s . This sealed tube neutron generator deliver s durin n/ s guaranteeit s g0 ove 1 r d life-time of 200 hours. e PHILIPTh S sealed tube type a 1860built-is ha 1 n Dushman type ionization vacuum gauge, a pressure regulating replenisher and a high voltage damping resis- m diameterm m long0 m 7 7 , 73 , s usuallit l oi filledn yi ,r to metal tube container. This typical bore-hole logging neutro na pulsin s tubha eg capability betwee A d schemati 100. an ^s 0 5 n c diagra a recenf o m t solid insulation tube packed SODERN neutron generator is shown in Fig.13 [28]. Therapeutic use of neutron generators assumes a minimum 14 MeV neutron yield of > 10 n/s. The relatively small size of the neutron tubes is an advantage in fast neutron therapy: they can be easily installed in the neutron collimators of the isocentric treatment geometry e problemTh . s relate o t tritiud m handlinn i g pumped neutron generators do not exist in the case of the sealed tube, so sealed tube intensive neutron generators are ideal for fast neutron cancer therapy. The MARCONI-ELLIOT HAEFELe Tth [30d an ] Y (the KARI KORONAr No e seale)ar d tube neutron generators that can meet the criteria of the high neutron yield and relati- ly long life-time e MARCONTh . T a mixebea, s mD ha dI containin a grefillabl e sealed tube with typical high frequenc a n sourc io homogeneouyd an e s field acceleration tube. The target material is Er, to achieve high thermal stability. The manufacture n rene ca e rsystem th w d thi s an i advantageous, r situationfo s s where regular chang f useo e d sealed tube s i requireds r canceFo . r therapy e sealeth , d e collimatoth tub n i e s changei r d regularly every week. The KARIN tube is manufactured by HAEFELY in Basel. This generator has an annular shaped Philips ion gauge (PIG) ion source at the ground potential, while e cylindricath r conicao l e th higl n htargeo voltags i t e th e n i potentia d an l

36 ALPHA -PHOTO MULTIPLIER SCINTILLATOR TUBE D/T ION BEAM TUBE ENVELOPE ALPHA ACCELERATING PARTICLE ELECTRODE HIGH VOLTAGE INSULATOR

VACUUM HIGH VOLTAGE PINCH-OFF ELECTRODES

TARGET

GETTER -FARADAC 1 Y FOCUS CAGE ELECTRODE SOURCE -14MCV NEUTRON

Fig. 14 A sealed tube neutron generator combined with APM head for time correlated measurements

Table 11. Compariso f sealeno d tube neutron generators

Type/Name Manufacturer Life-time Acc.voltage/Beam current Yield [kV/mA] [n/s]

A-3045/6 KAMAN 200 h 200 kV/A m 5 3xl010 A-3041-4 KAMAN 2xl06 pulses 5xl07 A-3043 KAMAN 100 h 5xl07 18601 PHILIPS 1000 h 125 kV/0.A 1m >108 18604 PHILIPS 400 h 200 kV/A m 8 >1012 TN-26 SODERN 2500 h 125 kV/0.1 mA 2xl08 MARCONI 200 h 200 kV/20 mA >1012 KARIN HAEFELY 400 h 200 kV/500 mA 5xl012 n TN-46 SODERN > 1000 h 225 kV/2 mA >io

center of the annular ion source and spherical collimator. A cylindrical pneumatic rabbit system (sample transfer) can be introduced into the center of e targeth t n aloninsulatina s axiit y gb s g polyethylene tub o t achieve a ehomoge - neous sample irradiation, especiall e samplth f s i i yrotatee d during irradiation in the hollow target. The conical targets are used almost only for neutron therapy. The target is produced from scandium titride. The tube can give > 1012 n/s yield durin s guaranteeit g d life-tim 0 hours40 f o .e

37 The long life-time criterion of the sealed tube neutron generators seems to be fullfilled by the new generation of SODERN (France) neutron tubes. Some parameter f theso s e neutron generator e summarizear s Tablen di . .11 The combination of the sealed tube accelerator with the associated particle method (APM) [28 r ] fielfo (Fig d) .14 analysi f o buls k mineral samples, verifica- tion of chemical and nuclear weapons [29,30], and three dimensional elemental analysis in solids [28] has great importance. A schematic diagram of such a combined generator is shown in Fig. 14.

4.3 INTENSE NEUTRON GENERATORS RecentlyneutroV Me 4 n1 , generators wit a yielh d highe s wern/ r e0 tha1 > n 12 develope r fusiofo d n related applications, neutron cross section measurements, production of long-lived isotopes, investigations on radiation effects, cancer therapy and elemental analysis of small samples. The original DC intense neutron generators were equipped later with pulsing system o measurt s e secondarth e d an y leakage neutron spectra. Such experiments require about 100 times higher neutron yields than thaf o today't s intense neutron generator o t achievs e requireth e d accurac f dato r yfusio fo a n reactor design. These programs became more important after the first successful experiment with DT fusion in the Joint European Torus in Abingdon (UK, 1991). The design of a fusion reactor needs accurate data for tritium breeding, nuclear heating, bulk shielding, secondary reactions and gas production. The accuracy of the energy and angular distributions of secondary neutrons - DDX measurements - for blanket and other structural materials should also be increased [31,33,34]. r calculation Fo e induceth f o ds activity, more accurate activation cross sections are needed not only for the materials of the major components but also for their impurities. Precise activation date ar arequire d e dosimth mainl r fo -y etry r reactionunfoldinfo d e neutroan th sg n fiel n i differen- d fu e t th part f o s sion reactor. During the early intense neutron generator period (second half of the seven- ties) when several such machine d beeha s n constructe r fo fusiod n oriente- re d search and neutron therapy, it became clear that these originally DC generators woul e morb d e suitabl r fusiofo e n studie a pulsin f i s g d systean mX helpeDD e th d neutron transport measurements e OsakTh . a OCTAVIAN generato e JAERth S d FN Ian r were equipped with nanosecond pulsin ge firs th unit t s a intenss e neutron generator, the RTNS-I at LLL. Ther e somar e e limitation e recenth f o ts generator n determinatioi s e th f o n data required for fusion reactors (low yield, small irradiated volume, large gra- e neutroth dien f o tn e fluencsample)th n i e . Therefore machinew ne , e undear s r

38 200 kV

c ._. o _

O *—

O 2-

1 -

0 7 0 5 0 3 10 TARGET LOAD [ kW ] Fig.15 Specific neutron yield vs specific target load for solid tritium targets [35]

100-•

a 10 JO E

t/i -2 K) -| ScH

cr ./; S3 K)-| 1.8 2 10-6. 0 70 0 50 0 30 XX) TEMPERATURE [*C] Fig.1 e equilibriumTh 6 pressure r somefo metal hydride systems temperatures v

construction or improvement, e.g. the conceptionally new Osaka generator and the s AlamoLo s supersoni s targega c t system. acceleratioV k 0 20 currenA A m n 0 voltagn 5 o t producd W an k e 0 1 a eloa f o d the target, which give e beas th yiels n/ s mf i selectei aboudD 0 (ser 1 tfo d e Fig.15) [34]. Sinc e thermath e l e stabilittritiuth f mo y targets e dependequilibth n o s- rium pressure of the metal titride system at a given temperature, the use of more

39 Tabl. 12 e Some characteristics of intense neutron generators [32]

Name Yield V ace'/ Itarget Beam diam. Application 1012[n/s] [kV/mA] [mm]

RTNS-I 6 400/22 6 irradiationP TO , RTNS-II 30 380/130 10 irradiation LANCELOT 6 160/200 50 irradiation OCTAVIAN 4 300/35 30 irradiation, TOP LOTUS 5 250/500 (50 cm2) irradiation Chalk River 4 300/25 10 irradiation Lewis/NASA >1 300/30 54 therapy Sandia 4000 (200/40A*) (200 cm2) irradiation " 6'x6' " 10 250/250 1800 radiobiology NRPB >1 600/10 6 irradiation DYNAGEN 3 500/12 20 therapy FNS 5 400/20 15 irradiation, TOP AWRE 2.5 300/12 10 — Bratislava >1 300/10 10 irradiationP TO , Kiev >1 250/15 10 irradiation INTTF >10 180/200 16 target devel. Dresden >1 300/20 15 irradiation Debrecen >1 200/20 10 irradiation Wisconsin >1 (gas) therapy Lanzhou <1 300/5 10 irradiation

e SandiNot th e hig r th eafo h generato) currenA 0 (4 t r

stable titrides than TiT has many advantages. The target cooling is a fundamental problem for intense neutron generators, and therefore the use of a technically perfec s targega t t assembl s i y recommende o increast d e yiel th f eo dpresent-da y neutron generators. At an equilibrium pressure of 100 Pa the corresponding temperature for the TiH, ZrH, ScH, ErH and YH systems are 390, 590, 700, 790 and 860 Celsius, respectively. The behaviour of the TiH, ScH and ErH systems is shown in Fig. 16 [35], and a survey of these generators is given in Table 12. Most of these neutron generators have solid rotating targets and minimal beam line components to achieve the highest possible neutron output. The mixed (DT) beams were used to increase the target life at LANCELOT and LOTUS. A gas

40 80 BEAM LINE WORKSHOP

TARGET ROOM I. TARGET

g /BEA,j p M PROFILE MONITOR

G.V. VACUUM PUMP

QUAORUPOLE LENS ANALYZES GA — R n GATE VALVE

EXPERIMENTAL PORT o O * 0

l J • ^ ' -A ^

ROTATING TARGET

0" BEAM LINE ^ /| Q.L. DEFLECTION MAGNET MOTOR- ALTERNATOR

DEFLECTOR G.V U U ACCELERATION \ Q.L.\ TUBE TARGET ROO. MTJ

\ ANALYZING /' MAGNET N SOURCIO E ACCELERATOR ROOM N SOURCIO E

facilityS FN t JAERI, a Fig.1e viewp th f To 7o Tokai-mura, Japan target is used only at Wisconsin University. The target problems, the tritium e handlinshieldinth d an gg require special building e cas f th intenso e n i s e neutron generators. RTNS-II, OCTAVIA werS FN e d constructeNan n theibuildinn i d ow r g together e supportinwitth h g electroni d mechanicaan c l workshops e relativTh . e amount of associated equipment inside and outside the generator hall can be estimated on the basis of the top view of the FNS facility in Tokai-mura (see Fig. 17) [36].

42 5.0 ION SOURCES: OPERATION PRINCIPLES, MAINTENANC TROUBLESHOOTIND EAN G

5.1 HIGH FREQUENCY ION SOURCES Low voltag T generatorD- e s employ thre en source io type f o s: radiofrequency [37-39], Penning [also called PIG (Philips Ion Gauge)] [26,27,40,41], and duoplasmatron (DP) [31,42-44] comprehensivA . e monograp n sourceio s n ato ha ho d s man been publishe y Valyb d i [44], discussin n i detaig l both theoretica d practicaan l l aspects. The advantage of RF sources is their high monatomic ratio (-90 %), havP D ed higan whil hG currentsPI e . e higTh h frequenc n sourcio y e propose Thonemany b d d co-workeran n s [45d ]an improved at Oak Ridge [37] is applied in different versions in many laboratories. The high frequency discharge in HF ion sources can be generated by an in- ductivel a capacitivel r o y z frequency MH couple 0 y10 d wito t oscillatoa h5 1 f o r power consumption of 100 to 400 W. The discharge generated by the high frequency electric field applied to two electrodes placed outside or inside the discharge tube is called either capacitively coupled or linear high frequency discharge. This type of ion source is used in the SAMES (called later AID), MULTIVOLT, etc., neutron generators (see Fig. 18) [37]. The discharge generated by the high frequency magnetic field in the discharge tube placed inside the solenoid of the high frequency oscillator is called inductively coupled or ring high frequency discharge. This type of HF ion source, used at KFKI, Toshiba, etc., neutron generators is shown in Fig. 19 [39]. n botI h case e dischargth s 0 diametem 10 m e 0 chambed 5 an r o t f abou o r0 3 t to 200 mm long is made of Pyrex or quartz glasses. The free electrons are accelerated by the induced alternating electric field d wilan l E oscillate samth t ea e frequenc e oscillatore th conditioth s r a y Fo .n of ignition of the ring HF discharge it was found that the maximum value H of the applied alternating magnetic field (H = H sintwt) must satisfy the equation

2E- IrT + ("V rH°

e angulath e s e tubei radiuth th ra wher ( f s ,i velocito e shig th r e hf o yfreque n e electronth e mas th f e o collisions i Th . cm y s fieldd betweean , n electronc s and gas particles lead to ionization only if E = eEA a E-, where A is the C 1 66 mean fre ee ionizatio th pat f s electroni ho - E n d f energyenergo san yE . 1 C

43 Coolin n (Algfi )

Pyrex glass

Discharge tube (Pyrex glass)

To oscillator (100 MHz)

Cement (polyvinyl acetate) Magnet coil

Cooling Gas inlet

Extracting electrode (Al) Fig 8 Capacitivefy.1 coupled high frequency sourcen io [37]

e s ignitio dischargTh 2nf.ga e = ) Th e a nth d f evoltago an depend L A I e n o s dependence of U, on f for different gases (at the same pressure) is shown in Fig. ) functio(f t a , r differ argo U e fo curves n Figth I = ga - n. 1 .L 2 s20 U sho e wth s pressuresga t s obvioui en t I . s thae th minimue t th d tha f an o t mi U valu f o e associated frequency increases with increasing pressure [44]. _2 Experiments show that the ignition voltage at pressures below 10 mbar is e dischargth independen n i e s natureth ga d pressurtubf d thae o an etan e th t f o e its value is primarily determined by the secondary electron emission coefficient of the discharge at the wall. In high frequency ion sources, the walls of the discharge tubes are important because, in a real case, the secondary electrons pla a yrol e e owinth impuritie e o wallt gth ,n o s whic e ar usuallh y causey b d electrode evaporation. The impurities will reduce the extracted ion current and thus the efficiency of the ion source. As can be seen in Fig. 18, the ions are extracted through a 1-4 mm diameter and 5-2 lonm 0m g channee harth dn i laluminiu m extracting electrod- e po usin a g

44 -Quartz -Extracting electrode (Al) - Anode(W)

Quartz

Focusing lens

versionA Fig higha 9 . 1 f o frequency sourcen io [44]

(V) p-30jjHg 500 Ar

AGO

300

200

100 Xe

(MHz0 5 ) 0 4 0 3 0 2 0 1 0 Fig.20 Ignition voltage vs frequency

45 (V)

1000

120jiHg 800

600

400

200

0 50 100 150 (MHz) Fig.21 Ignition voltage vs gas pressure

Inhomogeneous magnetic field

Homogeneous magnetic field

0 200 400 600 800 (Oe)

Fig.22 Extracted beamn io densitydischargesF H n i statics v magnetic field

tential difference variable from 2 to 10 kV. To decrease the power and gas consumption a permanen, t magnetic fiel s appliee i dth discharg o t d e volume either transverse or axial to the axis of the tube. By an axial magnetic field the ion concentration at the extracting electrode can be increased because the electrons will circulate on a helical orbit at a Larmor frequency

(9)

46 (W) 200 /j=1.56*10~2mmHg f=30MHi

160

120

H (Oe) 012345678 Fig.23 Dependence powere th f o consumption magnetics v fieldd extractedan densityn io highs v frequency power [44]

and consequentl e ionizatioth y n probability will increase. Furthermore e plasth , - ma with diamagnetic behaviour is compressed in the presence of a magnetic field. The magnetic field parallel to the electric field can be either homogeneous or inhomogeneous. It was found that the ion current density in the linear high frequency discharg s i substantialle y e presenc higheth n inhomogeneoua n i f ro e s magnetic field (see Fig. 22). e poweTh r consumptio e discharge luminositth th f e o plasmnd th an ef o ya show a resonance type increasing in a given interval of the static magnetic field (see Fig.23). This resonance phenomenon can be observed at a pressure of -3 2 p =1x1 5x1o 0t 0 mbar e resonancTh . n i e high frequency ring discharges (inductively coupled HF ion sources) has been observed in transversal magnetic fields at frequencies:

> ft 6 o t ) a T3 J = (H)

47 2 EXTRACTIO5. IONF NO S FRO SOURCEN MIO S Well collimated ion beams with small diameter, which are mainly required in neutron generators e obtaineb n ca ,d only from sources - equippeex n dio wit n a h tracting system. The probe-type and the diaphragm-type extraction systems are generally utilizee th neutro n i d n generators n sourceio n sn I .io (mostl F H y sources) with probe-typ n extractio io n ei thos d e an nutilizin e expandeth g d plas- a surfacem e ion-opticath , l syste s n i msimilaimmersioa o t r n lens (objective).

04.4

08.1 I

04.4 r 012

Fig.24 Probe-type ion extraction system (quartz sleeve and extractor electrode of SAMES neutron generators), Sizes are in mm

48 Fig.25 Schematic diagram f electrostatico immersion lens-type extractiond an the relation between parametersthe

In ion sources (mostly Penning ion sources) with diaphragm-type ion extraction, where the electrodes are shaped so as to reduce the space charge effect in the ion beam, the extracting system is similar to a quasi-Pierce-type ion-optical system. The scheme of a probe-type extracting system is shown in Fig.24. n immersioA n lens (objective) o electrodesconsisttw f o s , both with diameter D. One of them is of length h and closed at one end while the other electrode (ope t bota n h ends s i place) t a somd e distance e frorelatione firstTh th m - . be s twee e th parametern f thio s s simple immersion lene showar s n Fig.2i n 5 (graph). e seeb nn ca Ifrote mth figure thar h<0.785Dfo t e imag e th distanc th f s ,i eo K e negative t I mean. e havw s a evirtua l= 0.785D imageh m e distanc± r th ,fo = : K e < 0.785 h e e imagdistanc r De th th imagfo th s read i ed f an eo elan decreaseh s a s increases. e optimuth n I m case parameterth e e extractinth f o s g syste e chosemar o thas n t the total cross section of the ion extracting channel in the electrode is fully utilize e n channecurrenth io d n whil o e tl th ewal s i diminishedl . Thereforee th , magnification h of the immersion lens defined by K M = ZF (12)

shoul e choseb d o havt n a value e clos o unitt e y (K~2h) r otheFo . r value f thio s s magnificatio e transmitteth n n w valuecurrenio lo e d th s sit t r decreasefo d an s angle of divergence of the ion beam increases on leaving the channel. Under

49 0 1 8 6 4 2 0 Fig.26 Diaphragm-type extractionn io focusings it d an characteristics

thes parameterD e d e conditionsprobe-typan th f h o s e th e, extraction systee mar restricte e valuee rangth th f n o o ei t ds

0.6D * h * 0.9D (13)

The actual characteristics for J. ion current versus extraction voltage U . should be determined experimentally. e diaphragm-typTh e extracting syste s mi applie n i dPennin d duoplasmaan g - tron ion sources. This system can ensure a precise geometry with a well defined emitter n sourcesplasmio F n als H ae use. ca ob r surface t i dfo d an , In the Fierce-type optical systems the electrodes are shaped to prevent the divergence due to the space charge effect in intense ion beams. This ion-optical syste e principl s mth basei n o d e thae directioth t a charge f o n d beam betweee th n two surfaces of two concentric spheres with different radii is not affected by e spacth e charge (see Fig.26). The relation between the extracted ion current J. and the extraction voltage Uext is given by

JKU• i ext (14)

e mosTh t important parametere th metae lengt f th o le 1 hprob ar s e (alumini- e quartth um e f )lengto th z sond h sleevd h an e e ovee metath r l probe (see Fig.24). An extraction arrangement of a quasi-Fierce geometry is shown in Fig.26 with electrodes of approximately ideal shape.

50 Magnet (BaFe) Coolinn fi g Tin- 'Discharge tube (quartz) ^^ Extraction voltage Magnet ( BaFe) V Gas inlet \

Insulator Insulator (Al.O) ,

Extracting electrode (Al)

Fig.27 Capacitivefy sourcen coupledio F H with quasi-Fierce extraction

The quasi-Pierce-type extraction systems consisting of two conical electrodes are used in duoplasmatrons, Penning (PIG) and duopigatron type ion sources with oscillating electrons. e quasi-Pierce-typTh e extraction syste s i muse n sourcedio wits F wella H sh , especiall n thosi y e neutron generator e sacceleratoV termina th H wher d e an th el r

are placed in a tank under pressure of SF6 or oil. Fig.27 shows such an HF ion source with high (5-10 mA) extracted ion current. n arrangemena n I t t showa y applyin z b n Figi n , MH .19 a 5 gfrequenc4 ~ f o y 100 W output power and U =6 kV, an ion current of 5 mA with a gas consup- CXlT tion of ~ 15 cm /h has been achieved [31].

3 MAINTENANC5. DISCHARGS GA F EO E PYREX BOTTLE After several hundred hours of operation, a thin metallic layer is deposited e inneoth n e rdischarg th wal f o l e bottle, whic a reduction lea e o ca t hdth n i n atomic ion ratio of 40 to 50 %. Metals are good catalysts for the recombination of atomic ions; this should be considered in construction of ion sources. The recombination coefficien a relativn o t e r mosscalfo et metal s i abous t unity, while for Pyrex: 2 x 10 5 ; quartz: 7 x 10- 4; aluminium: 0.3. The surface recombination of the ions at the exit canal is decreased by applying a quartz sleeve

51 that acts as a virtual anode and focuses the positive ions into the channel in the presence of extracting voltage. During the operation, aluminium sputtered from the extraction channel is deposited onto the inner surface of the quartz sleeve, causing an increase in the fraction of molecular ion component to well over 40 %. This means that after a long-term operation (50 to 100 h) the glass e quartballooth d z an nshiel d surroundin e canath g l mus e cleaneb t r changedo d .

For cleaning, a solution consisting of 80 % HF (40 %) and 20 % HNO3 (100 %) is recommended e layeTh . r deposited froe inneth m r surface will disappear within 10-15 minutee dischargth f i s e balloo s i rotaten d aroun s axiit d s horizontalln i y the presence of a few mm thick layer of the solvent. After cleaning, the balloon should be washed carefully with distilled water. The same procedure is required to clean the quartz sleeve. The metal extractor tips ( soldered into the ion source balloons) e etche shoulb + HNC t F dno dH s wit e solutionth h . e brownisTh h deposi - carbot n laye a verl vapours oi i y froe adhe- th mr - sive layer on the surfaces of the glass, and it can be removed only by mechanical procedure. The aluminium extractor tip (canal) should be changed in every case. The mechanical cleaning of its surface is simpler but the diameter of the hole must remain. Cleaning the dirty glass parts of the HF ion source usually results in a shorter lifetime of the component. Use the "cleaned" balloons and quartz sleeves onl n i cas yf emergenco e y because, after cleaning, they will always giv a lowee r beam current. Before starting a long irradiation (> 20 hours) always change the extractor tip, quartz sleev d Pyrean e x bottle.

4 HIG5. H FREQUENCY OSCILLATORS e higTh h frequency oscillator e 15-10 th z rang e n e i couple0MH ar s eth o t d discharge balloon inductively, while at frequencies >100 MHz they are coupled capacitively. The high frequency oscillators are fed by a ca. 500-1500 V, 0.5 A anode power supply e activTh . e componenF triode H e oscillato th n a , f o tet t s i -r rode or pentode type transmitter electron valve. The circuit diagram of a simple but reliable oscillator of about 100 W output power is shown in Fig.28. The circuit is a three point type oscillator and it oscillates at about 27 MHz frequency. The coil L^ is the inductive coupling coil of the oscillator, surrounding the Pyrex discharge balloon of the ion source. The 5 nF capacitor and the coil 1^2 protec e anodth t e power supply agains e higth t h e frequencyth e d griTh an d . respectivelQ 0 22 e cathod ar higd ) yan h epowe Q k resistor 5 r7. (ca( s . 20-3) W 0

52 !KOOV max)

220 — CD— I—-— -- ,—-ft

———e -ft sourcen io Fig.2W 0 oscillator8 10 - Circuit z MH diagram7 2 a f o (V1 =OT100or GL8005;Lj 50mm= dia, CuAg;windings,dia 2x5 4mm dia,mm 18 = L? Cu; dia windings, 12 dia,x mm mm 3 0.8 18 Ly= 40 windings, 1.0 mm dia Cu)

wire wound resistors. The 60 pF resonant circuit capacitor is ceramic high volt- capacitorF n 5 ag d e inductance-freean ar s tub F n e 1 type e th ; e disk type. e hige shapth Th hf o e frequence b observe n oscilloscopy a ma yy b d e probe placed in the vicinity of the oscillator. The bandwidth of the oscilloscope shoult leasa MHz0 e 3 tb d . To pick up the signal from the oscillator, a simple wire, as aerial, or e coaxiath f o l d cablsmalen e usedee b e alway lth inpu y Us coil.e n 1:1ma o th st 0 probe at the AC input of the oscilloscope to avoid accidental damage to the vertical input. If the shape of the oscillation is not sinusoidal and shows some saturation, this is an indication of the abnormal operation point of the oscillator valvee gri th e tubd th r : o resistoe r shoul e changedb d . A push-pull type oscillator of 100 MHz frequency, using two triodes, is show n Fig.29i n e outpuTh . t powe s aboui r e th circuit d 200-25an t form , W 0a s grounded-grid, symmetrical oscillator. The metal ceramic plane valves are vented a bnorma y " cabinetsy b forcen 19 (lik fa r le d eai thee d th placean ar )fan y n i s d in two metal tubes. The feedback trimmers are air insulated disk type; the grid- to-cathode resistors are 50 W wire-wound. The current consumption of this oscillato s aboui r e t maximuth 200-40t a A m0m anode voltage e advantagTh . f o ethi s oscillato e easd th an distortion-fre ys i r e output power e changeb , n whicca dh

53 *"+

Fig.29 Push-pull oscillator for HF ion sourceswith 200 W output power dia,mm 24 CuAg; = dia dia, windings, mm 5.5 Ly mm 40 = 3.0 (Li 4 x 20 windings, 0.6 mm dia Cu; L^= 20 mm dia, 40 windings, 0.7mm dia Cu; V = V= GS-90B or GU-6B or TL2/300)

-*+UQ( 1100V max.) !0 Coupling ring 5T SP 5p SF0 H35 12V J7~ ID LS sg BU208

CGND

*UF Fig.30 Capacitively coupled HF oscillator with optical pulser (Lj = 100 mm dia, 2 windings, 4 mm dia CuAg; Ly= 8 mm dia, 50 windings, Vj=QQECu; dia,dia windings, mm 11 06/40)8 mm = 1.0 L^ Cu; dia 0.8mm

54 by the anode voltage. The diameter of the coil L. fits that of the ion source balloon used. This oscillato s utilizei re KFK th t Ida neutron generator. A capacitively coupled oscillatoz frequencMH f o abou 0 rs i utilizey20 t n i d the SAMES neutron generators. This push-pull oscillato s constructei e r th n o d a doublbasi f o s e tetrode e outpuTh . t n sourcpoweio s e i coupleer th bottlo t d e e doublth f o metao ebtw ye tetrodlus ringse eTh . allow n easa s y pulsing possibilit y groundinb y e suppressoth g r grid throug a h high voltage, high frequency switching transistor, whic s i h n drive a C modoptica D y b ne e l th cable r Fo . neutron generator, disconnect the wire between the suppressor grid to the collector of the BU 208A transistor and ground the collector of the transistor. The optical coupling between the ground and the HV terminal may be light cable or (screened) Perspex rod. The circuit diagram of the oscillator and the pulsing unit's receive s i showr n i Fig.30n e transmitteTh .e ligha th e tf b o r pulse y ma r commercial pulse generator-fed infrared LED.

5.4.1 Troubleshootin f higgo h frequency oscillators The discharge tube (ion source balloon) is made of glass, so the colour of the HF discharge gives information on: - the vacuum condition of the system s supple conditioga th e - yth f no - the power of the HF oscillator. The HF oscillator should be tested if the discharge in the ion source bottle is poo r totallo r ye vacuu absenth s t supplga m bu te th systey d operatman e correctly e operatio Th e oscillato. e th detecte b a f neo o n y b ca nr d a lamtub r o ep light. If a small neon lamp, fixed on a 20-30 cm long Perspex (or any other iso- s i placelator d ro d) neae oscillatoth r e coil)th e rth e o ,anodt (closth r o o t ee lamp should ligh F oscillatot H whe e th n r operates (see Fig.31) tubA . e light (18- 40 W of power) already indicates the output power of the oscillator placed in the F coiH l e vicinitth (Fig.32) f o ye tubTh .e ligh t e oscillatoth wil f li lighp ru t works. The intensity of the tube light depends on the output power of the oscillator.

Neon lamp

\ Perspex rod 20-2Scm)

Fig.3 e neonTh 1 lamp indicator oscillationF H f o

55 Brighter light u\U

Fig.32 The use of tube light for testing an HF oscillator

Incandescent lamp

n sourcIo e

Fig.33 Qualitative measurement e outputth f o power incandescenty b lamp

e incandescenTh t lamp also give a qualitativs e indicatio e outputh f o nt power of the high frequency oscillator: for higher power the light is brighter (Fig.33). If the tube light or neon lamp lights up in the vicinity of the oscillator bue discharg tth o lighn thern i s ti ee tube, this indicates that d vacuuba s i m (high pressure) in the whole system. n I normal operating condition e lighth s t emitted froe ballooth m s i vivin d reddish-violet or bright pink. A light blue colour indicates bad vacuum in the discharge balloon or that the D~ gas is dirty, containing air.

For testing the neutron generators the use of H2 gas instead of deuterium is recommended e differencTh . s i onle elac f th yneutroo k n productio e blinth dn o n target or constructional parts, which is an advantage when testing and repairing

56 neutron generators. (The bremsstrahlung still e exsists!dischargth f I )s i greye - ish-pin r e ligho extractiokth t d bluan ne current show a weas k plasman ca t i , happen that there is no more Y>2 ^ i bottle or that there are problems with n tne the pressure regulator valve (palladium leak, thermomechanical lear o needlk e a buil~ bottl D s n manometeri tha e evalve) th f I . , chece th pressure k th n i e bottle. ATTENTION l maintenanceAl : , troubleshootin terminaV H e d lth repaian g n o r mus e donb t e only after first groundinterminalV e H terminat th e no f th I g . s i l grounded, the 500-1000 V anode voltage can be lethal. The neon lamp and incandescent lamonly e useyb ma p d wit a hsufficien t long isolator (Perspex) rod. Whe e indicator t th nsho no F oscillation wH F o oscillatod H s e th , r shoule b d checked. Test the following components and parts: e electromechanicath f o l Al 1) l connection e oscillatorth f o s . 2) At power on, check: - Filament voltage and current - Anode voltage and current - Suppressor voltage and current (if it exists). 3) At power off, check: - Conductivity betwee e contactsnth , coild alon an e scondenser gth s - Resistanc e resistorsth f o e . e operatioTh 4) f vacuuo n m valve(s e oscillatorth f o ) . 5) The operation of the anode power supply.

For capacitively coupled ion sources of some 100 MHz, the twin tetrode QQE 06/4 s usedi 0 . o anodeBecaustw e connectee ar sth ea thick y b d, rectangular, silver plated coppe o anodr whictw ba n rspa e ca hth n pins e glass-metath , l solder- e anodeth n sometimeinf pi o g s cracks , e leadinth exposur e o t th vacuu gf o e m valve to air. If this happens, a larger white area can be observed on the glass e tubeth wal f ,o lwhic e s chemicai causehth y b d l reaction betwee e gette- th n ma r n e I airpush-pulth teria . d F oscillatoran H l l - uses d usuall n i inductively y coupled ion sources - the changing of the two tubes should be carried out simultaneously A tes. t circui r determinatiofo t e operationath f o n l characteristicf o s the vacuum tubes is recommended: the lack of oscillation in the case of a push- pull oscillator may have been due to the poor characteristics ( emission) of one or both tubes. For the repair of HF oscillators always use the same quality components: - Silver coated electromechanical component e useb shoul t d no d withou a reasot n (e.gF coilH . , anode-, grid- connections, capacitor holder).

57 - The wire-wound resistors should be changed to the same size and value (in ohms and watts) components. e condenserTh - F oscillatorH s e th use n si d shoul e b dinductance-fre e (mostly mica higd an )h voltage capacitors. e trimmeTh - r condenser e usuall ar r sinsulated ai y e actuaTh . l d sizdetailan e s related to the mechanical mounting can be copied from the original solution. - The some 100 W power of the HF oscillator is high; therefore, for the measure- n operatina mente electronin us o F st oscillato H gno co d (i.er . digital) multimeters. The simple ANALOG MULTIMETERS are recommended.

5 PENNIN5. SOURCEN GIO S The cold cathode arc discharge ion sources with oscillating electrons in the magnetic field are known as PIG-type sources. The operating principle of this sourc s demonstratej i eplanK d e an e samelectrode th , K n t eFig.34i a d e e Th ar .s negative potentia e ring-shapeth o t l R anodA dradiusf o e eTh electri . c field drives the electrons towards the anode. If a magnetic field B with a direction indicated in the figure is also present, the electrons will move along expanding helical trajectories. The maximum radius of the electron trajectory re,maorriox at given geometrical arrangemenn o s wel a s ta , l B depende magnitude d th an n o sE f o s the direction of the electron velocity to the magnetic field. With a sufficiently high B value, the r < R requirement can be assured, and thus the electrons CuLlLcLA from K, will continue to proceed towards the K-. electrode. The negative potential K electrod e th of e prevent e electronth s - whics h los a par tf theio t r energn i y

-o-

Fig.34 Operating principle e Pennine-typeth f o sourcen io

58 elastic and inelastic collisions - from reaching the surface of Kj, and they will return ~ towardelectrodeK e th s , elec. 1C Thi -d s an proces j K s repeatei se th t a d trodes and the electrons create an intense gas plasma by further collisions. If the energy of electrons decreases below a critical value, they can strike the anode. The number of oscillations n of the primary electrons can be calculated from the x mean freee path covered by the electrons in the oscillation [44]:

e distancth wher s i d e betwee e cathodnth ee anodsurfacth d e an ering . e value - th probabilitieTh f in o e pressurdependd th an e th p n o esf o s elastic (ai-c and elastic a)e collisions:

The number of the collisions will decrease with increasing pressure in the ion source. The mean free path x can vary for different gases. For example, for electrons of 100 eV energy in H~ gas, x = 500 cm (n = 20); in He gas x = 800 cm (n = 32); and in N~ gas x = 140 cm (n = 6). e Pennin th e cas f th o e n gI source e extracteth , d beam contains only abou0 4 t to 60 % of atomic ions as a maximum. In spite of the low atomic ion fraction, PIG ion sources are often applied in neutron generators, even for deuteron energies lower keV0 tha20 n , becaus f o theie r simple construction, cooling, power supply system, inlet gas-flow requirement, and long operating lifetime. The U a is usually about 5-10 kV; the I is in the range of mA. The current from the ion source e regulateb e e densitplasmn th th ca y f changin b o d- ya e anodth g e voltagee Th . Penning ion sources are usually equipped with permanent magnets, so they are idea sourcen io l r sealefo s d tube neutron generators. The extraction of the ions from the discharge in the Penning ion sources is usually mada diaphragm-typ by e e extraction system. e self-maintaininTh g discharg s i influencee e cathodth y b de material through e secondarth y electron emission induce y positivb d e ions strikin e surfaceth g . Al, Mg, and Be cathodes coated with oxides require low ignition voltage (U- = 300 to 400 V). The oxide layer can be regenerated by operating the ion source for 10 to 30 min with oxygen gas. The lifetime of oxide cathodes substantially increases if 2 to 10 % oxygen gas is admitted to the ionized gas. Low ignition voltage (400 i cathodesT d an s , requirei . U ) Metal , V Fe 0 e dcas sth f t80 o o als e suitabl n i o e

59 Vacuum gauge Hi

Anode Magnets Magnetic Ferrous circuit cathode Annular permanent magnet Nonmagnetic Pyrex Cathodes (stainless steel) tube Extracting electrode Anode

Base Rubber seal

Fig.35 Typical Penning ion source - schematic (left) and actual scheme (right) for use as cathode materials with high U- values are as follows [44]: Ni, Zn, Al, Cu, C, W, Mo and Ta, for which the ignition voltage is between 3600 and 1700 V. n mosI t cases a cathodeT , d alue an applie= 170 ar , s ; - however-Al 0V) (U , d Mg , minium alloys have also been used as cold cathode material because of their high sputtering thresholds [46]. A col dn sourcio cathod G e PI ewit h axial extractio s i shown n i Fig.35n A . 9 magnetic field of about 0.03 to 0.7 W/m , supplied either by a permanent magnet a solenoid y ob r s i applie, d betwee e electrodesG th sourcen PI e operate ar s Th . - s consumpga ing t a wits pressure a ga -e hP discharg th 5 n 2. i s o t e volum1 0. f o e h / NTP m c typicaA . 0 G sourc 50 PI n operatl o ca t e 0 2 etio f continuouslo n y over 200 h at about 2 mA beam current without the need for replacing the cathode or anode. A typical HV power supply feeding a Penning ion source is shown Fig.36. potentialV k n sourc io 0 Sinc1 e e th ,~ e requirethi sU higa s h voltag s i usuale - ly produced by a voltage doubling circuit.

5.5.1 Troubleshooting of Penning ion sources e problemTh s causing improper operatio e divide b f thin sourco n io ns ca d e into three groups: magnetic, mechanical, electric. The magnetic problems are caused by the decreased field strength of the permanent magne e o e thaelectros radiuth tth tf o sn trajectory increase - sresultin g in a lower discharge current I than during normal operation at the same U voltage. Similarly, every shunting of the magnetic circuit will have similar effects.

-CZh -0 + Penning Primary o- ion source from variac 0%

Fig.3 voltage-doublingA 6 powerV H supplyPenninga f o sourcen io

61 The mechanical problems may be caused by: - contaminated cathode surfac l e vapouvacuuoi eth e (dun i th rm eo t system), - alteration of the original anode-ring-cathode geometry, - a change in the geometry or surface properties of the extraction slit. e electricaTh l e problemb cause y n theb dca s: V poweH - r supply: transformer, rectifiers, condensers, limiting resistor RL- e Penninth n sourc n supplconnectorso io V V gH H r - o ee y (thth V n eo H : feedthrough), - HV cable between the power supply and the ion source.

Testing the magnetic circuit needs some magnetic field measuring instrument, e fielbu th n tsourcd io e strengttestee b e th t musdno n i hdirectly t . Some iron objects (bolts, screws, etc.) accidentally bypassing the magnetic circuits can be observed visually, while the magnetic short circuits inside the ion source can e founb d only after dismantlin e sourcegth . If the source is opened, carefully clean the ceramic HV leadthrough, the e anodcathodth d e e an removel ee surfacesb cathod oi deposi th n e ca n dTh eo .t only b ysandpape d polishinan r g paper. After dryin e th gsurfaces , clean them with organic solvents. When the components are dry, assemble the ion source. If the extraction slit becomes bored out by backstreaming secondary electrons, replac e plasmth e a extractio e w orific th s ne p usewit n wa cu i ea ndh t i inser f (i t origina source)n io l . Tes e mechanicath t l positio e extractioth f o n n electrode. The electric test of the power supply and the ion source requires a high voltage voltmeter (up to 30 kV) and a MQ meter, using a couple of hundred volts. V transformerH Tese th e te condensersrectifierth th , d an e scomponent Th . s can be checked with a normal, everyday multimeter through an HV probe. The resistors Rr and Rwj can be tested similarly with the same multimeters. Test the HV e th connector cabld an e s under working condition e s operationwella sTh . e th , e insulatiotestee b sourc th e t HVPn a atmospherith dd ca ef an So n c pressuref I . the system seems to be working electrically normally the fault should be searched for elsewhere, for example in the gas supply or in the vacuum system.

62 0 DEUTERIU6. M LEAKS

The ratio of atomic to molecular ions depends on the operating time of the ion source- e en pressur e th discharg, th s n ga i e e e e puritth volumeth f o d y an , terin e plasmath g carefullA . y designed pressure regulator mus e b tapplie o t d guarantee the stable flow rate and the purity of the gas during long operation timee threTh . e most commonly used deuterium leak n neutroi s n generator e palar s- ladium leak, thermomechanical leak and the needle valve. e palladiuTh e e mregulatioutilizeb th r n leafo ca dkf hydrogeo n s (isoga n - topes) flow which discriminates against other e gasesthermomechanicaTh . l leak and the needle valve are suitable for the regulation of all types of gases. As the neutron generator uses deuterium for feeding the ion sources, all of these e n sourcehigth io he e n vicinite o th founth voltagb s f n leaky i o dy ma es terminal.

PALLADIUE TH 1 6. M LEAK e s besregulatorth ga tf e "palladiuo th e s i s On m leak", constructe f o dpal - ladium-silver alloy [2]. Palladiu s temperature-dependenha m t permeabilite th o t y passage of hydrogen isotopes and discriminates against contaminant gases between 300 and 400°C. Characteristics of the I-Pd system make it possible to supply the n sourceio G loswPI d witintensite an necessarth h F R y y amoun f hydrogeo t r o n •5 deuteriu h / NTP m mc .0 gase a consumptio2 Permeatio t o a st 0 1 f f hydrogeo o n n throug a palladiuhwalm m l 1 thickdiameterlength0. m m m m m d -2 tub0 an ,f 10 o e, ness is about 5 cm• 5 /h at 400°C, which value is enhanced significantly by the use of Pd-Ag alloye ratf Th permeatioo e. f hydrogeo n n through palladiu s mi highe r than that of deuterium, the ratio Pj-j/^D ranging from about 1.2 to 2.5 depending on temperatur d metae conditioP th e ld th [47,48]an ef o n . The usual setup of a palladium leak consists of a palladium tube closed at one end which is heated either indirectly by a separate heater spiral or directly by resistance heating e palladiuTh . m meta s speciaha l l features: aboue litron te oe absorbef b e gra hydrogen on f palladiumm ca o y o usua b ds tw ga n l e construcTh . - tions of Pd leaks are shown in Fig.37. The indirectly heated palladium leak manufactured of glass is usually connected to heavy water electrolyzers. The directly heated Pd tube is placed inside a high pressur tank~ D e . £ A common disavantage of palladium leaks is ageing. After about 150-200 h o- perating time the transparency of the palladium will decrease at a given tempera-

ture and the leak gives a lower D flowrate. If this takes place the heating of 2

63 sourcn io e eth 0o t 2 (pressure ~1(T1-1(r2mbar)

Pd tube

Thin glass insulator

Heater spiral Up -6-10V AC/DC Up~1-2A

Manometer Outlet ( n Atmospheric pressure Filling inlet sourcn io e eth o t I /^~~*\ / . (Ibar) valve " p (p~10~1-10~2mbar) ^ glass cone joint

Insulated 02 tank feed -through p>1bar — Pd tube V I - 5 0. U p~ C S ——— D tank Ip -10-12A 2 (pressures Ibar) Fig.37 Directly d indirectlyan heated leaksd P

the Pd tube should be increased. As the heating in both cases (indirect and direct s i resistive) e e increaseb heatinth , n ca y gchanginb d e heatinth g g voltage or by reduction of the resistance of the heater. In the case of an indirectly heated Pd tube, the heater spiral should be shortened (cutting off 10-20 % of length). In direct heating, the palladium tube voltagw lo e e th e loa powef th itselo ds i fr supply (transformer) d tubP - re ee th : sistance can be decreased by sliding the isolated F, contact towards the lid of

2 thtankD e . Shortenin e originath g l tube lengta successfu s y 20-5i b h % 0 l solu- tior improvemenfo n e maximuth f o t ~ a transparencgive m D d leakt P A n e . th f o y

potential difference the energy is dissipated in the resistor at a rate I^R0, and so, if R = R /2, then I = 21 , i.e. the energy is doubled. e glass-warTh e palladium leaky breakma sf I , the.lowe do y r pressur - evacu - n sourcio r - o esid em u shoul e testeb d y usuab d l vacuum testin f i therge ar e improper (gray coloured) gas discharges in the HF ion sources. The vacuum testing of directly heated palladium leaks is carried out in a similar way. In the case n overpressuroa f e tank th e th e sealintann ,i th e k f greateo g~ D r r ba tha1 n should be adequate. All of the seals at the connection of the manometer, the tank closin e heateg th valv d r an eelectrod e shoul e checkeb d d regularly.

64 In the case of defective operation of the palladium leak (if the operation of the ion source shows some problems related to the gas supply) test the following:

- The colour of the discharge of the HF ion source: if the discharge is weak and is slightly pink, the gas supply is not adequate - e.g., the electrolyzer doet worno s- kex properly s i e s fillinth ga , 2 g Y> valv e s i leakinth e d an g hausted or the heating power is not enough (contact failure). - The palladium (tube) leaks: this can be observed if the discharge in the ion source bottle canno e extinguisheb t y switchinb e dfilament th f of gA hol .n i e e thith n wall palladiu e detecteb m n y usuatubb ca de l vacuum leak testing methods, and the repair requires soldering with silver at the workshop of a precision goldsmit f thero replacemenhi n s i e e defectivth r . fo t Pd e

6.2 THE THERMOMECHANICAL LEAK AND THE NEEDLE VALVE Thes o mechanicatw e l leaks havo majotw e r problems: then faio t closca yl e properl r theo y y canno e openedb t . Their repair require a wels l equipped precision mechanical workshop and an experienced tradesman. The electrical troubleshooting of a thermomechanical leak is similar to that of a palladium leak. The high flow rate in the case of a closed valve can be caused by stain or dirt in the canal of the leak. The dismantling, cleaning and assembly should be carried out with precautions. Anybody who is not familiar with the structure of these leaks must not open them for cleaning or repair. The best solution for the repair of these elements is to send them back to the manufacturer for repair and readjustment.

6.2.e thermomechanicaTh 1 l leak valve The operation of the thermomechanical leak is based on the differential thermal expansion coefficients of different materials. The leak itself consists of a metal ball held by a ceramic rod against a precision seat with an orifice. The orifice and the ball form the valve seal. The outer cylinder, the container wal e le directlceramiar th aroun d e d balyro th can d l heated electricallye Th . expansio e cylindeth f o ns greatei r r tha e ncerami th tha f o t ce th rodsea o ls , between the ball and its seat will be gradually opened. The leak rate of the thermomechanical leak valve depends on the temperature e outeth o f e rleak th wal f , o e electrili.e th n .o c current user heatinfo d e th g wall. In the case of a heating current switch-off or power failure, the ball will be reseated to the seat and the valve closes. In the case of an electric power cut-off, this valve will protect the neutron generator against exposure to atmo-

65 Gas outlet (ion source) metal ball ceramic rod

Direct heated metal cylinder

Good electrical contact 0-1 V AC

Fig.38 Schematic diagram thermomechanicala f o leak valve

_ E E «- a ^^ -12

-8

—r~ (AC)V m [ ] 0 3 10 0 2 Fig.39 The leak rate vs heating voltage of a thermomechanical leak

spheric or D~ pressure. The operational principle of a thermomechanical leak is £ show Fig.38n ni . e thermomechanicaTh l a precislea s i k e componen d thereforan t e th concure - rent exposur o corrosivt e e gase d atmospherian s c moistur y damagma ee valveth e A . typical leak rate (torr 1/s) or gas consumption (bar ml/h) is shown in Fig.39 as a functio e heatinth f o ng voltage outeth f r o e a thermomechanica wal f o l l valve.

6.22 Maintenanc troubleshootind an e f thermomechanicago l leaks e thermomechanicaTh l leak valves shoul e connecteb d e vacuuth o t dm system (ion source) in a correct way relative to the direction of the gas flow. The di-

66 s flo ga s usualli w e rectioth y f o nindicate e valveth n .o d Flus e th hthermome - chanical leak valve wity nitroge dr he cas th f r uncertaiotheo e(o y ngasn i dr )r n operation. Chec e electrith k c contacte leath k e f heatino th svalve s A g. current of the cylinder is usually higher than 10 A, the electrical connections should be checke d cleanean d d regularly s accuratA . currenC A e t measuremen e rangth n ei t of several tens of amperes is not simple (it can be done by a digital clampmeter C heatinfoA r g current), calibrat e heatinth e g voltage alon e th gcylinde r while it is working properly. The voltmeter contacts should be separately connected to the d cylindercontacba a n casi tf : o ebetwee e filamenth n t transformee th d an r leak body, this voltmeter will indicat e conditioth e econtactsth f o n . Opening the welded body may kill the whole leak valve. In case of lack of heating test the following: e primarTh - e ystepdow th sid f o e n transformer (voltag d currentan e ) e secondarTh - y voltage acros e leath sk valv t e nominashoul(i eth e b d l value) e voltagTh - e drop acros e secondarth s y d coithermomechanicaan l l leak contacts. There should be a small (~ 1-10 mV) voltage drop along a good contact. e contactIth f e defectivear s , clean the tesd man t again. If there is a grayish-pink discharge in the RF ion source, check the vacuum and deuterium connections. If these connections are in good condition, there may e somb e cracks alone leak e bodth th gf . o y

6.2.3 Needle valves Needle valves are used for controlled admittance of fine gas flows into the ion sources. They can be operated either manually or electromechanically. Re- liabl d populaan e rd needlan I H e e BALZERn source0 valveth io 01 r e ar fo N s SEV EDWARD typesK variable 10 Th . V SeFC leak needle valv f VARIAeo N include movsa - able piston witn opticalla h y flat sapphire which form a variabls e seal completely free from friction, seizing and shear. All these valves can be used for very s n finsourcesstreamio ga s e pressur e r e finga adjustmenfo th Th s e. e n th i e f o t adjustment of the stream is performed by adjustment of the position of the needle e osapphirth r e piston e movemenTh .e needlth s i controllef eo t d througa h threaded a threadeshaf r o t d shaft-and-lever mechanism havin a gmechanica l advantage of up to 10,000 to 1 [49]. I valvH a BALZERn exchangeable0 a er 01 Fo N SEV , easy-to-clean dust filtes i r fitted into the lateral small flange port by which the valve can be connected to the gas bottle of the ion source. The main parts of the valve are made of stainless stee r rustprooo l f e materialvalvth ed an sea ,s i madt f a closeleado e n I .d position the valve needle is forced into the preformed conical seat by a spring

67 17-

1 HOUSING 11 GUARD RING 2 KNOB 12 LOCKING WASHER 3 SCALE DRUM 13 NEOPRENE O-RING 4 SOCKET 14 VITON O-RING 5 NEEDLE 15 SEAL 6 FILTER 6 SCRE1 W 7 NUT 17 COVER 8 SPRING 18 THREADED PIN 9 COVER 19 THREADED PIN 10 BALL BEARING Fig.4 typicalA 0 needle valve

SCALE TURNSCStO n • • • 10 i 1- 0 10 =b. — (~r- E VN OK H2- 103 10 EVN010H1 < - -s ^,'? Z ^ 10 1 f^iff ^•£;=• •*^ X* 101 0.1 /- — / mO 1(T3 10~2 10" 10 10° GAS FLOW (mb-l/s) FOR AIR

Fig.41 Gas flow rate curves for BALZERS needle valves [49] (Sk scale= marksmicrometer-likethe of scale drum n- number of turns of flow rate adjusting knob)

68 which ensure a constans t closing pressure. Over-tightenin e valvth g e during clos- y damaginma e gneedl th e e and/o e valvth r e seat. Moving e directioaxiallon n i y n the needle is coupled to the actuating knob over a ball bearing. The fine thread e knooth f b ensures thae needlth t s i e lifted reproducibly froe seate th mTh . scale ring and actuating knob are coupled to each other for easy adjustment. This allows a simple setting of the opening point at any time and marking a specific gas flow as 0-point for accurate reproducibility. Usually, the adjustment of this valve tends fro a mmaximu a minimum o t 0 mbaconductanc s 13 f mo 1/ r. ca f o e _Q 1x10" mbar 1/s NTP (i.e. the leak rate in closed position). The construction of the needle valve is shown in Fig. 40 [49]. The gas flow rate - the air admittance - of the EVN 010 HI (and EVN 010 0 H2) is shown in Fig.41. The conductance of a needle valve depends on the viscosity of the gas to be admitted into a vacuum chamber. If gas other than air is used, the original calibration curve can be utilized only by taking into account the viscosity of the given gas related to air. Table 13 shows the viscosity of the most frequently used gases [49].

Tabl. 13 e Viscosit e mosth f t o yfrequentl y used gases

Gas r\ Viscosity [Poise] K = r\ air / rj gas

Air 180 1 Nitrogen 173 1.04 Carbon monoxide 172 1.05 Oxygen 200 0.9 Carbon dioxide 145 1.24 Hydrogen 87 2.07 Water vapor 93 1.97 Helium 194 0.93 Argon 220 0.82 Crypton 246 0.73 Xenon 225 0.8

a laminae needlr th s flo Fo n ga i rwe valve e followinth , g e b equatio n ca n used for correction of the flow rate Q of the admitted gas:

= *air / V ^ga ^aix Qsx K air = r ImbaW r '/*]

69 Example 1: a neutro r Fo n generator ,n sourc io wher y deuteriue b e th ed shoulfe s me ga b d ), the preset conductance of 10" mbar 1/s for air (Sk scale mark of needle valve EVN 010 HI is about 13, based on Fig. 41) will be higher than for air owing to the K value of hydrogen (2.07). As Fig.41 shows the gas flow is higher than 0.001 mbar 1/s, a flowrate below this t reliabl se a vacuu f e i b yvalu mn s ca en thegaug floga e ca b nwe s used i e Th . calculated on the basis of

[mbaS x p r A 1/s = ] Q (18)

where the gas flowrate is Q [mbar 1/s]. the pressure change in the vacuum chamber is Ap [ mbar ] and the pumping speed of the vacuum pump is S [ 1/s ].

Example 2: e pressur th e Ivacuuf th n i em syste s mi increase d fro a mvalu f 2x10o e o t " 3x10 " e pumpinth mba d an rg 1/se 0 valuwil: Q speeth 10 ,be f l o es i d

= 10" 0 4= 10 10" [mbax 6Q r 1/s] (19)

t s flowse relativel Ga e sb n belo" yca mba 10 s weasil1/ r y meanb ya f o s vacuum gauge. Experienc s showha e n tha a " tmba lowe10 s 1/ r. r ca limi f o t can be reached, but that the long term stability of the valve begins to decline at gas flowrates below 10" mbar 1/s.

6.2.4 Maintenance of needle valves Needl et requir valveno y maintenance an o d es n i e coursgeneralth , en I . of time, however, some contamination effects will become apparent. Very fine dust particle n ca stilsl pass e throug th e inlefilte th f e o needltth h r porf o et valve and they can settle in the gap between the needle and the needle seating. Adjustment of the low gas flows becomes difficult and the leak rate of the closed valve deteriorates. These symptoms indicate that the valve requires cleaning.

e dismantlinTh d cleaninan g g procedur f o eneedl e valve s i s describer fo d I typH BALZER e 0 e manufacturer'valv01 th n N i e SEV s Manual [49].

70 a) Dismantlin valve gth e (see Fig.40) ) fro e (9 actuatinm th p ca e -g th Tak knof ) of eb(2 - Remove locking washer (12) from spindle needl) (5 e - Unscrew knob (2) with scale drum (3) counterclockwise from sleeve-socket (4) - Remove screws (16 d sleeve-socke)an ) (4 t - t Pulcarefullou l y spindl ) togethe(5 e r K (shapewit d h an sprin) ) rin(8 gg (14) from valve body (1) - Undo the screw (7) and remove filter (6).

b) Cleanin e dismantlegth d valve parts Wash needle valv ) th e seatinghol(5 th een i e d bod ) ,witan (1 y filte) h (6 r chloroethan r o similare , usin a gsof t brush. Carefully remov y depositan e s from the needle usin a gsof t too f wooo lr plastico d e needl.th The b eru n vigorously but carefully with a cloth drenched in solvent cleaner. Make sure the needle surfac s cleai ed smoothnan .

IMPORTANT NOTE: DO NOT BEND THE NEEDLE DURING THE CLEANING !

Rinse all parts in alcohol or acetone (in an ultrasonic bath if possible) and dry them. Blow out the seating with clean compressed air or hair dryer. Be sure thae seatinth t g surfac s i score-free d smoothan e . Wip e holeth d eseal an s - ing surface d valv f an o sleev se) (4 bode) wit(1 y h lint-free cloth drenchen i d alcohol, remove all dirt and dust. Finally rinse O-ring (15) and K-ring in alcohol or acetone (if possible in an ultrasonic bath).

c) Reassemblin e valvgth e Slightly grease O-ring (14 d K-rin)an g (15) with Flombin greas r (o silicoe n grease slidd an )e them onto spindle (5). Be sure to protect the needle from contact with lubricant ! Carefully insert spindle needle (5) with washer under the spring into the valve bod ) unti(1 y l needl ee seatin restth n i sg - Slide spring (8) onto spindle (5) - Fasten sleeve (4) to valve body (1) using screw (16) Screw actuating kno ) wit(2 b h scale dru ) clockwism(3 e onto sleev) (4 untie l groove of spindle (5) protrudes clearly over ball bearing (10) - Attach locking washer (12) - Close the actuating knob (2) with cap (9).

71 d) Adjustin e spindlgth e clearance - Take off the cover-cap (9) from the actuating knob (2) - Remove locking washer (12), guard ring (11) and ball bearing (10) - Adjust clearance with threaded pin (21) - Assemble in reverse sequence. e) Testin e needlgth e valve after cleaning Open and close the valve about 10 times to compensate for any rough spots. The leak-tight valves should fulfill two requirements: - Leak tightness of the valve seat in closed position (needle/seating) - Leak tightness between valve interio d valvan r e surrounding.

Tightness of the second type is less important because such leaks merely result in gas losses to the atmosphere if the needle valve is used to regulate the gas consumption of ion sources from gas tanks of more than 1 bar pressure. If the valve operates correctly it must be possible to alter the gas flow smoothly, withou te directioe th stepsrotatioth f I f .o s i nchangen e valvth d e must react immediately e e valueth calibratioTh f .o s n curve r 1 fo 0.0s0. 1r o mbar 1/s should be reproducible within about 10%. f) Leak testin e valvth ge with vacuum gauge The leak-tightness of a valve can be checked with the vacuum gauge which is normally installed in any vacuum system. - Valve seat: Evacuat e closeth e d needl pressurw lo e d a closvalve an eo th t ee gas inlet flange wit a hblankin g plate. Ope e valvth n e slowly about five turno t s evacuate alse valvth o e interior. Clos e th valved waian et untie pressurth l e stabilizes. Remov e blankinth e g plate abruptl d observan y e meterth e f I ther. s i e a leak at the valve seat there will be a pressure rise in the vacuum chamber which will be indicated by the meter. - Valve interior: The valve interior is sealed by a K-shaped ring, the K-ring. Again, the gas inlet flange must be closed by a blanking plate and the valve openede valvth f eI . interio s i r leak-tight e samth , e pressure e musb reachet d irrespective of whether the valve is open or closed. As mentioned above, a leaking valve interior only means a certain loss of the gas admitted. Compared e resultwitth h s obtainable with sensitive leak detectors, this testing method yields a relatively low detection limit. For qualitative testing this is normally sufficient because leaks, which would be detrimental to the vacuum e detecteb system n ca ,d a easilvacuu y b y m gauge. Leaking valve seatn ca s only be repaired by the manufacturers. Leaks in the valve interior are mainly cause contaminatey b d r damageo d d seals.

72 g) Testing leaks with halogen or helium detector - Valve seat: Connect the closed valve to the leak detector and admit test gas into the gas inlet port via a fine (dust) filter. - Valve interior s :inle ga Cove te porth r t wit a blankinh g platd slowlan e y open the valve about five turns. This also evacuates the valve interior. Now, blow the outside of the valve with test gas. n e i goovalvth Ifs i de condition e leak-tightnesth , s values must e reacth h technical data. Similar haloge r heliuo n m leak testing procedure e useb n i dn ca s thermomechanical valves.

6.3 CALIBRATION OF LEAK VALVES: GAS CONSUMPTION MEASUREMENT SOURCEN IO F SO S e th f caso abnorman eI l operatio f o n leak valves, their transmission shoul s consumptio n sourcee ga testedio b d e e s usualli Th th s. f o ny measurea n i d practical unit of ml/h NTP ( Normal Temperature and Pressure ), i.e. 20°C and 1 bar; the gas consumption unit is ml bar/hour, which can be converted into the usual flow rate, pumping spees consumptioF ga H d e e mbar/ 1 unith Th f o f t o . sn or Penning ion sources at the neutron generators is in the range of 1 to 10 ml/h NTP, whil e duoplasmatronth e d othean s r high curren n sourceio t s consumo t 0 1 e 100 ml/h NTP gas. This means, the leaks should assure the gas flow into the ion sources in the range of 1 to 100 ml/hour NTP gas. The U shaped manometer - which e usuaith s l equipmen r determinatiofo t e pumpinth f o n ge vacuu th spee f o md pumps - can be used for the measurement of the gas consumption and the character- s adjustinga e th istic g e s f determinatioconsumptioo th ga valvess r e Fo . th f no n

Rubber Vacuum gas balloon system (H2.D2)

Fig.42 e measurementSetupth r fo of flow sourcen io rate f o leaks

73 ^^J shaped v— 1* >— ' glass tube mounted r^~^ vertically |

E- mm scale 1- = - Silico- l oi n

~ rr. L. j =j

Fig.43 The U shaped glass tube manometer used to determine the pumping speed and gas consumption of ion sources through the measurement of the flow rate

n sourcesio e oth a goof,s adjustinga d g shapeU valve e th ,d - manometeex n a d an r tra gas container (usually plastic balloon) is needed (Fig.42). e setu Th U shapep wite dth h glass tube filled e witcalth h r -silicofo l oi n ibratio s leak d pumpinga an f s o n g speed measurement s showi s Fig.43n i n . The valve connecting the upper part of the U shaped manometer serves to equaliz separato tw e pressur th ee th e n verticai e e manometerth l legf o s .

6.3.1 Measurement Regulate the gas leak for the required value of the gas flow. This can be detected, for example, by the optimal colour gas discharge in the HF ion source. e valv Th tubU e s openi e econnectine floth o leg Th f w. tw o s rate th ge (gas consumption of the ion source) can be measured in such a way that the volume of the gas entering froe plastimth c balloo s leanga k intvalve n th oio e e (intth o source a give n i )n time wil e determinedb l . Clos e valvth e e which tubU bypasse e e th e manometeleg f th so s d staran a r t stop watch. The gas entering into the leak valve will flow from the left leg of the U tube, causing a pressure drop against the right leg of the U tube. The lev- e silicoth l f wil oi o e nlef l th l te rist a hane d d sid decrease an righeth t ta e hand side. While waiting for an easily observable volume of gas consumption (say ) 1-stoe 3stoml th e psilico pth l e f watchlevefaloi th o nl f lI . too a ktim, At e the flow rate through the gas leak (the gas consumption of the ion source) will be

X = (20) 4At

74 where d is the inner diameter of the U shaped glass tube. If the diameter d and l levet timoi A n hours h i lee A change th flo e th ,wd th s measurean i erat m ec n i d s consumptio n ga sourc io ane e th de th wilf e giveo nb l n ml/i n h NTP. This unis i t suitabl r othefo e r purpose s wella s e : consumptionth base n o d e normath , l operation time of an ion source (neutron generator) can be calculated from the ^2 8^ container volume. The gas flow rate value may also be calculated for other pressures. In the a needlcas f o e e valv e adjustablth e e flow rat s i usualle y give t a lowen r pressures, e.g n i mbars. . This measurement make t i possibls o t chec ee characteth k - ristics of leak valves given in manufacturers' data sheets and their recalibra- tion after repair. U shape e dTh tube fille o t hald f height with s i silicoals a l ousefu oi n l toor determinatiofo l e pumpinth f o n g spee f o dvacuu m pumps. Pumping speed measurements will be found in Section 9, which describes the vacuum systems and vacuum component f neutroo s n generators.

75 7. DEUTERIUM ELECTROLYZERS

Deuterium gas needed for the operation of neutron generators is commercially available t manbu , y laboratorie o supplt s s producy ga thei ~ D er n acceltheiow r - erators e deuteriuTh . m r o consumptioPennin F R n sourc io ga s f i eo belon 0 1 w ml/h NTP, while a duoplasmatron needs about 20-100 ml/h. These gas consumptions cae satisfieb na simply b d e laboratory electrolyzer e e KFKToshibTh th . d I an a neutron generators even hav en electrolyzertheiow r s with palladiue mth lean o k HV terminasourcen io F supplo .H t l e th y The Toshiba neutron generator uses a heavy water electrolyzer of a few ml heavy water with a built-in palladium leak. The construction of this Pd leak with heavy water electrolyze s showi r Fig.4n ni 4 [50. ] e productioTh e nelectrolyze th rat e f adjusteb o e n y settinca b rd e th gcon - ductivity, the pH value, of the heavy water electrolyte. The solution is usually tabletH alkalineKO sr y puttino b inte : heav th D o a gcouply KO e waterf o th e f I . electrolyzer produces more deuteriu s tha e mga consumption sourceth nio e th , f o n the water level will sink in the central glass tube and the cathode contact will t offcu . e b This electrolyzer works normalle conductivitth e f i electrolytth y f o y s i e e th cath f o - " p ti rangecm e p A th " properly.u par Q f t o t2 se 0. , e i.eth .n i e electrodod s immersei e e electrolyteth n i d , whic n sometimeca h - de sH causKO a e posit formatio e th cathod n o n e surface. This insulating layen prevenca r e th t

o leat k

Joint

Cathod~ e Anode +

Mercury _^ cut

Electrolyzer

Fig.44 Heavy water electrofyzer of the TOSHIBA NT-200 neutron generator

76 a ® a ON 0 6 a OFF * Si

LED bar Sx1.2K

0 2 ELECTROLYZED R d LEAP KD AN

INFRA LIGHT TRANSMITTER

Fig.45 Optical fiber isolated leakd P control electrofyzerd an e th r fo deuterium terminal HV sources supplyion the of on

e electrith flof o wc current froe electrolyteth m , especiall e neutroth f i y n gen- eratof actio o a lons bee t r ha grnfo nou time. This disadvantag e simplth f o e electrolyzer s eliminatewa s y KFKb d I [51]n i thei U shaper d electrolyzer-Pd leak system y usinb o cathodes, tw g e uppeTh . r cathode (electrode No.2 in Fig.45) switches on a relay which remains switched on throug e loweth h r e cathodelectrolyzeth f o e r (electrode No. 3n i Fig.45) until the electrolyte level sinks under the level of the tip of the lower cathode. As

77 the electrolyte leves conga l e -e riseuppeth th s o t o rt agai e cathodp u du n - e e n sourcth e relaelectrolytio th sumptio e - yed th switchean f o en n leveo s l e operatinth sink o st agaige du conditionn . Using this hysteresis principlf o e the KFKI electrolyzer a heav, y water electrolyzer-Pd leak system with optica- fi l ber connection for the regulation of the deuterium flow into the ion sources has been constructed , PenninHF e s suppl.th g ga Thi f e o ysth e useb r syste fo dn mca and duoplasmatron ion sources. The circuit diagram of the system is also shown in Fig.45 [52]. e electrolyzeTh e r hysteresith work n o s s principle described abovee Th . d tubP s i controlleee heateth a switchin f y o b rd g (Darlington) power transistor. e on-ofTh e ftransisto th rati f o s oi controlle r d D phototransistothrougLE a h r optical link by an astable multivibrator at the ground potential. The optical connection can be a simple 3 mm diameter Perspex rod or plastic (insulating) fiber cable which endure e higth sh voltage differences betwee e electrolyzeth n n (o r the HV terminal) and the LED driving pulse generator (at the ground potential of the control desk). The effective heating current of the Pd leak is controlled by the duty-cycle of the astable multivibrator pulse generator. The heating current of the Pd leak is displayed on the electrolyzer by a simple linear 5 LED bar in-

dicator. Similarly2 D leve e le th th statu,"hysteresis f o s " switching relas i y LEDd re show a a gree . d y b nan n e electrolyzer-PTh d leak system e ar vers y economical solutione deuth r -fo s terium supply of a neutron generator. e deuteriuTh m suppl f commerciao y l neutron generator e solveb n dca s alsy b o separate heavy water electrolyzere neutroth e f sitth o en t a sgenerato r laboratories. An electrolyzer for filling the ion source deuterium vessels can be con- structee basi th f Fig.46 o sn o d . This electrolyze n producca r e 100-500 ml/P NT h deuterium gas with a simple (sometimes not water cooled) electrode stucture and e th n servdeuteriu ca s a e m e supplieneutroth r fo nr generato d othean r r utiliz- ers. The whole glass system of the electrolyzer can be manufactured by a skilled glass-blower or made of commercial laboratory glassware. Before startin e heavth g y water electrolysis e electrolyzeth e deuteri,th d an - r um vessel shoul e evacuatedb d simplA . e mechanical pum s usei p d after openine th g valves V|, X^, d V-,. The mercury level in the long (800 mm) vertical glass tube an indicates whe e wholth n e syste s mi evacuated . Closin e valvth g e V^, e mercurth , y a manometer n i level s a , , show e tightnese th systemsth e f mercuro Th s. y level should not sink if the vacuum-tight system is properly sealed. If the Hg level doet sinno sk within abou minutes0 1 t e heavth , y watert pu electrolyze e b n ca r into operation. To get clean deuterium gas, heavy water and deuterized alkalis + NaOD O (D , KOD, LiOD s electrolyta ) e usedb n . ca eSinc e amounth e f hydrogeo t n

78 Liquid nitrogen trap Mercury traps

vessel n io r fo sources (ca. 11)

Water cooled jacket

Pt electrodes

Fig.46 Electrolyzer fillingr fo deuterium tanks of neutron generator sourcesn io

s negligibli H g NaO KO e solutio2 e th r 1- H o comparen i f e o n100-20 th o l t dm 0 D-O, the utilization of normal alkalis to set the conductivity is not usually a problem. Depending on the thickness of the Pt tube electrodes and the Pt wires, the power supply is a 20-50 V/2-3 A unit. A full wave rectifier with variable output voltag s i suitablee e outpuTh . t voltag s i usualle y regulate a varia y b dc oe primarr th tria n o cye ste th sidp f o edow n transformer.

79 The electrolysis should start with low current. The voltage (and the current) of the electrolysis should be increased gradually. To prevent the conical taps from falling out, the deuterium vessel should not be filled up to the atmospheric pressure. When the electrolysis stops, the Vj and ¥2 glass taps should be closed, and the O-ring sealed connection between the electolyzer and the deuterium vessel can be vented into the atmosphere by the valve V^. With two identical deuterium vessels e th deuteriu, me th acceleratosuppl f o y r will e electrolyzere smootth b f d uninterrupteo an he us .e th y b d

FLOAE 7.TH 1 T REGULATOR ELECTROLYZER e regulatioTh d stabilizatioan n e deuteriuth f o n n msourcio pressure eth n i e need some control circuits e differenrelateth s o valvest ga de tcombinatio Th . n of an electrolyzer with a mechanical float regulator can serve as a reliable gas

Fig.47 Construction of the float regulator-electrolyzer [53] (numbers are explained in the text)

80 supply for the ion source. The regulating electrolyzer, described below, has been used satisfactorily for a long time at Bratislava [53]. e constructioTh e regulatinth f o n g electrolyze s i showr n i Fig.47n e maiTh . n component is a float (4) filling almost the whole volume of the cylindrical cham- r (1) be e electrolyze.Th s i closer d e froe th uppeth m n i e flangr th ) sid(2 f eo e center where there is a nozzle (3). The float has small points (5) round its cir- cumference that preven e floath t t jacket from directly touchin e inneth g r surface of the chamber and at the same time allows the float to move freely along the e floath e f tchamber e o th n e uppea th centether th axif d o s li f i so rn e I r . elastic rubber sea le inned th oute(6) an n r .O r e loweth wall f e ro sth edg f o e float are the platinum electrodes (7) of the electrolyzer. The chamber and the float are placed in a glass vessel (8) partly filled with the electrolyte D2O with the addition of KOD. The material of the float, the nozzle and the float chambe s Perspei r x (Plexiglas). In order to understand the operation of the regulating electrolyzer, let us assume that the whole free space of the float chamber is filled with the electro- e floa th s i pushe lyteto s , d upwards t tile i closenozzl lth e inle y th f b so e t its elastic seal. If DC voltage of appropriate polarity is applied to the electrodes, the electric current starts to flow between them, and deuterium gas productio ne inneth start t ra s negative electrode e deuteriuTh . s graduallga m y fills the space between the float and the inner surface of the float chamber, pushing the electrolyte out. Wit a hgradua l e electrolytdecreasth f o e e e th leve n i l float chamber, the force P, by which the float presses the seal to the nozzle, will also decrease. The force P can be described by

2 2 ) PI - 2 (p d * - P= ^ + D sG h- (21)

where = diamete e floaD th f t o r s = specific density of the electrolyte h = height of the electrolyte level measured from the bottom of the float G = weight of the float = diametee nozzl de th inle th f o ef t o r P= deuteriu2 s pressurmga e abov e electrolytth e e = deuteriu , p m pressure nozzl e th inle th f o et a e

81 The size and mass of the float as well as the diameter of the nozzle inlet shoul e choseb d o e thas followinnth t g relation r eacfo s . h (21eq e ter ar )f mo fulfilled:

(P'P2 d > » G d an DG 2 sH> (22)

e e floatheighth th wher f s .i o t H e If the construction of the regulating electrolyzer obeys these conditions, tha certaie t a forc nP e electrolyt eh decreaseleve= h l o mucs s h tha a leakt - age appears betwee e rubbeth n e nozzlth r sead e an l edg d deuteriuan e m flows from the float chamber inte vacuuth o n source mio e leakag Th e syste. th e f mo grow s with further decrease of the electrolyte level until the gas quantity q that flows through the leakage into the ion source equals the quantity Q originated by the electrolysis, i.e. until Q = q (23)

becomes valid. If the electric current in the electrolyzer circuit is changed, s thachieveega quantite y th electrolysib d f o y s i changes e cor th s a well d-y B . responding change of the electrolyte level in the float chamber the leakage between the rubber seal and the edge of the nozzle is also automatically changed n i suca e degrequantit ) th hs q wel (a s w equilibriuf ea o lyne tha a t s mi achieved, i.e. a state at which the eq.(23) is valid. When the electric current is switched offe electrolytth , e level graduall e t levea whicth ylo t h risep u s the nozzle actually gets completely closed e timTh .e behaviour e equilibriuth f o m settling e equation aftes givei e th chang th y rQ nb f o e:

) 2 D - q(h)]/ - Q [ 2 dh/dw4 (D = t (24)

where the meaning of D and DQ is evident from Fig.47, and q(h) is the function of e escapinth g deuterium quantity froe floath m t chambe n sourcio r inte eth o during uni e telectrolyt th tim f o e e ebehaviou Th leve . h f l o thir s functio s i characn - teristi r eacfo c h constructio a regulatin f o n g electrolyzer, sinc t i depende n o s the geometry and the mechanical properties of the nozzle and the seal as well as e geometrth e floatn th o f .yo e experimentaTh l function a giveq(h r fo n) geometr d constructioan y a f o n regulating electrolyze s i showr n i Fig.48n e rougTh . h surface rubber float shows a faster rising function while the fine surface silicon rubber float gives a relatively easy way to change the gas admittance characteristic of the regulating- electrolyzer [53].

82 18 22 26 38 40 h[mm]

Fig.48 The ion source gas consumption q vs height of electrolyte for the rough surface (1) and fine surface (2) silicon rubber seal on the top of the float

Fig.48 shows a relatively easy way to change the gas generation. It can be done by adjustment of the electrolyte level. When a regulated current source is used for the supply of the electrolyzer, the gas admission into the ion source wil e stableb l e advantageTh . f thio s s float regulating electrolyzer are: n sourcee io stabilizeb e e s o t th supplt regulaI enableth s -ga y f b de o y th - s e currene electrolyzerth tioth f o nn i t . - The insulated distance control of the current is relatively simple. - If the electrode current is switched off the ion source will be closed.

83 . REMOT8 E CONTRO HIGE TH H F VOLTAGLO E TERMINAL n n sourceio sourcio e e th e, Th power supplies, beam handling facilities, etc., placed on the HV terminal of a neutron generator require normal control during operation e th n relate.sourc io Regulatiod an ede th device f o n s needa s control throug e acceleratioth h n high voltage e terminaTh . l contro a neutrof o l n generato : e carrieb by t n ou ca dr - mechanical control using insulators (manual driv r y Perspeo nylob e d nro x fiber), - electromechanical control (motor gear driven insulating rod or nylon fiber), - insulation transformer controe poweth f terminalV o rlH supplie e th ,n o s - optical insulatio e controterminaV th H n i ne l th linl f unitso e , - computer (microprocessor) e terminacontroth f o l l units with serial optical links between the ground and the HV terminal.

8.1 MECHANICAL CONTROL This is the simplest solution for the distance control of mechanical leaks and variacs at the HV terminal. Insulating (Perspex, Bakelite) rods are fastened to the shaft of the needle leak or variac and there is a suitable knob to turn the insulating rod at the ground potential. Since the insulating rod - especially in high humidity - may conduct, the ground side of the rod should be grounded between the the manually touched knob and the HV terminal to avoid an electric shock to the operator. If the mechanical transmission is made of nylon fibers, similar precautions shoul e takeb d o t avoin n electria d c shock A schemati. c rep- resentatio f thio n s simple distanc s showi e Fig.4n i n . 9

HV TERMINAL

GROUNDED TURNING i-KNOB

Fig.49 Insulating rod control of the regulation elements at the HV terminal

84 8.2 ELECTROMECHANICAL CONTROL e electromechanicaTh V terminaH le s th e i similacontrounitlth t a o f st o rl e mechanicath l control. However, instea f manuao d l control e rodr th fibers(o , s ) are driven electrically by servo motors, DC motors with gear, etc. The insulation between the HV terminal and the ground is the same. Since the servo motors and especiall motorC D e sth y with producy geama r a highee r momentum e mechanicath , l overload of the insulating rod, fiber or variac, and needle valve should be avoided This is why position limiting switches are used for the electromechanically driven variacs, potentiometers, needle valves, etc. These switches protect the system against mechanical overdrive and, in the case of the lower or upper mechanical limit, allows the opposite direction drive only. The limit switches e usuallar y microswitches wit(NormaO N h (NormalC N Open d an )l Closed) contacts. The principle of such a mechanical overload protection is shown in Fig.50.

INSULATING ROD VARIAC ETC. NO

HV TERMINAL NO o DOWN MOTOR

o NO LIMIT SWITCHES

Fig.50 Principle of mechanical protection of the electromechanical drives

This mechanical protection utilize a twis poweC C D nmotoD r e supplth r r fo y gear driving insulating rods or fibers. The motor can be switched to the for- war r backwaro d d directio e UP/DOWth y b n N switch. Thi wity s ke hswitc a nor s i h - l opema n position r forwarFo . d motio e mototh n r wile powereb l y positivb d e voltage, and for backward direction it will be connected to the negative voltage. The two limit switches break the positive or negative supply when the position of the rod (fiber) reaches the upper or lower limit. These switches ensure the return of the gear in the opposite direction after the positive or negative supply has been switched off. This type of electromechanical remote control ( from the control desk ) is utilizee Toshibath t a d , KFK d MULTIVOLan I T neutron generators.

85 TERMINAL POTENTIAL + 150 hV

INSULATING n n n n n n n TRANSFORMERS E VARIACE CONTROTTH OTH N O S L DESK

Fig.51 Insulating transformer control of HV terminal (SAMES J-15 J-25and neutron generators)

8.3 INSULATION TRANSFORMER CONTROL Thi n ssourc io typ f o ee contro a e solutiocontro th a s Pennini lr f o fo nl g ion n sourcextractioa r o e r duoplasmatroo n F voltageR r Fo - n .source in io n e th s sulation transformer contro s e i leshiglth s h o t reliablnumbe e du f eo insula r - tion transformers connected to the HV terminal. The KAMAN (TMC) A-111, A-1254 and e SAMEth A-71 d an S1 typ neutroD e n generators utilize this metho r insulatiofo d n remote terminalV e contro equipmena e H Penninth cas th f e f o eo th n n lI .io gt a t source a simpl this i s e metho dF typ R whil ee requireth e a numbes f separato r e insulating transformers: as many as five or six, as at the SAMES J-15 and J-25 neutron generators. The probability of sparking along the surface or between the primar d secondaran y y a side singln i s e transforme s i rmuc h lower e th tha n i n five or six insulating transformers. The utilization of a single, high power insulating transforme s i typica rr neutrofo l n generator d smalan s l accelerators. Figure 51 shows a block diagram of the insulation transformer remote control of the SAMES J-25 [54]. e onlTh y advantag f insulatino e g transformer e easth ys i svaria c controt a l the ground potential.

86 660

1k o 100n=*\\ : = i--= - U? 100n

OOV AC HV TERMINAL

Fig.52 Optoisolated triac control of mains transformers powere th o t controlterminalV H t a oo e disadvantagTh f thio e s typ f o eremot e control e n additioi th hig, o ht n probabilit f o yelectrica l breakdown alon e e th gsurfacth transformers f o e s i , that repair of the insulated secondary coils is impossible in practice.

8.4 OPTICAL INSULATION CONTROL e V terminacontroH Th f o l l equipmen y b opticat l fiber insulatioe b n ca n ANALOG or DIGITAL. Analog control is mostly used for neutron generators. How- ever, some sophisticated neutron generators (RTNS-II, OCTAVIAN, etc.) use microcomputer control. Analog control of equipment at the acceleration high voltage (HV terminal) requires power supplies and other units where the regulating input needs DC voltages. Thi s e inputh usua s i t l remot ee th mediuinpu f o t m frequency power supplies and other electronically controlled devices. In a simple power supply using one mains transformer, the variac type input voltage regulation can be changed to C voltagD a e input controlled triac control circuit. This e usuath circui s i l t power electronic circuit e requireth : C inpuD d t contro l e produceb voltag n ca e d aftee opticath r l insulatio simply nb e circuits. A typical triac control circuit with optical inpu s i showt n i Fig.52n , based on the AID (SAMES) T type neutron generator [55]. This circuit has a pulse generator unit at the ground potential which drives the infrared light-emitting diode. This light source transmits the light pulses to the optical receiver photo transistor by an electrically insulating optical fiber cable. The optoreceiver is on the HV terminal. The output pulses of the optoreceiver are integrated and the integrator output drives the DC input of the triac (or thyristor) controller chip. The pulse generator works at constant repetition rate ( constant pulse fre- e quencdutTh e . pulse) yyth cycl sf o e wil e b ltransforme d into voltage th t a e integrator. Figure 52 shows this universal triac controller circuit utilized by SAMES. The transmitter can be a TL494 pulse width modulation circuit or a NE555 timer circui n astabli t e multivibrator mode [56]. Fro ma practica l poin f s i viewo t ver t i ,y importane transienus o t t sup- pressors at the small transformer, giving a phase control signal, and parallel to the triace e operationausuath Th . f lo typesy e thyrisan lTh . amplifiere b - y ma s tor controller circuit may be one of the commercial type 1C proposed by the semiconductor manufacturers. The optical fiber can be substituted by Perspex rods; the optotransmitte d receivean r e recommende ar a rmatche e b o t d paid r (lik- op e tical gate).

8.5 COMPUTER CONTROL Increasing applicatio f microcomputeo n r technique s promoteha s e developth d - a mencontro f o t l syste r processemfo s relate o t daccelerators n A up-to-dat. e

88 LJ Umg cr A.C. J ON J Ug o POWER POWER OFF 1/1 < UQ \ \ I 350 kV

. U U f eU x 1 UJ ex. STEP MEASU- J ON AUX. J MOTOR J '\ lEOj^a / f POWER V RING OFF VACUUM / 1 | x y£ f yr _rf^ ^ * V k 0 30 (V^__p— FIBER OPTIC CONNECTION MAIN 1 [~ \ —— -t- VACUUM CON-MEA- BEAM U- C TROR L ING Ij-\i AUD TEC MEASURIN G i /L™/ CONTR. CONTRO L AN D r AkIj A r

Fig.53 Block sourcen io diagram e th controlf o CAMACy b e neutron th t a generator using intelligent CAMAC crate controller [51]

• i. >. --> o >* oj w ^ -s- ofc X 0 ^ < i • D u. _ >~ «i "c«> 3 fc M £ CO zl I 0 U)(7) ^'D 54 * A graphic computer ^ ^ terminal r ^

^ ^x^ i V'micrcx:omputer \/ / for electronics for communication matching and transfer 0

NEUTRON GENERATOR

Fig.54 Hardware system block diagram neutrona f o generator control

89 +12 V

Fig.55 Optical transmitter of analog signals from control console to HV terminal

computer control syste r sucfo m h purposes represent n on-lina s e closed loop system consisting of one central computer and additional microcomputers for special tasks. Such a system reduces operating expenses and is necessary for testing and operatio e automatioth e f acceleratorso th n f o n e microprocessoTh . r syster fo m neutron generators requires solution of problems related to the operation and constructio a (cascade f o n ) generator e opeTh .n loop e utilizesystemb n t a ca ds the manual control of the generator, while the closed loop systems can handle the entire controe neutroth f o l n generator operation. e modulee acceleratoTh th f o s r (neutron generator) contro e almosar l t standard: CAMAC or, more recently PC cards. A block diagram of a neutron generator (with duoplasmatron ion source) control is shown in Fig. 53 [57]. In this neutron V termina generato e n H potentiae sourcth io th t s e e a floatini elf th th o r l n o g acceleration voltag + extractioe n voltage e opticaTh . l fiber cables allow excellent insulation between the HV terminal and the subterminal of the ion source. The sensors and controllers are at different potentials, in the HV area and the ion source area, so they are connected by different fiber cables. The intelligent CAMAC controller at the ground potential controls the ion source power supplies and the gas supply. The measured parameters - e.g. arc current, arc voltage, n e sourctransmitteio ar pressur- e eth n i e degroun th bac o t dk potentiale Th . software systeth s beef ha mo e n designe o t meed t requirements like startine th g generator operation, controllin e operationath g l parameter e generatorth f o ss a , wel s operator-aidea l d contro d safetan l y contro breakdown i l n situations.

90 +12V

3 TL072

-*- 12V Fig.56 Optical receiver of analog signals at HV terminal with a switch-on option

+ 12V

-12V Fig.57 Optical transmission circuit of signals from HV terminal to ground (HVterminal controlto desk)

similaA r system using CAMA Cs showi daty n Fig.5i nwa a 4 e [58]loweTh . r part of this diagram shows the microcomputer configuration (Z-80 based). The measured ion source parameters and the output control values are using the CAMAC data way in digital form. The analog values measured at the HV potential are connecte a multiplexe o t d o obtait r C nbetwee AD mor e th en input e analoTh . g output s alsi oC demultiplexedDA e oth f . e transmissioTh n techniquee electronith d an sc equipmen e vicinitth f n i o yt the neutron generator were designe d developean d d taking into consideratioe th n noise immunit d protectioan y n from corona discharge d sparkinan s g alon e systh g- tem. The analog and digital (binary) signals are transmitted and received by

91 + 12V

BC18_ - 2

Fig.58 Optical receiver signalse th f terminalV o H from e th (analog meter driving frequency voltageto converter controlthe in desk)

glass fiber optics e electroniTh . c components e th terminause n i d l blocke ar s radiation hardened. The optical transmission system utilized for the transmission of the measured controan d l voltage . Fig.556 s showi d s5 an shown Figi n 5 e 5 transmitsth s - ter-receiver configuration e relateth operato o t d r console (ground potential) V terminal-to-grounH 8 sho 5 while th wd ean Fig7 5 s d optical insulation circuits. The circuit in Fig.55 is a voltage-to-frequency converter with infrared LED output. This circui s utilizei t d e grounV th termiH bot n e o dhth - potentiat a d an l e opticanalTh . l receiver, frequency-to-voltage converte s i show rn i n Fig.56. This circuit also has a comparator actuating a limit switch. This limit switch is utilized as a mains switch-on device at the HV terminal. The analog output of the 1 LintegrateM33 d circuit utilize n i frequency-to-voltagd - re e th o t e d le mod s i e mote control input of the power supplies on the terminal [59].

92 9. VACUUM SYSTEMS OF NEUTRON GENERATORS

1 IMPORTAN9. T TERM UNITD SAN VACUUN I S M TECHNOLOGY [60]

9.1.1 Terms e termTh s use n descriptioni d d technicaan s e pumplth d compodatr an sfo a - nents in catalogues are standardized. The more important terms given below have been chose o t assisn t those less experience n vacuui d f o m e technologus e th n i y the catalogues and to extend their utilization. An absolute pressure gauge is a pressure gauge used to determine the pressure from the normal force exerted on a surface divided by its area. An absolute pressure gauge is independent of the gas type used. Absorption is a type of sorption in which the gas (absorbed) diffuses into e soli th e bulr liqui o dth f o k d (absorbent). Adsorption is a type of sorption in which the gas (adsorbed) is retained at e surface solith th r liqui o df o e d (adsorbent). Backing pressure is the pressure at the outlet of a pump which discharges to a pressure below atmospheric only. Compressio ne ratith rati os i o betwee e outleth n e t inleth pressurt d an e pressur a pum f r specifio e pfo c gas. Concentratio f moleculeo n e numbeth s i sf molecule o r s containe n adea n i - d quately chosen volume divided by that volume. Intrinsic conductanc s i conductance e speciath n i el case wher e orificth e e or duct connect o vesseltw s n i whics h Maxwellian velocity distribution prevails. e cas th f moleculao en I r flow, intrinsic conductanc e e producconth th s f i eo - t ductance inleth e conductancf t th o epor f e o ductth te f transo th esectio d - an n mission probability of the molecules. In this flow range, it is independent of the pressure. Degassin a desorptio s gi n whic s acceleratei h physicay db l processes. Desorption is the liberation of gases absorbed by a sorbent material. The liberation can be spontaneous or can be accelerated by physical processes. Gas diffusion is the movement of a gas in another medium due to its concentration gradient. The medium may be gaseous, liquid or solid.

Flow: Viscou s throug ga e spassag th a flo a s hduci f w o e t under conditions such thameathe t n free pat is verh y smal in comparisol n witsmallesthe h t internal dimensio a cros f o ns e sectioducte th flo s Th i thereforf w. o n e dependene th n o t viscosity of the gas and may be laminar or turbulent. In the case of viscous flow, the resistance is a function of the pressure.

93 Turbulent flow (eddy flow) is a viscous flow with mixing motion above a critical Reynolds number (For circular cylindrical pipes2300)= e R ,. Laminar flow (parallel a flow viscous i ) s flow without mixing motiot a n small Reynolds numbers. Molecular flow is the passage of a gas through a duct under conditions such e thameath t n free pat s i verh y larg n comparisoi e n wite largesth h t interna- di l mensions of a cross section of the duct. In the case of molecular flow, the resistanc s independeni e e pressureth f o t . Gas is matter in a state of aggregation in which the mean distances between the molecules are large in comparison with their dimensions, and the mutual arrangement of the individual molecules is constantly changing. Gas in the stricter sense is matter in gaseous state which cannot be brought to a liquid or solid state by compression at the prevailing temperature. Gas a e controlleballasinle th f o ts i t d quantit a gas f o y, usually inte th o compression chambe a positiv f o r e displacemen o prevent t s pumpa to s ,condensa - tion within the pump. Gettering means bonding of gas, preferably by chemical reactions. Getters (getter materials) often have large real surfaces. Leaks in a vacuum system are holes or voids in the walls or at joints, cause faulty b d y materia r machinino l r wrongo g handlin e sealsth f .o g Lea ke throughpu s th througrat ga s a i efunctioe a s a th i hleakf o f t tI o n. type of gas, pressure difference and temperature. The unit for the leak rate is: mba0 1 s'1 r= 1 . W 1 = 1 s' 3 m a P 1 Maximum tolerable water vapour inlet pressure, p , is the highest inlet WO pressure at which a vacuum pump can continuously pump pure water vapour under ambient condition f 20° o sd 101 Can 3 mbar. The mean free path is the average distance which a molecule travels between two successive collisions with other molecules. Outgassin a spontaneou s gi s desorption. Partial a pressur specifier e vapouo pressuro th t s s e i ga erd du ecomponen t oa gaseouf s and/or vapour mixture. Permeatio e s passagthrougth ga s f i no a e hsoli d barriea liqui r f o o dr finite thickness. Permeation involves diffusion and surface phenomena. The pressure of a gas on a boundary surface is the normal component of the force exerted by the gas on an area of a real surface divided by that area. e legaTh l pressure units areI uniS :te th Pasca (abbreviatios a d l an ) Pa n bar as a special unit designation for 105 Pa. 1 Pa = 1 N rn2 1 bar = 1000 mbar = 105 N m"2 = 105 Pa e uniTh t commonly use vacuun i d m technolog e millibarth s i y .

94 Quantity of gas (pv value) is the product of the pressure and volume of a specifie e prevailinth t da quantits ge ga b temperature f o v o t valup y s i e e th f I . used as a measure for the quantity of substance or gas, this must be an ideal gas whose temperature must be specified. e resistancTh e reciprocath e s conductanceei th f o l . The Reynolds number is the nondimensional quantity

Re =^ (25) where Densit= p f fluiyo d = Averagv e flow velocity 1 = Characteristic length (e.g. pipe diameter) n = Dynamic viscosity R230< e 0: Lamina r flow 400> e 0: TurbulenR t flow The saturation vapour pressure is the pressure exerted by a vapour which is n i thermodynamic equilibriu s condenseit f mo e witdon he prevailinphaseth t a s g temperature. s (sorbate ga a Sorptioa liqui solif e takinr o y th o db d )p s u i gn(sorbent) . Sorbents are also called sorption agents. e standarTh d reference e conditioconditio th a solid s i f ,o n liquir o d gaseous substance determined by the standard temperature and standard pressure. The abbreviation is NTP : Normal Temperature and Pressure.

StandarC 0° d= temperatureR a 273.1= l f T 5K :

Standard pressure: pn = 101325 Pa = 1013.25 mbar The throughput is the quantity of gas (in pressure-volume units) passing through a cross section in a given interval of time at the prevailing temperature, divide thay b d t time. e throughpuTh puma e throughpu th f o ts s pi pumped ga f o t . l partiaal f e o total Th m pressurel su pressur e th s s i epresent . This ters mi used in contexts where the shorter term "pressure" might not clearly distinguish between the individual partial pressures and their sum. e ultimatTh e pressur e valuth s ei e whic e pressurth ha standardize n i e d test dome approaches asymptotically, with normal operatioe vacuuth f mo n pumd an p without gas inlet. Distinction can be made between the ultimate pressure which is due to noncondensable gases, and the ultimate pressure which is due to gases and vapours (ultimate total pressure). s wite concentratioga th Vacuuh a e stat th f o es mi f moleculeo n s being less than that of the atmosphere at the earth's surface. Vacuum is the state of a gas

95 Tabl . Vacuu14 e m ranges

mbar Molecule concentration

Low vacuum (GV) 1000 2.5xl025 - 2.5xl022m-3 Medium vacuum (FV) 1-10"3 2.5xl022 - 2.5xl019rn3 3 7 High vacuum (HV) 10 - 10"' 2.5xl019 - 2.5xl015mf3 *y 1C *5 Ultra high vacuum (UHV) < KT < 2.5xl013m'J

The molecule concentrations refer to a temperature of 20°C.

wit a pressurh e below atmospheric pressure ,r pressurai i.e e .th e prevailint a g the respective location e vacuuTh . m range e e giverelationar sTh n Tabli n. 14 e- ship between pressur d concentratioan p e f moleculeo n n (fos r ideal gases: is ) T K n - P K = 1.3807 x 10"23J K1 (Boltzmann constant) Thermodynamic= T temperature Vapou a substanc s s i phasr ga n ei e whic s eithei h n thermodynamii r c equilibrium wits liquiit h r solio d d phase (saturate e broughb d o n t vapour)tca r o , thermal equilibrium by compression (condensed) at the prevailing temperature (un- saturated vapour). Note: In vacuum technology, the word "gas" has been loosely applied to both the noncondensabl e vapourth a distinctiod f i an , t requireds no ga e s i n . The vapour pressure is the partial pressure of a vapour. The volume flow rate is the volume of gas passing through the duct cross section in a given interval of time at a specified temperature and pressure, divide thay b d t time. e volumTh e floe averagth w s i rat eS e volume flow fro a mstandardize d test dome through the cross section of the pump's intake port. Units for the volume flow rate are m s" , 1 s" , m h" . The water vapour capacity C«y is the maximum mass of water per unit of time which a vacuum pump can continuously take in and discharge in the form water vapour under ambient condition f 20°o d s101 an C 3 mbars g i giveh"e t n I Th i n. 1 relationship between water vapour capacity D^Q and maximum tolerable water vapour inlet: pressurs i W P e

- j g n i (26) 7 21 = o W C

S = Volume flow rate, in m3 h- 1, at inlet pressure

pw = Maximum tolerable water vapour inlet pressure, in mbar T = Thermodynamic temperature of the water vapour pumped, in K

96 Table ISA, Multiples and units Power of Prefix Desig- 10 nation

IO9 giga G IO6 mega M IO3 kilo k 1 io- deci d 2 io- centi c 3 io- milli m 10"6 micro /" 9 io- nano n 12 io- pico P

Table 15B. Vacuum technological quantities [62]

Quantity SymboI unitS s l Recommended units

Pressure P N/m2,Pa bar, mbar Total pressure mbar Pt Partial pressurs ga f o e p. mba) N r ,p (e.g R p . constant "i" 2 2 Saturation vapour pressure mbar PS Vapour pressure Pd mbar Residual total pressure pr mbar Residual gas pressure P mbar mbar Residual vapour pressure Pm Ultimate pressure mbar "end Mean free path r, A m m, cm Collision rate z 1/s s"1 Collision rate related z" " 2 1/cm m s " s , " m s" to area Particle density n 1/m3 m -3 , cm-3 Throughput V N m/s mbar 1/s Leak rate s Nm/ mbar 1/s Pumping speed s m3/s 1/s, m3/h Conductance c m3/s m /s, 1/s Impedance R s/m3 s/m , s/1 Conductance L 1/s 1/s

97 Tabl . Conversio16 e n tables (a) Temperature

K °C OF

Kelvin 1 °C+273.15 5/9 (°F+459.67) °Celsius K-273.11 5 9 (°F-325/ ) °Fahrenhei K-459.65 ° 9/ 5 9/ 7t 2 3 + C 1

(b) Pressure Pa mbar bar torr atm

1 N/m2 = 1 Pascal 1 0.01 10'5 7.5xlO'3 - 1 mbar 100 1 10'3 0.75 1 bar 105 1000 1 750.06 0.98 1 torr = 1mm mercury 133.32 1.33 1 - 1 atm 101325 1013 0 76 1.013 1 2 1 at = 1 kp/cm 98066.5 981 0.981 735.6 0.97 f wateo m r 1 9806.65 98.1 0.098 0.097 0.1 f ps lb/ft1 1 = 2 47.88 0.47 0.36 i ps lb/in1 1 = 2 6894.8 68.95 51.71 0.068

kp/c= t a m mW s psf psi

1 N/m2 = 1 Pascal - - - - 1 mbar - 0.0102 2.09 0.014 1 bar 1.82 10.19 2089 14.50 1 torr = 1mm mercury - - 2.78 0.019 1 atm 1.033 10.33 2116.2 14.69 1 at = 1 kp/cm2 1 10 2048.2 14.22 1 m of water 0.1 1 204.82 1.42 1 lb/ft2 = 1 psf 1 1 lb/in2 = 1 psi 144 1

98 Table 16 (Cont.) (c) Volume flow rate, Conductance

m3/s 1/s m3/h fl3/min

1 m3/s 1 1000 3600 2118.88 1 1/s 10"3 1 3.6 2.119 h / m 1 2.78xl04 0.278 1 0.59 1 ft3/min 4.72xlO"4 0.47 1.69 1

throughput(dV p ) , Leak rate W mbar 1/s torr 1/s cm3/m (NTP) a mP 3 W l / 1 s= 3 1 10 7.5 592 1 mbas 1/ r 0.1 1 0.75 59.2 tor1 r 1/ls 0.133 1.33 1 78.9 cm1 3/m (NTP) 1.69xlO"3 0.017 0.013 1

sourcn (e Io )s consumptio ga e throughputV (p n ) unit l atm/houm 1 = 2.8x10r " t Normaa mba s 1/ r l Temperatur Pressurd an e e (NTP)

9.1.2 Units [61] e basiTh c units were establishe n Octobei d r e 195Firsth 4t a General Con- ference for Mass and Weight in Paris. This uniform system was given the interna- tionally binding initials "SI" (froe Frenchth m , Systeme International d'unites). The recommended units are mostly multiples of fractions of the SI units (Table s 15A)exceptionA . o t thiss , some units which have become customar n i certaiy n branche e recommendear s d (see Table 15B) r convenience.Fo a conversio, n tabls i e given in Table 16.

99 VACUU2 9. M PUMPS When selectin e mosth g t appropriate vacuum a systeneutro r fo m n generator the following important factors should to be taken into account:

(a) The value of the ultimate pressure to be attained for normal operation:

The aim of all vacuum systems for a neutron generator, as for a charged particle accelerator, is that the accelerated deuteron ions should reach the tritium target without collision. To do this, it is necessary to keep the pressure in the acceleratin thaw e gmea lo th ttub o ns e fre r emolecule ai pat f o h s should exceed the length of the accelerating tube. Dat t a differena e seeb - n Tabli shontn 7 ca w1 pressuree s a e tha th - t s above requirements for the mean free path values for neutron generators at pres- e " -10rangth 10 suref "somo n ef i o mbas e e ar metersr . These ultimate pressure values, in Table 17, have to be taken into account, at least when selecting the most appropriate typ f vacuuo e m pump [63].

Table 17. Mean free path vs pressure

Pressure (mbar) Mean free path (cm)

atmospheric 6x10" 1 5xlO"3 10"3 5x10° 10"6 5xl02 10"9 5xl06

s (bintakGa ) e (du e o t n sourcelakageio e th , , etc.e accounteb )o t r durinfo d g the run of the accelerator:

e pumpinTh g n importanspeea s i d t consideration s requireIt . d value shoule b d determined, first, by the regularly occurring gas intakes. Note that the pressure e systeth n mi e changeb must no td during operation r fo example: a neutro t a , n yield of 10 n/s in the case of a commercial neutron generator with a beam current of 1-2 mA, the Y>2 8as intake is 4-5 cm /h.

100 (c) Economic considerations:

- The greatest possible simplicity in handling the pumps and the entire vacuum system (start, stop); - The lowest possible probability of failures (the pumps may break down, espe- ally during long irradiations); - The least possible reconstruction requirements during maintenance (to repair an ye syste partth f mo s migh e difficulb t t owin o tritiut g m contamination); - The best possible accommodation for the existing laboratory conditions (various supplies, e.g. for electric power, water and liquid nitrogen; replacement possibilities of failed parts of instruments, workshop background for maintenance and renewal, etc.). Fig.59 shows the best known pump types, mad wely b e l known manufacturers.

Taking into account both the pumping rates and cost during operation, the manufacturer f o sneutro e followinnth generatorn f practici o e ge on us epum s p " mbacombination10 - r " 10 ultimat o t obtais e eth n pressure (necessara r fo y neutron generator): - oil diffusion pump - rotary pump - turbomolecular pump - rotary pump - Ti-ion getter pump + rotar( s y pump).

The working principles of these pumps, vacuum gauges and other vacuum technical equipment e mos th s wel a s t, a limportan t e operatiodetailth f o s f vacuuo n m systems containing the above fittings, are discussed below.

Ultra high Medium Low High vacuum vacuum vacuum vacuum

MoM«.ll ecular flow -Vi 5COU s flow - ^ h .Diaphragm Rotiirv i/one Multi - staqe dry Ro plunqey t jr r \— Roots

Turt 0 ITolecular -- Diffusion Sublimation h- Adsor ptio i i— - -H Sputter ion t— H h- Cryoaenic H

Fig.59 The operation pressure range of different vacuum pumps

101 Gas ballast Inlet inlet valve Exhaust valve Pump Gas ballast fluid inlet Sliding vanes

Fig.6 rotarye Th 0 vane vacuum pump

9.2.1 Vacuum system based on a combination of oil diffusion and rotary pumps e rotar(aTh ) y vane pump This is probably the most widely used pump and is well established as a backing pump in many compound systems. n eccentricallA y placed slotted rotor a cylindricaturn n i s l stator (Fig.60) drivea directl y b n y couple r threed(o o electri) tw ce e slot ar th motor sn I . sliding vanes which are in continuous contact with the walls of the stator. Air drawn in is compressed and expelled through a spring loaded exhaust valve. e vane d Th roto e an sealesa ar e rstatoflui y th s b dimmersei d r d filan m d in the fluid to provide heat transfer to the pump casing. Rotary pump fluids are usually selected from high quality minera l witw vapouoi llo h r pressur d gooan e d lubricating properties. Suggested l aretypeoi f :o s Shell Rotary Vacuum Puml Oi p Gluvaco0 91 R l Alcatel 1DO F MOVIXEL A PV100 TURBEL0 10 A FS ELFLL BARELF F 100 0 10 S C P B INLAND 15 INVOIL 2D SHELL VITRE0 A10 TOTAL CORTIS 100

Two-stage rotary pumps are frequently used, in which the exhaust from the- first stag is internalle y seconthe connecte inlethe of d t to dstag e (Fig.61).

102 Inlet

Exhaust

Fig.61 Cross section two-stagea f o rotary vane pump

Fig.62 The operation of the gas ballast in a rotary vane pump

Gas w. •HH molecules Inlet flange Vapou— t je r Cooling. coils

Condensed .__«• vapour 3 c 9 i o*» B Exhaust flange oy iQ «i .(Backing line) Hot | [ L®—fi vapou * o-*.* » - r - * 1—

im^lJ^TTTT-^p^in iii>rlij~iiTlr~-^ r>..._p flujd Ejecto t je r Heater

Fig.63 Schematic cross section of a diffusion pump

103 This solution improves the ultimate pressure of the pump by reducing the back leakage where the rotor and stator are fluid sealed. To reduce the condensation of vapours during the compression cycle, gas ballasting can be used, where a controlled quantity of a suitable noncondensable gas (usually air s )admittei d during compression (Fig.62). In Fig.62, vane A is about to close the crescent-shaped chamber B containing the gase d condensablan s e vapours being pumped s thiA .s volum s sealei e d froe th m inlet, a controlled volume of air at atmospheric pressure is admitted at point C. r intakai e e Th d raiseprevent e an pressurth sB sn i econdensatio f vapouo n y b r openin e exhausgth t valve befor e conditionth e f condensatioo s e reachedar n .

(b) The diffusion pump The operation of the diffusion pump can be seen in Fig.63. The pump fluid (usually oils with large molecules) is heated electrically and the oil vapour streams through the chimneys. Then the vapour molecules - emerging from the (ring-shaped) nozzles with supersonic speed - are directed toward the cooled pump wall, where condensation takes place e vapouTh . r condensate e reflowbottoth o t ms of the boiler. This process can be maintained continuously by permanent heating. Air molecules in the pump - and in the connected vessel - can be mixed by Brownian movement (diffusion) between the oil molecules while moving towards the wall l ; probabilityal y thus collisionb n i - , a r velo t moleculeai - ,sge - n ca s city component directed downwards because their mass number is much smaller than that of the oil molecules. Gases compressed in this way in the lower part of the pump can be removed by the backing pump. The condition of the operation is that the mean free path of the oil molecules shoul e greateb d r thae distancth n e e froe wallnozzlth th m .o t eThi s condition can be reached in such a way that - before switching on the diffusion pump - the whole vacuum system should be evecuated to a forevacuum (10 -1 -10- 2mbar) by applyin a gforepum p (usually rotary type s backina ) ge diffusioth pum o t p n pumpn I . addition, the rotary pump must be kept in operation as long as the diffusion pump work n i ordes o t prevenr e disintegratioth t e lowesth f o tn vapou t streaje r m system because of the accumulation of gases from the diffusion pump at the exhaust passage towards the backing. On disintegration, the velocity of the air molecules in the upward direction is the same as that in the downward direction. This is interesting becaus a jet'e s disintegration a gradualead o t s l breaking dowf o n the pumping ability. e necessare th valu Th f o e y forevacuu e exhausth t a m t passag a diffusio f o e n pump is at least 10- 1 -10- 2 mbar. However, the pumping speeds of the rotary pumps in this pressure range are usually quite low (see later). It is usual to insert a so-called booster-pum s intak ga n betwee pumpo i ps e highi etw th f e .i s th n

104 In older vacuum systems, an overheated diffusion pump, only at the pressures 0 1 mbar a 1s booster- use0 s wa 1a , d n thiI . s2 way e exhaus,th t pressure th f o e booster increased inte mbath o r region, wher e th rotare y pumps also have high enough pumping speeds. In contemporary diffusion pumps the booster stages are originally built in the housing (stator). Fig. 63 shows that the outstreaming of the oil vapour from the pipe directed towards the exhaust passage - operates as an oil vapour ejector pump.

Table 18. Technical data for diffusion pump fluids [64]

Mineral oils Silicon oils Pentaphenylaether BALZERS 61 71 5 17 SANTN DC705 A C O4VA

Theoretical vapour pressure [mbar] 2xlO"7 2xlO"8 2xlO"8 4X10'10 IxlO'10 2 ] Viscosit/s m [m y 171 410 39 175 1000 (at 25°C) Chemical resistance good good better better best Thermal resistance good good better better very good ^ 7 Pressure rang0 1 e -2-5xlO'6 10"3- 10"7 10-10"' 10"5-10"8 10'3-10~8 Price low low medium medium high

Suggested types of diffusion pump fluids are as follows (Table 18):

Mineral oils These oils are manufactured by molecular distillation of crude oil and they e suitablar r generafo e l applications a dow" pressur mbar10 o t n f o .e They will not withstand repeated exposur o atmosphert et statho ea becausn i e e exposurth e e will produce carbonaceous compounds with high vapour pressures which decrease the performanc e pumpth f o .e Their e depositinneth n r o e sacceleratosurfac th f o e r system will produce conducting layers.

105 Silicon oils These e oilexceptionallar s y stable compound t a higs h temperature d thean sy will provide an ultimate pressure between 10- 5 to 10- 9 mbar. Their deposits on the electrode n a accelerato f o s r will produc n a insulatine g layer. These fluids are poor lubricants.

Pentaphenylaether These fluids have exceptionall vapouw lo y e rar thermallpressur d an ye very stable. They will not backstream in a properly designed pump and baffle, and they e chemicallar y stable. Their e break-ue surfaceth electrodeth n f a o n o pf s o s accelerator will produce a conducting layer. They are good lubricants, but they e expensivear .

—Vacuum chamber

^| lonization gauge

Roughing valve High vacuum Isolation -

Fore-line Diffusion trap pump

valve

Rotary pump

Fig.64 Typical diffusion pump vacuum system

(c) Combination of diffusion and rotary pump A typical system consisting of an oil diffusion pump and a rotary pump can be seen in Fig.64. As mentioned earlier, selection of the appropriate diffusion pump is determined, first of all, by the gas consumption of the ion source. On the other hand, for the selection of the rotary pump - to be fitted to the diffusion pump - the pumping speed of the diffusion pump as well as the maximum permissible pressure at the exhaust have to be taken into account.

106 For example, in the case of a diffusion pump having 1000 1/s pumping speed: if a pressure as low as 10— 4 mbar is assumed for the pumped volume (in fact, it shoul e higheb d r tha nh / pumpinthat) m ,6 g3. the speea n s necessari d e th t a y exhaust side with a 10" mbar pressure. Such a pumping speed can be achieved even e simplesbth y t rotary pumps witn practici a pumpinh h / em g8 (se 6- speee f o d the pumping speed diagrams).

Cooled vapour traps: The so-called foreline traps - liquid nitrogen traps - between the rotary and diffusion pump protect the diffusion pump from the oil e vapourotarth f yo r pump. This muse impedeb t n ordei d o prevent r t mixine th g rotary pum l wite oi pmucth h h finer diffusion pump oil. Thi s i necessars - be y cause such mixing would spoie diffusioth l l (thoi n e molecular size being much larger for diffusion than for rotary pump oils). A vapour s traanotheha p r very important role t i prevent: e vapourth s s (especially steam), generally present in vacuum systems, from getting into the rotary pump. A rotary pump cannot completely remove such water (or even steam). Prac- tical experienc a e traf showo p e improvesus thae t e vacuua leasth th t y stb m haln ordea f f magnitudeo r . e FREON-1Th 2 cooled trap e ar advantageous n laboratoriei s s where liqui- ni d trogen supply is either difficult or impossible to obtain. The manufacturers of diffusion pumps usually stock this type of refrigerator operating cooled traps as well. Fig.65 shows some simple form f cooleo s d vapou rs i quittraps t I e . practi- a vacuu t pu m e inleo t th tra l tn i pca chambe e diffusioth f o r n pump; this reduces the risk of getting diffusion oil into the accelerator tube and helps to limit the different vapours present in the vacuum system. Further questions on the correct use of a vapour trap are discussed later.

Buffer chamber: Sometimes it is worth while to insert a buffer chamber with a 5 litre - usuallo t s 3 volumy f madf o estainleso e s stee - l l betweeoi e th n e rotardiffusioth e cas d th yf poweo e an eFig.6n pump se I . ( r) 4 scut-off e th , buffer chamber will play an important role and take over the duty of the rotary pumpa shor r fo ts time (unti e diffusioth l n pump cools down). s Owinrelit -o t g atively large volumen pumca t pi , a outwhile, r r fro,fo ai e exhaus th m t porf o t the diffusion pump that is still working and reduce the contact of its oxygen l vapoursoi wit e t stilth hho l .

Isolation valves: Diffusion oils in contact with air may absorb a lot of hu- d gaseoumiditan r yai s froe materialth m , which wile releaseb l d again whee th n pum d results i startean p n extri p s su intake dga a . Therefore, after stoppine th g vacuum system of the generator, it is worth while to isolate the diffusion pump

107 —— » ^ Vacuum Vacuum (- ^ _ chamber or _ —— » • — — . Rotary ^Liquid diffusion __ . _-_^ — . pump — = pump Nitrogen — ^ _- Liquid - — — .^Stainless steel vessel -- Nitrogen — — _ — - — —-Glass vessel —: ^ - Rotary — L^^J^~ 1 numo

\ Inlet for liquid rnr 11 Nitrogen — — '"- — — — ^Accelerc tube . D* 7 —— —~^IE~^3— — — Liquid \ Nitrogen \ Inner cylindrical vessel

Fig.65 Liquid N~ cooled traps

Fig.66 Different types f isolationo valves [65]

108 from the other parts of the vacuum system by valves (see Fig.66). These valves are usually quarter swing or butterfly valves. The acceleration tube and the target beam line are usually isolated by gate valves. In this way, air is admitted only inte smalth o l space rotar th e inler e th f admit ai yo jus t a e t pumth - y b p tance valve, while the other parts of the system keep the vacuum. After switchine vacuuth f of mg system e backinth , g valve prevent e vapourth s s e warming-uth of p trap from streaming back toward e diffusioth s n pump e rolTh e. of this valv s similai e e hig th o that hrf o tvacuu m isolation valve. For the safe running of the vacuum system, it is worth while to control the cooling water e diffusiosupplth f o y n pump continuousl a pressur y b y e switch; this switche e pumth f pof s heating automaticall e wateth f i yr supply cuts off. A relay indicating the pressure level - built into the electronic unit of the vacuum gauge (vacuum gauge controller n als a e useca similab r o - fo )d r purpose; this relay stops (or inhibits switching on) the heating of the diffusion pump if the gas intake exceeds the upper limit set in the controller. If a neutron generator - with its vacuum system - is operated for a long s i advisabltime t i , o t controe l continuousl e temperaturth y e diffusioth f o en puma thermocoupl y pb r thermao e l switch.

Pirani lonization gauge gauge Vent <3> valve

Target— assembly

Backing valv) e(1 Vent valve (1)

Rotary Rotary pump pump

Fig.67 General vacuum system for a neutron generator

109 During operation of the vacuum system of a neutron generator consisting of a rotar s a wels ya diffusiol n pumps (see. Fig.67) e followinth , g important instructions shoul e followedb d :

A. Switchin e vacuuth n go m system ) Switca , firste rotaron hth , y pump (backing e vens valvopeneth i t3 d e an d s valvclosei 1 e d automatically) e rotarth t r y Le fo 5-1pum .n 0ru pmin - utes with gas ballast valve open. During this time the pump house warms o operatiot p u n temperature (the t vapourcondensno o d s e later), andn i , addition, it will be possible to release the gases and condensed vapours froe forelinth m e trap. After this "cleaning", the foreline trap has to be filled by liquid nitrogen. b) Ope e backinth n g valv. 2 e c) Ope e isolationth n valv. 1 e e ) heatine Switcdiffusiod th th n f o o hg n pump whe e correcth n t forevacuum e wholth valu n i e - esyste s reachedi m- . ) e After another perio f o 10-1d 5 minute s elapseha s - dthi s depende th n o s heating-up time of the diffusion pump - the cold trap (on the top of the diffusion pump) has to be filled up with liquid nitrogen.

B. Switching off the vacuum system ) Closa e isolatioth e n valv. 1 e b) Empty the cold trap (by warming or compressed air). During this opera- tione th greate, e r th adsorbepar f o t d gased an vapours - re s wile b l leased, and the diffusion pump that is still working will transport them e forelinth o t e trap (within about 10-20 minute - sdependin e sizth e n o g of the traps). c) Switch off the heater of the diffusion pump; let it cool down below operation temperature. d) Close the backing valve 2. e) Switch off the rotary pump (now, the valves 3 and 1 will automatically be turned off and on, respectively). This vacuum system - as can be seen in Fig. 67 - involves the possibility of the tritium target exchange and the exchange of the old ion-source components - using a second rotary pump - without switching off the diffusion pump. In such a casee th require, e dvacuu th parf o mt syste - aftemr closin e th isolatiog n valve 1 - can be exposed to the atmosphere. After the necessary changes have been made, the upper part of the vacuum manifold can be evacuated by the second

110 rotary pump in order to reach the required forevacuum value and, finally, it can be connected - by opening the isolation valve 1 or the isolation valve 2 - to e higth h vacuum manifold.

9.2.2 Vacuum system base Ti-ion do n getter pump The sputter ion pump is a getter ion pump in which ionized gas is accelerated towards a getter surface, continuously renewed by cathodic sputtering. The basic sputter ion pump consists of two flat titanium cathodes, a cylindrical an- n axiaa d l magnetian e od c fiel s showa d Fig.68n i n . A typica i getteT l r pump o flaconsisttw t f rectangulao s r titanium cathodes wit a hstainles s steel anode between them consistin a larg f o ge numbe f opeo r n ended boxes (see Fig.69). The pump, mounted inside a narrow stainless steel box and attached to the vacuum system, is surrounded by permanent magnets. The anode is operated at a potential of some kV whereas the cathodes are at ground potential. The cold discharg s i initiatee y strab d y electrons produce y colb d d field emission. Thes- el e ectrons execut a ehelica l motion aroun e magnetith d c field lined thean sy oscillate to and fro in the axial direction between the cathodes, as in the case of PIG ion sources or Penning vacuum gauges. The positive gas ions formed by ionization bombar e titaniuth d m cathode, sputtering titaniu o fort mm getter filmn o s the anode and the opposite cathode. The titanium reacts with all active gases, forming stable compounds, and a considerable number of bombarding gas molecules wil e buriee cathodeb lth n i d . Noble gases (He, Ar, Ne) are pumped by burial under the layers of titanium e pumth n o p wall d anodean s , whil e otheth e r gase e e ar buriescathodeth n i d . Unfortunatel s a furthey r sputtering takes place e previouslth , y buried molecules can be released, giving rise to instability in pumping. Various solutions to the pumping of noble gases have been attempted, for example the use of differential cathode materialsf o e us ,s i tantalums i e where titaniu th e on d on ed an man , slotted cathodes where the bombarding ion arrives at a glancing angle. However, the most successfu s bee e ha triodl th nn pum io e p configuration (Fig.70) n whici , h the whole pump body (B) is grounded and, being at the same potential as the anode cylinder (A) n auxiliar,a acts a s y anodee ionsTh ,e .diod th producen ei s a d pump, now graze the titanium lattice (C) giving a high sputtering rate, the sputtered titanium forming preferentiall e pumth n po ybody . Energetic neutral particles created by ions glancing off the cathode are bur- e surface pumth th ief n po e o dreflectee ar bod r d o pumpey e anodean dth y t a dAn . positive ions arriving at the pump body are repelled by its positive potential t touc no e surface th ho d d . an Burie r implanteo d d noble gases covered with fresh

111 c D

Fig.68 Schematic diagram singlea f o cell sputter pumpn io

Fig.69 Operation diodee th f o type sputter pumpn io

• Titanium atom O Gas molecule ©Positive ion B Electro0 n ® Neutral particle

Fig.70 The triode-type ion getter pump

112 titaniu e lefar mt generally undisturbed, leadina highe t o t pumpingne r g speed for these gases. Before the titanium ion getter pumps start, it is necessary to reach the fore vacuue vacuuth n mi m e systeacceleratinth f mo g a rotartub y b ey pump cooleA . d trap betwee e rotarth n e ysputte th pum n pumd io ran pp adsorb e vapourth s d an s the humidity of the air. A e srotar th soo s ya n pump reache e ultimatth s e n vacuumgetteio e r th ,pum p shoul a e switcheshorr b d fo t n o timede startinTh . i T getteg e currenrth f o t pump at the forevacuum pressure is usually high, which leads easily to over- f procedur of n pumps io d e heatinswitcan e Th th e.n o f hshoulo g e repeateb d d sev- era le timeth loadin s a s g e currenth gette f o tr pump reache e th normas l working current range. During this timee isolatioth , n valve betwee e rotarth n y pump and the getter pump should be closed, and the rotary pump should be switched off, exposing to the atmosphere its inlet through an air admission valve. The normal work of the ion getter pump - in a leak-free vacuum system - will be indicat- a decreasin y eb d g getter pump current.

9.2.3 Vacuum system based on turbomolecular pump e principl e Th moleculath f o e r pume facs th s parbasetga i p thae n o dth -t ticles to be pumped receive, through the impact with the rapidly moving surfaces a orotorf n impulsa ,a require n i e d flow direction e e rotorsurfacth Th . f ,o e usually disc shaped, forms e witstationarth h y a surfacstator f o e, intervening s transportei space s e backinga n th whici s e o t th dhge origina th port n I . l Gaede molecular pump and its modifications, the intervening spaces were very narrow, which led to construction difficulties.

wttr STATOft OCCS

ftOTCftOKC

EXHAUST

Fig.71 Schematic diagram turbomoleculare th f o pump

113 Using a turbine form of blading of the rotor, the so-called "turbomolecular pump s developewa " a technicall s a d y viable pump. A typical double flow pump can be seen in Fig.71. The rotor revolving at a very hig hy b moleculaspee - d r carrieai r e impactth s - s toward e backinth s g pressure channel. In the case of lower vacuum needs - e.g. for neutron generator - s mainly rotary pump e ar utilizes s a dbackin g pumps e turbomoleculaTh . r pumps have advantagth e e ovel diffusiooi rn gette io s wel a s nra l pump n mani s y respects t pu int :e o b operatiothen ca y n quicklminutes)w n fe onl(i e a yyth ; vacuum system is much less polluted by oil vapours; an unexpected exposure of the vacuum to the atmosphere does not damage the pump.

Important instructions on the use of turbomolecular pumps are as follows:

a) It is worth while to insert a cooled trap between the turbomolecular pump and the rotary pump; this improves the ultimate vacuum and protects e th turbomolecula e th rrecipien pum d an p tl vapoure oi th froe f mth o s rotary pump. b) Before starting up the turbomolecular pump, it is important to check the operatio f o waten r cooling n I mos. t vacuum systems, water cooling starts automaticall e operatio th e turbomolecula th d an f yo n r - pumin s i p terlocked by a pressure switch on the drain leg of the water cooling pipes. If the water cooling is not automatically switched on by the forevacuum pump, wait until the pressure in the vacuum system has reached the forevacuum levele interlocTh . k e forevacuuth circui f o t m controller will inhibi e manuath t l switch-o e turbomoleculath f o n r pump. c) The use of an isolation valve (quarter swing or butterfly valve) is advisable between the inlet of the turbomolecular pump and the vacuum e manifolneutroth f o dn generator. This butterfly valv s i commoe n i n most neutron generators. The operation of the isolation valves is manual, electromagnetic or pneumatic. For smooth operation of the neutron generator, the vacuum in the manifold (and in the acceleration tube, beam line, etc.) shoule th housine e turbomoleculakept b t dth bu f ,o g r o t pum s ha p e venteb o t atmospherd o t avoi el infiltratiooi d n from bearing lubricat- nitrogey dr f o s advisable i ne us g oilse in Th .. d) After switch-off e turbomolecula th e rototh f ,o r r pump remains whirling foa perior f o 10-1d 5 minutes. During this s i timeadvisabl t i , o t leavee the rotary pump running and only at the end of this period to vent the pump housing to atmosphere. Experienc s showha e n that uninterrupted operatio f turbomoleculao n r pumps gives a longer lifetime than frequent switching on and off.

114 It is important to know that structural changes may take place in the bearing materials of a turbomolecular pump even if the pump is not used. Therefore, it is advisable to review bearings every three to five years.

3 PRESSUR9. E (VACUUM) MEASUREMENTS

Neutron generators of various types use the following vacuum gauges:

1. Thermal conductivity gauges (Pirani and thermocouple gauges) for forevacuum measurements; 2. Ionization gauges (thermionic ionization or Bayard-Alpert type and cold cathode or Penning type ionization gauges) for high vacuum measurements.

The principles of these gauges are summarized in the following.

9.3.1 Thermal conductivity gauges s i wel t I l known r thahighefo t r pressures (greater than about 10-2) 0mb the thermal conductivity of a gas is independent of pressure (in this case the mean free path of the molecules is much smaller than the dimensions of a typical vacuum vessel). At lower pressures (from 0.5 to 5x10"* mb), the thermal conductivity of a gas is proportional to the gas pressure. The Pirani and thermocouple gauges operate efficientl t suca y h pressures.

W. GAUGE HEAD 3 R BRIDG , R2 , ERl RESISTORS P1,P2 BALANCE POTENTIOMETERS M AMMETER

Fig.72 Circuit diagram variablea f o resistance thermal conductivity gauge (Pirani vacuum gauge)

115 The Pirani sensor consists of an electrically heated filament (the heating is direct electric curren - sometimet s regulate e Fig.72)se e - temperaturedTh . , d thereforan e alse resistanceth o e filamenth f o , t depende thermath n o sl conduc- e surroundinth tivit f o y g air e resistanc.Th e e e changfilamencalb th f n -o eca t ibrate r pressurefo d e thermocouplTh . e gauge e operatebasi th f simila o sn o s r principles, wher e temperaturth e e filamenth f o s i emeasure t a thermocoupl y b d e [66]. Both type f gaugo s e hav n accuraca e f abouo y 10%± t . In the case of a neutron generator, Pirani or thermocouple gauges - as forevacuum meters - can be connected as follows: e acceleratinClosth o t e g tubr measuremenfo e e forevacuuth f o t m (usualln o y e vacuuth m manifol e neutroth f o d n generato; ) r - Between the high vacuum (e.g. oil diffusion, turbomolecular) pump and the rotary pump for controlling the forevacuum needed for the high vacuum pump; - Somewher e targeth o t te chambe r measuremenfo r e forevacuuth f o t m aftea r target exchang e emaith whilne th evacuu m e acceleratoth syste f o m s i r running (see Fig.67).

Cathode Ion collector

Anode

UK UA (+50VH+200V)

Fig.73 Operation principle of thermionic vacuum gauge current;ion electron= = (I f current)

9.3.2 lonization gauges

) Thermioni(a c ionization gauges The construction of the thermionic ionization gauge is similar to a triode electron valve (see Fig.73) e operatioTh . n principle followingth s i e : Electrons emitted by the incandescent cathode are accelerated towards the anode and they ionize air molecules along their way. Positive ions produced in this process are collected by the grid having negative potential and, therefore, the grid current intensity in the circuit will be proportional to the gas pressure in the gauge.

116 " mbae norma10 Th - r range0 l 1 n generali operatio , e t th A . n ni rang s i e 3 8 higher pressures e incandescenth , t filamen y burma ne t otheth out n r o ;hand e th , extension of measuring ranges towards smaller pressures would be limited by secondary electrons released from the collector by X-rays generated by the electron bombardment of the anode. The electronics, connected to the gauge, usually has built-in filament protection, which interlocks the switch-on of the gauge whenever the pressure is higher than 10 mbar.

+ 2kV

Fig.74 Operation principle of cold cathode ionization vacuum gauge

(b) Cold cathod r Pennino e g ionization gauge The construction of this gauge can be seen in Fig. 74. The operation princi- s i vern source io e y pl e similath s wito e t samrth h e name. Electrons emittey b d cold emission from the two flat cathodes as a result of having some kilovolts on the ring anod s emolecule ga ionize gauge th th en e i s heade positivTh . e ions produced in this way run up towards the negative electrode and therefore they change - proportionally to the gas pressure - the electric current in the anode or cath- e circuit od e magnetiTh . c fiel B d(generate a permanen y b d t magnet) forcee th s electrons betwee e cathode th e anod th nd an e ont a olon g helical patd therean h - fore the efficiency of ionization and the sensitivity of the pressure measurement increase (Penning principle, like the ion sources of the same name). The ionization vacuum gauges with cold cathode can generally be used within e rangth e " fro" mbadow10 10 m o rt n havin a g direc3 t meter 7readout; however, this range can be extended by the use of current amplifier down to as low as 10"12 or even 10"13 mbar. In spite of vacuum gauges with hot cathode, the cold cathode ionization gauges have the advantages of a much higher lifetime and needing simpler electronics for operatio d readoutan n .

117 On the other hand, there is the disadvantage that, from time to time, the gauge needs cleaning. This is because the cold emission depends strongly on the cleanlines e cathodth f o se surface. Meticulous care shoul e b takend , especially in vacuum systems with an oil diffusion pump. Oil deposits and any other contamination on the surface of a cathode decrease emission capability, and therefore highe d highean r r vacuum wile "measuredb le th contaminate y b " d gauges i t I . worth while to clean the dismantled gauge head by washing it with hot water and household detergent, distilled water, alcohol and any other organic solvent.

9.4 PRESSURE MONITORING AND LEAK DETECTION

9.4.1 Leak rate measurement The pressure in a vacuum chamber that is to be evacuated by any pump sys- wilm te l decrease with time (after switchin e system)th n e curs o gshowth a , -n o n e connectionth l e systeal e th vegasketn i Fig.75th t sf a mo e s wel f ar I s. l fitted to result in hermetic sealing, and the inflow is very small, the pressure will decrease froe initiath m p l value s describea , y curvb d e (1)e pressurTh . e decrease continues until the ultimate pressure - whose value depends on the pump- e inhermeticitth g speee vacuud th an f do ym syste s i reached m- e timTh . e necessary to attain the ultimate pressure is determined by the pumping speed of the pump.

t Fig.75 Pressure change vacuumn i chamber pumpings v time

Curv ) show(2 ee cas th s e whe e usuath n l ultimate pressure reacheb n - ca la ed ter: after switch-on of the gas ballast of the rotary pump. This phenomenon indicates the presence of different vapours. Curve (3) describes a situation when the ultimate pressure can't be reached even afte s ballastinrga switchine th g n o valveg . This indicate a highes r than usual intake from the outer atmosphere. In this case, the pumping speed of the t higno pum hs i penoug o t decreash e furthe e e pressure th b vesseth r o t n i l e evacuated. The dynamic equilibrium will be reached when the gas outflow - produced by the pump - is the same as the leak current, that is the intake from the outside.

118 In cast advisablno e s (3)i o t t leav i e,e vacuu th e m syste n i moperation A . leak test shoul e carrieb d d out e defectiv:th searc r fo h e gasket d changan s e them r properl(o w ne y foa sealingr ) e installationeutrow oneth ne t a A !n f o ngenera - tor, after assembling the vacuum system, it is very important to determine which of these three cases exists. This kin f pumpino d g speed determinatio a tes s ta - n - may also be needed several times, especially if some changes have been made on the vacuum system of the accelerator (e.g. inserting a new cooled trap; replacing valves; dismantling, cleanin d re-fillingan a oily diffusio f (b go ) n pump; etc.). e besTh t metho o chect d e hermeticitth k a vacuu f o ym systey b measurin s mi g the leak rate ,e don b whic s followsn a e ca h : Measur e pressur a momenthed th et a an n ^ | p t et separat e pumth e p froe mth evacuated vessel e pressurTh . e wil lr intakai increas e eth ethrougo t slowl e hdu y the leaklaterd y measur t ma an momen,a , e pressure w th e^ i t ^ p e definitioy r b intake ai s i e , /? ,Th n

where V is the volume of the evacuated system. If the pressure difference is c measured in mbars, the time in seconds and the volume in litres, then /J will be obtained in mbar 1/s. In practice (utilizing a pump of usual power and a vacuum system with rubber r s o seald i lowe equagaskets)n an orde(i o ? st / f rl o magnitudef i r , ) than 10"* mbarl/s, then the hermeticity of the system (i.e. the effectiveness of the sealing) can be considered acceptable. If the hermeticity of the system is not good enough, further leak tests shoul e carrieb d d out.

9.4.2 Pumping speed measurement verA y importan e vacuut th par f o mt system desig s i ndeterminatio e th f o n pumping speed, which must be known in order to attain the required ultimate pressure. a basiThi s i sc problee differenth t a m t accelerators n ,sourc io wher e eth e naturally consumes gas. When a new neutron generator is put into operation - and also later, when a modification has been made - the pumping speed of the system should be tested. When connecting a pump to a vacuum system of volume V, the pressure change with time during operation will be:

— — S = (28) V e pumpinth whers i S ge speede negativTh . e sign correspond e facth t o t s that

119 the pressure decreases with time f I constan. t pressur s i maintainee a needly b d e valve controlling the air intake from the atmosphere, the pumping speed will be:

dV S = ar (29)

In practice, when the pumping speed is to be determined, a definite change in atmospheric pressur r volumai e e durin e correspondinth g g time b e o intervat s ha l measured e donb . y usin Thi b n ee ca th equipmengs t show n Fig.76i n . This setus i p also use r measurinfo d n sourcgio s consumptioga e n [66].

ext

Fig.76 Equipment for the measurement of pumping speed

The pressure p in the vessel under evacuation is adjusted by a needle valve positio N n equilibriue I valvO th . e V f eith o n e pummth p remove e amounth s f o t air inflows into the system through the needle valve. Turning the valve V in the F positionOF r U froshapeai e lef, th me t th d sid f o glase s tube will flow into e vesseth l under evacuation; therefore e liquith , d level will risn thio e s side. Usin a gstopwatch e timth , e e intervameasureb n ca l d during whic e liquith h d level rises from the lower to the upper mark in the U shaped glass tube. The volume - havin e externav th gr oai f l atmospheric pressur - ebetwee e mark th nn als ca so be determined. t i follow, pV Fro e expressios= th mv thae r th thavolump ai t nt f o e GXl flowed inte vesseloth :

ext V = (30)

120 and therefore the pumping speed:

Pextv

Typical pumping speed versus pressure function e ar shows n Fig.7i d n an 7 l diffusiooi n a a Fig.7rotarn d r an pumpfo 8y , respectively s i characteristi t I . c for both cases thae reath t l pumping speed become e sultimat th zer t a o e pressure, i.e. when the pumping causes a dynamic equilibrium: then the pump removes just as much air from the system as the air intake due to the ineffective sealing. The pumping speed is given - for both types of pump - at the constant section of the pumpin a rotar gr speefo a ydiffusio r h pum/ dfo m functios pn 1/ n i pum n d (i n an p in general). These pumping speed value e manufacturer - givesth y b n e measurear e - sth t a d inlet of the pumps, but the realistic pumping speed far from the inlet is below this catalog value because the resistance of different elements of the vacuum system decreases the pumping speed. This can be seen in Fig.79, where changes in the pumping speed functions for an oil diffusion pump are shown with different cooled traps. For neutron generators, it is advisable to carry out the pumping speed measurements of the vacuum system according to the arrangement shown in Fig.80. e mosTh t importan o chect t l tasal k s firsi kf whethe o t e requireth r d vacuum e guaranteeb e n availablth ca y b d e pumping system alon e wholth g e e lengtth f o h accelerator tube (having dimensions negligible compared to the mean free path) wite deuteriuth h m consumption necessar r normafo y l operatio e acceleratorth f no . For pumping speed measurement it is advisable to connect the U tube to the n amoun a n sourca needl io t ga a f le e deuteriuso vi e tth e o inlet f valvo d t an me gas flow into the system which corresponds to the actual gas consumption. If, for economic reasons, T>2 cannot be used, hydrogen gas may serve the same purpose, like at the installation of a new or repaired neutron generator. Most commercial neutron generators with a deuteron beam of 1-2 mA have a deuterium gas consumption equivalent to 4-5 cm /h at NTP. Therefore, the pumping speed measurement shoul t e s carriea thiinleb d ga t s t ou dwit h very great precision. The pumping speed of the pumping systems can change for the following reasons: - Decreased heating power of the oil diffusion pumps: This can be an electrica e lth problefailure t y arisbu mma , e e wateth eve f i nr cooling loop is very close to the heater of the pump or if the oil level in the pump is

121 0 1 ' 0 1 - 0 1 10 10" PRESSURE (robar)

Fig.77 Pumping speed characteristics of a twin stage (DP) and a single stage (UP) TUNGSRAM rotary pump [65]

in3 CRYSTA0 16 L

XX I ~~X ^ / CRYSTA0 10 L CRYSTA3 6 L "">»

' /,'' NN J f 10' / If I / II I

If 1 I m 10-' 10-' 10-7 10-6 10-5 10"* 10-3 10-2 PRESSURE (mbarl

Fig.78 Pumping speed vs pressure characteristics of ALCATEL diffusion pumps [64]

1600 ~1500l/s A 1200 -9001/s B 800 ~700l/s C

400 0 K 4 p(mbar)

Fig. 79 Pumping speed of a diffusion pump (A: without traps, B: with water cooled trap C: with water and liquid A cooled traps)

122 Pirani gauge ION ACCELERATOR SOURCE TUBE

HYDROGEN GAS NeedlH e valve Valve OIL DIFF PUMP

Fig.80 Arrangement pumpinge th r fo speed measurement neutrona n o generator

less tha s neededi n . Norma l leveoi ls 8-1i l higm 0m h measured froe bottomth m e diffstackoth f . n rotarI - y pumps e loweth , r pumping speed stems mainly from problems wite th h oil level, so this should be checked regularly. n getteio n rI - pumps e contaminateth , d cathode surfaces (mainl y oilb - y de ) crease the electron emission, which may cause a fairly large decrease in the pumping speed.

9.4.3 Leak detection As mentioned in Section 9.4.1, if the measured intake exceeds a value of about 10" mbar 1/s, then the vacuum system is not tight enough and therefore a leak detection becomes necessary. There are many technical methods available to detec a leakt , starting from single vacuum e gaugemosth to t smoder n mass spectrometers [67]. However, before startin y leaan gk detection s i wort t i ,h checkin e wholth g e system again. Most problems are caused by rubber gaskets (it may happen that the gaskets are not fitted properly to the metal surfaces in the joints, or the gaskets migh e missinb t y mistake)b g s i advisabl t I . o chect ee e th th kvalve d an s air admittance valves A particularl. y careful tesy tan shouln o e carrieb t d ou d e neutroth par f o tn generator where some alteratio s beeha n n made recently, e.g. target exchange, exchange of glass balloon in the ion source, placing or replac- w vacuune in a g m gauge ,e checetcth f kI . e seemunsuccessfulb o t s , some method of leak detection should be carried out. Some method e easil b - whics n y ca haccomplishe a neutro n i d n generator laboratory - are described below.

123 The simplest solution is the use of a vacuum gauge, which is a normal acces- y vacuuan f sor Pirano n i my e us isystem gaugee r Th lea.fo s k detectio s i basen d on differences between thermal conductivities of different gases (e.g. the conductivit f hydrogeo y s i mucn h higher than thaf air)o e Pirant Th . i gauge usee th s thermal conductivit e pressurth n i y e measuremen f gaso t : therefor e th - leae de k tection should be carried out as follows: - Usin a gcylinde f hydrogeno r , wite th hcorrespondin g pressure regulatora , ^ shoulH narrof e blowo b d t wje n e vacuuont th e wal f th oo lm vessel, where th e leak e beinar s g looked for. Test e carefulljointse th e Th gasketth . d y an s hydrogen creeps inte vacuuth o m vesse t thaa l e twal th are lf o a wher a leae k exists. Therefore e Piranth , i gauge will sho a w highee th r o t pressur e du e higher heat conductivity of the hydrogen, indicating the position of the sus- pected leak. - lonization vacuum gauges may also be used to search for a leak if vapour of some organic liquids (e.g. ether a s "tesi use )s a td gas" n I thi. s casef o , course a vacuu, e attaineb m o t valu ds ha t whicea leas s i th withi e rangth n e 3 of applicability valid for the given gauge (this value is between 10" Q " mbar10 d , an dependin e typ th f gauge)o en o g . Both methods described above are suitable for a quick search for rather coarse leak sites (i.e. intens r intakes)ai e ; this quicd rougan k h e methoth s i d most expedien o detect t t whas happeneha t a neutro o t d n generator during dismantling and reassembly.

Vacuum chamber

Rubber ring ——. 1 HALOGEN x sssfcHI LEAK 2^^-JJ—— — DETECTOR FREON \ anrt Pirani GAS Pennin\ /T g ISOLATION aauoJ ^ \ TeT gaug/ ^O e VALVE KjU*1 Needle valve vent valve Piran1 i gauge OIL DIFF ® J PUMP \ ts3Li Col< i d 1 Rotar I P y — "-PNJ- trap -P ^ / pump Valve Vaive

Fig.81 Vacuum test stand with halogen leak detector

124 n getteWheio n ra n pum e vacuu s th usei p n i dm system, this pum n alse ca pb o used for leak detection. This is because the pumping speed of an ion getter pump is different for different gases: for example, it is three or four times smaller for argon thar oxygenfo n . Thereforee th t je f i argoe s , tes th s ga i usen t n i d ammete e pumth f po r power supply will sho wa highe r power consumption. Leaks causing smaller inflow - whics h canno e detecteb t y commob d n vacuum gauge - s wil lo t mechanicaoccu e du r l imperfections (e.g. imperfect welding). This could occur e.g . target-holdew whecolw ne ne da n a tra s r i assemblepo r d onto the vacuum system. It is advisable to test them - with the help of a separate simple vacuum system - before installation. Such a vacuum system equipped with a halogen leak detector is shown in Fig.81. e measurinTh g e halogegaugth f o en leak detector contain n indirectla s y heated platinum probe - like a cathode - emitting alkali ions, and their current is measure n ammetea s indicatei y b r do y audirb dn oemissioio signal - e in Th n. creases considerably whenever halogen gas gets into the measuring gauge. This phenomenon can be used for leak detection in any vacuum system. The test gases containing halogen are usually fluorocarbons which may be easily purchased on the market. FREON-12 is the usual gas used in cooling machines, refrigerators and air conditioners a boilin s ha g t i poin: f 30°Co t , while FREON-11 a liquid s i 2 , wita h boiling poin f 93°Co t e utilize.b Bot n r halogeca fo hd n leak detection e FREOTh . N liquid or gas gets easily into the vacuum chamber through the leaks in the wall (and improves sealing) of the vacuum vessel from the outside and therefore an increase n currenio d t wil e detecte b le electronics th y b d . In order to achieve the highest possible sensitivity in the leak detection process, it is advisable to place the halogen gauge as close as possible to the vacuum vessel to be checked for leaks (short interconnection). Vacuum gauges mounted on a vacuum stand can give - even before the halogen-using in-situ leak detection proces - rougs h preliminary informatio e th expecte n o n d t i leak f i , exists. A needle valve can be used to simulate a leak, and therefore to check the leak detector.

Aluminium / holder o - ring

Fig.82 Typical O-ring seal for interconnection

vacuuThe m testevessebe to dl should courseof , welbe , lvac the fitte -to d e lea th system k a s wel uu i r advisablr teste thi mfo t lFo i us .s fittin o - t eO g ringr examplfo , e likon ee that show n i Fig.82n e diamete e s Th rini .th g f o r

125 e th f o p to e openin e th th determiney b t e diametea ge vesseth th d y f an b o l dr vacuum test stand. High vacuum plastics (like Viton - availabl) e fro a mcoupl f o e manufacturer e useb s wella dn ca . - s It shoul e noteb d d thae halogeth t n leak detecto n alse ca connecteb ro o t d the vacuum system of a neutron generator. The most advantageous place for the gauge is near the forevacuum pump. In this case, the test gas will probably reach the detector. The most sensitive leak detection can be performed by a mass spectrometer leak detector, n essentiawhica s i h l piec f o equipmene n i almost t y laboratoran y involving vacuum equipmens associateit d an dt technologiesn a e . b y Thima s existing residua s analyzee ga leaH l kr o rdetector , accordin a o t whetheg s i t i r magnetic secto r quadrupolo r et i typedoet t neeno bu s,a verd y high resolution because it is only necessary to separate the two most commonly used gases, hydrogen and helium, at mass numbers 2 and 4 respectively. e conventionaTh l mass spectrometer leak detecto s i normallr a yportabl e unit having its own vacuum system including a rotary and a turbomolecular pump (liquid nitrogen trap), gauges and electronics, as shown schematically in Fig.83.

nass valve spectrometer Neutron 1 (jenerator h——dP U*sT— Electronics -o- panel Cold Penning trap ——^ \ / ft gauge —.^N^™ Speed control He Rotary Backing gas pump Valve valve OIL DIFF -H^- -*- PUMP Cold Vend Y f trap valv? eZi

Fig.83 Utilization of mass spectrometer as leak detector T=^^-TI L-L Gas intat I CollecHCollec m

Focusing electrodes

Fig.84 Schematic diagram quadrupolea f o mass spectrometer

126 e mosTh t specialized mass spectrometer leak detecto s i usuallr y quadrupole, in view of its small flange-mounted gauge head and easy operation. The ion beam produce y electrob d n sourcnio collisios e i accelerateeth n i nd injectean d d into a quadrupole separation system having four electrodes of hyperbolic cross section ( see Fig.84 ) [69]. A constant high frequency voltage U (f = 2.5 MHz) and a superimposed DC voltage V are applied to the electrodes. As the value of U and V are increased simultaneously, the mass will be scanned from an initial value up to a maximum, whereby the ratio U/V must be kept precisely constant. e masTh s numbee ionth sf o emerginr g throug e separatioth h n syste d beinman g detected must satisfy the following conditions:

M = —————————— (32) 2 2 7.2 f rQ

wher s hale i distancth f r e e betwee opposito ntw e electrodes. e lineaa resulth s f A ro e tvoltag dependencth a lin , d -V ee an mas th M f s o e ear mass scale may be obtained by a linear scan of V and U. e n sourcth io spectromete n e i filamenee Th th f o s ti rswitche n wheo e dth n A n vacuuow pressur e s vacuumit th n systei " ed mmba 10 an s mlesi x r s5 tha~ n system of the neutron generator under test is to be tested with a helium gas jet. e lea s Th i detecteke audi th y ob d indication e analoand/oth y b gr e meteth n i r readout. The sensitivity of this detector can be improved by increasing the gain of the electrometer amplifier and/or reducing the effective pumping speed of the high vacuum pump by adjusting the speed control valve but still maintaining the maximum pressur t whica e e filamenth h t operates.

127 10. BEAM ACCELERATION AND BEAM TRANSPORT SYSTEMS

10.1 ELECTROSTATIC LENS The electrostatic lens is the most usual focus device in low energy accelerators and neutron generators. The energy of the ion beam extracted from the ten w f keVfe o ssourc a . s i Thie se focuse b bea n mca dn electrostati a easil y b y c lens. The magnetic lens is very rarely used in neutron generators. The preaccel- eration lens is usually an electrostatic diaphragm, immersion or unipotential lens. The unipotential lenses is also known as the Einzel lens (its original German name) [69]. A typical single gap diaphragm lens acceleration tube is shown in Fig.85. This acceleration tube has an extraction gap (focusing gap: due e focusintth o g behaviou e th firs f to r diaphragm n acceleratioa d an ) n e gapTh . V acceleratiok e 0 electrodeth 15 f sizo r ne fo svoltag s i indicatee e th n i d figure [70]. An immersion lens is an electrostatic lens which has equipotential regions on both sides of the lens. It can be constructed of diaphragms or cylinders. Fig.86 show n immersioa s n lens consistin o equatw f l o gdiamete r cylinders. The immersion lens - depending on the polarity of the voltage drop of the lens gap - can be analogous with convex and concave or concave and convex lenses [71].

Ion source base AcceleratioV H n

Vacuum Focus gap manifold

"Extraction canal 100mm

GND

Fig.85 Immersion lens equivalent single gap acceleration tube for 150 kV neutron generator

128 U,>U2

Fig.86 Electrostatic immersion lens (consisting of two cylinders) with its optical analog

10.2 UNIPOTENTIA EINZER LO L LENS The Einzel lens is an electrostatic lens in which the energy of the beam doet changno s e because symmetricath f o e l potentia e electrodesth f o le uniTh .- potential len s n (electronfocuseio e th s ) beams with positiv- r po negativo e p U e e middltentialth f o es electrode. Fig.87 represent e opticath s l analog convex- concave-convex lens system of the Einzel lens for a deuteron beam. The Einzel len s i symmetricals e foca th e :imag th le objec th lengtd e an n tsido h e have th e sam d focusinf evalue e f focaTh . ge th l powe a lengtd unipotentia r an rfo h l lens is show n Fig.88i n .

Fig.8 e unipotentialTh 7 (Einzel) lens with opticals it analog

129 I ••

I I 0.6 - - I 0.5 -•

Fig.88 The focusing power and focal length of the unipotential lens

The focal lenses of the low energy accelerators (neutron generators) are usually immersion or unipotential lenses. In a unipotential lens the diaphragm or cylinder close to the ion source serves as an extracting electrode. The exit electrode of the unipotential lens is connected to the potential of the first electrode of the (multielectrode or homogeneous field) accelerator tube. The focus voltag e acceleratioth e n sourco s i e n high voltage terminal (see Fig. 11)s A . the focus electrodes and the accelerator tube electrodes should be aligned, the beam position at the exit of the acceleration tube is almost off axis. The steer- d positionininan g e beath mf o g after acceleratio s i usualln- bi y y carrieb t ou d n sourcesio e ase b e P currensub dth D n , ca quadrupold - tan G PI , e lensesRF r Fo . stantially increase a wid n i ed rang y b eincreasin e extractinth g g voltage [72], e i.ecurrenth . s i proportionat e energth f o o extractet yl d ionsr typicaFo - . ex l traction geometries e th maximu, e mth relatio s i voltag givey b V nn e [73] spacinp ga e centimetersn i gth , wher s ' i d d e 0 1 .x 5 V= 4 1/2

10.3 TROUBLESHOOTIN ELECTROSTATIF GO C FOCUS LENSES If the properties of the electrostatic focus lens are unsatisfactory the following shoul e testedb d : - Vacuum in the system, - Focus voltagfocuthe s on eelectrode , - The correct operation of the ion source and the extracting system.

130 Bad vacuum (oil vapour, high pressure in the focusing space) can easily lead to surface contamination of the focus lens electrodes. In the case of high pressure, a glow discharge can be built up between the focus electrodes and can contaminat e e electrodessurfacl th th vapoure oi e f cas o th f e eo surfac eth n ,I . e layer may insulate the electrode (silicon oil!), and the distribution of the potential betwee e lenth n s electrodes wile changedb l . Thin leaca sd e th easil o t y deterioration of the focusing properties of the lens. Observation of the focus electrodes - after a long term operation of the vacuum system of the accelerator s i important- e cas th f blackish-browo en I . n electrodes, clean them with organic solvents and polish their surfaces. After dismantling the focus lens, the polished surfaces shoul e washeb d f possibl(i d n organii e c n ultrasonia solven d an t c bath) and carefully reassembled in the vacuum system. Test the high vacuum feed- e througfocuth f so h electrode e th surfac f I .s i contaminatede , e cleafeedth n- throug s wella h . Tese conductivitth te isolatioe th assembleth d f an o ny d system. A thick conducting carbon layer on the surface of the insulator of the focus electrode (from hydrocarbon oil contamination in the system) may produce a bypass currena e changfocus th d n an i e tefocu voltageth s s poweA . r suppl s i usually y protected against breakdown o t discharge e du s s betwee e focuth n s electrodese th , voltage test of the focus voltage may give an indication on the surface contamination in the focus electrode region. If malfunctio e focuth f so n power suppl s i detectedy , troubleshootind an g repair should be done on the basis of the instruction manual, in which the necessary procedures will be described in detail in the section on high voltage power supplies.

10.4 THE ACCELERATION TUBE The acceleration tube of neutron generators consists of annular electrodes separated by a hollow cylindrical insulator of glass or ceramic. The insulators and the metal electrodes are bonded together either with epoxy resin or with polyvi- nylalcohol (PVA) glu o t fore m vacuum tight joints. Ther e variouar e s simpl- de e sign r acceleratinfo s g tubes with homogeneou r inhomogeneouo s voltw s lo fiel- f o d age accelerators [74-77]. The transport of the beam over long distances is simple if the currents do not exceed a few milliamperes. The homogeneous field acceleration tube, i.e. the multigap tube, consists of more acceleration gaps using diaphragms, cylinders, hollow cones, etc.n a s ,electrodesa y b d fe , whice ar h equiresistance voltage divider resistor chaine inhomogeneouTh . s fielr o singled , two, etc., gap acceleration tube consists of cylindrical or conical immersion lenses (see Fig.85). This tube is common in fixed acceleration voltage machines such as the sealed tube neutron generators. Two-gap acceleration tubes are used

131 in TMC A-lll and KAMAN A-1254 neutron generators. The first gap is for focusing and the second for acceleration. Recently, accelerating tubes with either a single-gap [74,78-80] or special multigap arrangements [81-84] are used in improved neutron generators. The single-gap, high gradient system requires particular attention to the purity of vac- e th manufacturin uud man e electrode.th f o g During operation a pressur, e higher than 10" Pa must be assured in the cavity, free from hydrocarbon molecules. A carefully polished and cleaned titanium electrode is needed to maintain a field strength of 200 kV/cm. Secondary electrons - ejected by positive ions from the neutrae electrodth n o l e bea atometh d msurfacan e n acceleratei sar - e d towards the high voltage terminal e intensitiesTh . , maximum kinetic energies d meaan , n energy value f electrono e sdeduceb n ca sd froe measurementth m e bremsth f o s- strahlung. These data can give information on the optical behaviour of accelera tion tubes with different field structures [85-86]. The acceleration tube in a neutron generator also: - Accelerates the extracted D ions; - Keep bean s io together focusem o D ; e it sth r - Holds the vacuum in the neutron generator; - Holds (sometimes e voltagth ) e divider resiston sourceio e rth . chaid an n The electrodes should screen the insulator walls of the tube to protect them from oil vapour or carbon layer deposit due to the ion beam scattering on the oil vapour in the vacuum. The manufacture of an acceleration tube requires a well equipped workshop with special tools for glueing and bonding the components of the tube as well as for its alignment. The bonding is usually carried out under pressure. The use of PVA requires a polymerization process under high temperature in a special oven. e resi d glue Th squeeze an nar o efor t t a mfillee ou th de insid t th d bot an n eo h e metaoutsidth f o l e insulator joint. This n fillecausca t e problems becaust i e is situated at the point of the highest electric tension and because it outgasses into the vacuum during the operation of the tube. In some tubes the accelerator electrodes are not flat diaphragms but are strongly dished or are cone shaped to shield the fillets and insulators against the electric charges, scattered oil and other vacuum dirt. The electrodes of an accelerator tube - especially the multielectrode homogeneous field accelerator tube - wils l increas e vacuuth e m resis- e acceleratoth tanc f o e r a vacuutub s a e m n sourccomponentio e eth (ths s A .ga e inlet) is situated at one end of an acceleration tube and the vacuum pumps are situated at the other (at the ground potential), the accelerating electrodes of an acceleration tube are usually perforated to form a lower vacuum resistance e vacuu between sourcth io d e man enth pumps.

132 200mm

m m O H Steps in the rings

Fig.89 The structure of a homogeneous field accelerator tube

The structure of a homogeneous field acceleration tube is shown in Fig.89 [69]. This tube consists of conical acceleration electrodes and ceramic insulating rings. The cone shaped electrodes protect the walls of the ceramic insulator rings against contamination. e tube'Th s lifetim e s i limiteth e r fo d following reasons:

- Secondary electrons produce y ionb ds froe th mresidua s ga (nol t onll oi y vapour!) in the acceleration tube accelerating and bombarding the electrodes and insulators; - Electrons bombardin e e bonepoxth th g filledt A ya betweePV tresi r e o nth n insulator e metath d l an electrodess ; - Poor vacuu e tubeth mn ;i - Heavy ion r moleculao s r ions forme r electroo n d io frone mth bombardmen f o t residual gas and from electron bombardment causing sputtering of the metal accelerating electrodes; - X-ra r ultravioleo y t photons causing photoelectric emissio f electronsno ; - Field electron emission from the metal electrodes.

s i probablIt f etheso l thaeal t processes contribut a greate o t r elesseo r r e operatioth exten o t tf o conventionan l acceleratio s i n clea t i tubes t r bu tha, t electrons, whatever their source, plae majoth y r e rolee lifetimaccelth Th . f -o e eratio ne increase b tube e followinn th ca sy b d g means:

133 - Maintaining a good vacuum in the acceleration tube in the order of x 10~ 3 o 10"t 1. 5; 3Pa - Using the electrode shape which shields the insulator ring and bonding fillets fro e electromth r heavo n n bombardmentio y ; - Constructin e electrodth g e system from aluminium; - Using the secondary electron suppressor -300V to -400V before the target to protect the residual gas and electrodes from secondary electron bombardment. f o deflectin e us ge Th magnet- r o electrostatis c deflector platey protecma s t the accelerator tube as well [87].

10.5 TROUBLESHOOTIN ACCELERATIOF GO N TUBES n improperlIa n y working acceleratio n beanio mtube th ecurren w tlo wile b l or diffused e firs n Th source .tio o e chect don b d thin e tes s i an th o .ekt tg The steps are as follows: 1) Chec e vacuumth k . 2) Chec d e tesresistoan th kt r e acceleratiodivideth f o r e nresis th tube s - A . tance between two electrodes is usually in the order of 10-100 MQ, the use of a megaohm-meter is recommended. Check the contacts between the electrodes and e resistoth r chain A . faulty contact betwee e voltagth n e e th divide d an r electrod n beamy io e disrupisolatema th e : e th td accelerating electrode will e chargeb a positiv o t d e potentia t i wil d l an l deflec e originath t n beaio l m fro e originamth e beal th axi mf o s line. ) 3 e Tesinsulatioth t n betwee e acceleratinth n g electrodes wit a high h voltage insulation tester. Testing voltage of several kV is recommended. In case of sparks alon e e th th gsurfacisolato f o er rings, some meta r carboo l n tracks can be observed. Clean the outer surface of the isolators with polish- ing paper and organic solvents. 4) Observ e inneth e e acceleratior th sid f o e n tubee th n .sourc io Whed e an eth n extraction system can be easily dismantled, open the ion source end of the acceleration tube. Ligh e e th th insidtacceleratio f o e ne th tubeelec f I .- trodes (not only the accelerator electrodes but also the extraction and focus) show some contaminatio l e nforvapou oi th f o brownish-blacsuc ms n i a hr k layers clea e electrodth n e surfaces. Polishin d an washing g with organic solvents (N-hexane, petroleum-ether, acetone, etc., depending on the materials used) shoul e donb d el deposite oi electrodecarefullyth e n th e o fors f Th o m.s n givca e some information about wha n s beamakini tio me th gcurren t decrease. e erodeTh d surfaces indicat e electroe effectth th e f o sn bombardments along the acceleration tubed giv n indicatioan a e, e pooth rn o nvacuu r othemo r contamination sources. Check for continuity between the inner electrodes and the outer contacts.

134 11. PRINCIPLE BEAF SO M FILTERS

11.1 ELECTROSTATIC AND MAGNETIC BEAM DEFLECTION The accelerated deuteron t alsd beaionD an o ms, othen containD~ io r , + D s J Z i i ) froe residuath mO r grease specieo , l lN oi ( s. gase e d vapouThith an s sf o r vapour will cover the surface of the target. The target lifetime in a neutron generator depends very strongly on the quality of the vacuum. The deflected beam result a morn i s e clear emitted neutron spectrum. e beaTh m filter f o neutros n generators utilize electrostati r o electromagnetc - ic separator e straight-linr th o both n i s , e Wien s i welfilter t I l . known that in an E electrostatic field perpendicular to the direction of the ion beam, as show n Fig.90i n e deflectioth , y nalon e plategth f lengto s s givei 1 hy b n

vye = 2m (33)

where E is the electric field strength (E = U/d). For small angles the deflection angls i e 0 eEl Ul (34) e mv where U is the accelerator voltage.

Electrodes

Fig.90 Electrostatic deflection f chargedo particle beams

135 Results sho e wsmalth thay l b tangl e electrostatic deflectio e beath n m components cannot be separated for e/m. However, the neutral beam components (oil vapour, residual gas) will be separated from the ions. Electrostatic beam deflection is normally used to produce a pulsed beam at the neutron generators. e magneticallTh y deflected a circula beas ha mr trajectory e th (Fig.91 d an ) = \/p deflection angle is given by

1/2 = lB(e/2mU0) (35) e magnetith e th cwhers i fiel B ed strength. Equation (35) shows thae deflecth t - tion angle of an accelerated ion beam depends not only on the U acceleration voltage, B magnetic field strength, but also on the specific charge e/m of the given ion. e WieTh n filter utilizes perpendicula # fields B magnetie d Th .an cE r deflec-

e compensateb a - electrostatin e ca a giveth tio n f y o io nnb d c deflectione th o s : m C Wien filter form a straight-lins n selectorio e d . Wheonl= f o a ysingl d nn io e specifi y pasc ma scharg througm e/ ee exie th filtehth t f rsli o t [88] e valu.Th e 2 s constani ov f d depend (B/Ean U tU voltage2e n o s= ) . m : The Am/m mass resolution of the Wien filter - as for other (magnetic) analyzer - s strongly e depende entrancth th widt n f d o o exihsan e et th slitf o s vacuum chambee e filterlengtth th f d .o h an r

Pole

Magnetic poles

side view

Fig.91 Schematic representation of magnetic deflection

136 Magnetic pole piece

Electric deflector

Separated ions

Electric fielr E d 'v

* ) Fm= Fe

Ions with V>VQ Fm>Fe ie light ions

Fig.92 Schematic diagram of a Wien filter

The capability of a Wien filter is determined by D, the dispersion between e selecteth d (M-AM) an mas e e dispersioM filtes th Th s i determine. f r o D n y b d e expressionth :

ld-a-E-AM D = (36) 4-U a-M

where 1^ is the distance of the exit slit from the filter, a is the length of the filter (magnetic poles and static deflectors) and M is the mass number. A schematic diagram of a Wien filter is shown in Fig.92.

11.2 TROUBLESHOOTIN ELECTROSTATIF GO C DEFLECTORS e cas f th impropeo e n I r operatio a stati f o nc deflector (pulser) e problemth , s can be of mechanical or electrical origin. The mechanical problems are mainly in the vacuum chamber of the deflector, so it is necessary to open the vacuum system

137 and inspect the deflector plates. A broken contact or deflector plate holder can e b foun e delectri th easily d can , e contactdeflectioth f o s n voltage shoule b d tested. During conductivity measurements, the occasional resistive conductivity between the deflector plate to the vacuum vessel should also be tested. Sometimes the isolators can be covered by conducting (evaporated metal or carbon) layers, so a proper resistance test should be carried out at high voltage of the same or- e originae valuth th f des o ea rl deflection voltage. Every breakdown betwee e deflectoth n e grounrth platd dan e mean a reflectios n HV pulse along the high voltage cables connecting the HV supply to the deflector chamber. The cable end should be tested carefully because an unterminated cable d coulen d carr a ydoubl y hige reflectioth h o voltagt e n du e fro e shorteth m d cable end. This means the cable short circuits are usually at the HV power supply end (connector) of the high voltage cables. Test carefully the HV vacuum feedthroug r conductivitfo h insulationV H d an y .

11.3 ANALYZING MAGNETS OF NEUTRON GENERATORS e analyzinTh g magnet f neutroo s n generator s followsa e ar s :

- ion species selectors for the D , D~ and D-, components of the accelerated beam; - selector f o charges d particles from neutral particles e.g l vapour.oi , oxygen and nitrogen molecules from the residual vacuum.

As the magnetic field selects the ion species by e/m, even a slight deflection - larger than the diameter of the beam on the original place - may select the charge d unchargean d d component e beam th e selectio f e Th chargeo .sth f o nd and neutral beam components increases the target lifetime, protecting the target surface fro l vapoumoi r contaminatio e vacuuth n i nm system. The analyzing magnets are made of soft iron. The main problem in the construction of an analyzing magnet is procurement of the proper soft iron. Iron sheets with carbon content less than 0.06 % are suitable material for deflection analyzing magnets [89]. The power consumption of the coils is usually in the range of a couple of hundred watts, so correct cooling of the coils is recommended. As the metal wires of the magnet have positive temperature coefficient, regulation of the magnet current is advisable. Two magnet constructions are shown in Fig.93 n (ani Fig.97) d e firss Th i eas .t o t constructy , havin o haltw gf disc shaped pole pieces. This solution allows a relatively high deflection angle, with the given e polTh e . piece ° 70 s- hav0 coi6 a shapelo t dat p eu a tha s easi t o t ymanufac -

138 55

R50

Coils:

Fig.93 Schematics of a 6(f deflecting magnet (sizes in mm)

FILTER VARIAC (I

ADDITIONAL CURRENT REGULATOR

Fig.94 Power supply of the deflection magnet shown in Fig. 93

139 1N 4001

Fig.95 Current regulator circuit magneta f o power supply

ture, and they can be machined on a lathe. The manufacture of the rectangular components does not even require a milling machine. The coil is about 5000-8000 a wirdi e m [90]windingm 6 .1. f o s This coil requires a 150-180 V DC power supply with an output current of 3- 5 A. The power supply can be made from a variac (Trj in Fig.94) controlled isolation transformer (Tr^) ratin kVA1 g e r rectifiercooledai Th . e ar ,s mountea n o d suitable radiator. The wire of the filter coils is made of 2.0 mm dia wire. The condensers are parallel connected and inexpensive (similar to those used in TV power supplies) n additionaA . l current e regulato attachee b systemth n o ca t rd, for example, instead of the F3 fuse. This current regulator may be constructed on the basis of power transistors cooled effectively by a fan. A principle diagram e poweoth f r suppl s showi y n Fig.94ni . The power consumption of the relatively large coil is about 500-600 W in the case of separation of a 200 keV ion beam at a deflection angle of 60°. Water e coolincoith l f o gdoet compensatno s e th e growt th r resistanc fo n e i h th f o e o warmint simplA e coi . du lup ge current regulatoe attacheb e origi th n ca o -t rd - ful l l na wave rectifie - power r suppl , ycorrespondin e marketh o t gd placen i s . Fig94 .

140 RED-I + 12V ilOOnF c 22k 100k -12V M )( 1.5M -W- 2.2k 100nF MH T

100k TL 074

2.2k 2k F n 0 10 SEfl -»-M -w- + 12V <—T 2.2k 3.9k -CD-

33k 100k

2k -9f-

GREEN

IN (.151

Fig.96 Operator alarm circuit for the magnet current regulator shown in Fig. 95 Coils

f Fig.9deflection3( A 7 magnet r neutronfo generators ) [91 (sizesmm ] n i

142 The current stabilizer series circuit is shown in Fig. 95. It is based on the commercial voltage regulator circuit LM 317. The LM 317 load current is shun- e usuaP transistoth PN y l b d r te equivalent complementary Darlington circuite Th . magne t e regulateb currene helica th n ca y b tl d potentiometer connected parallel to the 1 Q reference resistor at the output of the LM 317. This resistor should a hig e hb power resistor witw thermalo h l coefficient e resistoTh . - r as itsel s i f sembled into a cylindrical hole of the radiator of the regulating LM 317 and 2N3055 transistor seconA . d 2N3055 transistor ZeneV (wit 6 3 hr e basidiodth sn i e circuit) protect e currenth s t regulating circuit against transients e voltagTh . e e currenth dro n o pt a regulatovoltmetee reab y b dn ca rr connected paralleo t l the current controlling unit. The optimal voltage drop on the current regulator . V i0 s3 betwee d an 0 1 n An advanced voltage regulator circuit, the LM338K, may replace the LM317 and 2N3055 based circuit in the above described magnet current regulator circuit. As the absolute maximum e voltagratinth f o ge difference betweee th e inpud th n an t e outpusame th e overvoltag s th i ,t e protection (ZF3 d 2N3055an 6 ) e b shoul t no d changed. The operator alarm circuit - connected to one of the above current regula- s showi tor - sn Fig.96i n e circuitTh . n poweV twi, 12 ow npowere s r it supplyy b d , has three voltage comparators, three indicator LEDs and an acoustic alarm device [90]. e voltagAth s e drop acros e LM31th s r LM338o 7 n outpua r Kfo t curren f abouo t t 3-4 A should be more than 10 V during normal operation, the first comparator detects the < 10 V voltages. In this case the green LED lights up and the acoustic alarm gives a continuous alarm signal. These warnings inform the neutron generator operator e tha resistancth e tdeflectinth f o e g magne s increaseha t d e warmin th e coild thae duth voltago an th t ,ef t o g e drop alon e currenth g t regulator is below 10 V. As the output voltage of the magnet powering rectifier can be raised by turning up the variac (in Fig.94), the operator should turn up the variac. If the operator turns it up, and the voltage drop on the current regula- tos highei re alarth , rm V tha signa0 1 n e l th greeD goe stopd LE nsan s outf I . the operator turns more than is needed, and the voltage drop is over 27 V, the alarm starts to sound again and the red LED comes on. As the voltage drop should be between 10 and 27 V, if the red light comes on, the variac should be turned downwards. When the voltage drop on the current regulator reaches the absolute maximum e regulatoratingth f o se ZF3rth d 2N305 an , 6 circuitV) 56 (3 bases d protection circuit start o t shune s currenth t t regulator e thirth ; d comparator will start the second red LED flashing and the continuous alarm signal starts to be pulsed. This indicates that the absolute maximum permissible voltage has been reached.

143 e magneAth s t coi l s resistancit warm , up s e increases; normall e currenth y t regulator circuit should be set below the upper voltage drop limit (27 V). The voltage drop along the current regulator circuit is displayed by the 100 uA analog mete s showa r . Fign i n 95 . A deflection angle as small as 30° is usually enough to separate the ions e unchargeth d an d component e beamth f simplA o s. ° analyzin30 e g magne s showi t n in Fig.97 e deflectioTh . n coi s i dividel d o intparto allot tw o sa w mor e effective cooling of the coil. The same solution applies in the case of the previous deflection magnet: however s i mor t i e, difficul o t wint a dsemicircula r coil than a rectangular coil. The coil is about 3000 windings if the power supply of Fig.9 s i utilized.4 Both e othemagnetus y r coilma s s with thicker wired lowean s r resistance t bu the, y suffer froe mth disadvantag e that they would need larger buffer condenser bank d highean s r electronic current regulation. The power supplies of the magnets - if they are regulated - can be calibrat- r efo bead m energy, utilizin e monoatomith g e cneutro th ion f o s n generator target beam a properlline n I . y regulated magnet power supply e changth , n beai e m energy e (duchangth f eo acceleratint o e g voltag r o extractioe n e voltageb y ma ) detected by the change of the deflection current needed to achieve the maximum beam current.

11.4 VACUUM CHAMBERS OF DEFLECTING MAGNETS e vacuuTh m chamber f deflectioo s n magnets e vacuushoulth t fi dm system (the beam line and the target holder) of the neutron generator. If the the monoatomic

- -—1— Deformation

— — Deformation

Target 2

Fig.98 Vacuum chamber of deflection magnet using oval tubes

144 beam component is selected, then the molecular and three-atomic components usually thermally loae wal th f do lthi s chamber e correspondinTh . g chamber surface should be cooled properly. The deflection magnet vacuum chamber is usually made of stainless steel, but other suitable metals (e.g. copper) can also be used. The vacuum chambers of the analyzing magnets are usually connected to the beam line of the neutron generators by elastic joints, like bellows, which make alignment of the beam line easy. verA y simple vacuum chamber constructio s showi n n Fig.98i n e chambeTh . r was manufactured of two stainless steel tubes with their original circular shape pressed into an oval form and welded to each other, and fitted to the 30° deflection magnet. The undeformed ends of the circular tubes can be fitted easily to the beam line components of the neutron generator. Vacuum chambers for different magnete manufactureb n ca s a simila n i d r way.

11.5 PROBLEMS WITH ANALYZING MAGNETS For a correct analyzing magnet the shape of the beam is narrow. The beam profile observed by quartz or other visual detectors depends on the energy spread of the beam and the ripple of the magnet current. The high ripple (> 1 %) in the magnet curren y comma te froe malfunctioth m e currenth f o nt regulating transistors (see e.g. Fig.95.) e breakdowth ; n between collecto o t emitter r will shorten e regulatinth ge originath circuit d an e l ,fulth rippll f o wave e rectifier will caus e a spreaedeflection th n i de punchthrougTh . a typica s i h l problem with power transistors and it can be detected by testing the diode base to emitter and base to collector. Both junctions will indicate correct operation, while the emitter to collector shows total short circuit. The extraordinary warming up of the coils shows the short circuit within the coil. The cold and warm resistance of the magnet coil should be tested and checked and regularly recorded in the log e neutrobooth f ko n generator during maintenance work.

145 . QUADRUPOL12 E LENSES

Magnetic or electrostatic quadrupole lenses are commonly used as post- acceleration ion beam lenses at almost all accelerators. For neutron generators w energlo suc s ya h accelerators e magnetith , c quadrupol e mosth s ti e frequently used lens. Electrostatic quadrupole lense e morar s e simple t thebu ,y need power supplies and high voltage feedthrough into the vacuum system. Furthermore, a high vacuum is required in the system to avoid corona discharges. The relatively low energ n beamio y n neutroi s n generators need simpl voltagw lo e e magnet power supplies, and the lack of vacuum feedthrough makes the magnetic lenses much simpler to use. The only drawback of the magnetic quadrupole lenses is the relatively heavy weight of the coils and the iron cores of the lenses.

Fig.99 Electrostatic quadrupole lens focusing in the horizontal or vertical plane

Fig0 Magnetic.10 quadrupole lens focusing horizontale th n i verticalr o plane

146 Fig.101 Approximation of the hyperbolic electrode by circles (cylinders) (electrostatic quadrupoles)

Fig.102 Approximation of the magnetic pole faces by steps

Fig. 103 Focusing particle trajectories of a quadrupole doublet

147 e quadrupolTh e lens consist f o fous r hyperbolically shaped pole facer o s electrodes. The quadrupole lense focuses in one plane and defocuses in the perpendicular plane. Thus, several such lenses must normally be combined to make a useful lens system. Usually quadrupole doublets and triplets are in use at neutron generators. Sinchyperbolithe e c pole face or electrodes difficulare s to t fabricate e shapth , s approximatei e circley b d r stepso s . e focuth If se quadrupolth plan f o e vertica th e n r i o horizontalenl s i s l plane e beath , m direction e electrostati th z axis( e smagnetif th o ) d an cc quadrupole lenses will have the shape shown in Fig. 99 and Fig. 100 (45° rotation!), respectively. In practice the hyperbolic pole shapes are approximated by cylinders (electrostatic quadruplets) and by steps (magnetic quadrupoles). These approximations are shown in Fig. 101 and Fig. 102. As the quadrupole singlet focuses in one plane and defocuses in the perpendicular direction, minimum doublet e thereforar s e always usede particlTh . e trajectories started at the center of the object (x = 0 and y = 0) in a double focusing doublet, as shown in Fig. 103. The doublet is a double focusing system. The calculation of the ion trajectories along the quadrupole lenses can be found in Ref. [91]. The exact calculation and lens design is beyond the scope of this Manual. A magnetic quadrupole doublet is described in Section 12.1.

BIASEE 12.TH 1 D QUADRUPOLE LENS e biaseTh d quadrupole len s i poweres d asymmetrically o e thas particl,th t e beams are focused and steered. The repeated and, hopefully, always convergent alignmen e beath mf o t line elemene steerinth y b tg e quadrupoleeffectth f o s n ca s be eliminated. Instea f mechanicao d l adjustment s i quit t i e, practica o t mak l n a eelectri - cal alteration of the effective position of the quadrupole by unbalancing the voltage e segment th e curren th f o sr (o st throug e coils)th h . The circuits of the quadrupole biasing network are very simple [92]. The magnetic quadrupole lens has four equivalent coils and in the simplest case the current flows in series through the coils. When the symmetrical supply of the coils is altered, the different pole pieces will be magnetized differently. The balanced e quadrupolsupplth e f change b o yn externa a n y ca eb d l bypass branchf I . the serial connected coils and an external potentiometer are connected parallel e potentiometeth f e ano centrath d p ta ls i connecte r e e sericenteth th - f o o t rd ally connected coils, the balanced bias of the lens can be altered by sliding the potentiometer. This is the principle of the biased quadrupole lens. The circuit diagram of a biased magnetic quadrupole lens is shown in Fig. 104.

148 Fig4 Circuit10 . diagram biaseda f o magnetic quadrupole lens

Fig5 Circuit10 . diagram quadrupolea f o doublet

The biased quadrupole lens (both magnetic and electrostatic) requires in practice split power supplies. We now discuss a magnetic quadrupole doublet lens for neutron generators up to an accelerating voltage of 150-200 keV. The focusing distance of the quadrupole doublet is about 60-80 cm and the beam can be steered at this distance both horizontally and vertically in a range of a few centi- meters e poweTh . r e a coilsuppla currenth voltags i se f b o y tn e ca sourc t (i e sourc s wella d einstea a an cente) f o d r tapped power potentiometer n electronia , c equivalent (potentiomete d emittean r r follower s i use)d [93] e circuiTh . t dia- e grath quadrupol f mo e doublee wit on d htha f s an showi o split t n Fig i 5 tn 10 . power supplie s showi s n Fig.106i n .

149 + 24V

+ 34V

CURRENT SYMMETRY

Fig.106 Split power supply (1/4) of the quadrupole doublet shown in Fig. 105

e change e b currene bea th e th n focu y mTh f ca b o d ts potentiometers control- line outputh g t voltagee poweth f o rs supply whil e e biasinupper th th o e f - ro g lower - pole pairs is done by the symmetry potentiometer controlling the bypass of the upper or lower pairs of coils. The strong focusing property of quadrupole lenses makes them particularly e alonsuitabl us e bea th r g fo me linn neutroi e n generators e commerciaTh . l neutron generators, which are mainly manufactured for fast neutron irradiation and activation usuallanalysisnot yare , equipped with this strong focusing component. As the magnetic quadrupoles are simple, they are therefore recommended in addition to the beam deflecting or analyzing system for upgrading commercial neutron generators.

12.2 THE BIASED MAGNETIC QUADRUPOLE DOUBLET The quadrupole lens described above has been designed and constructed for energV ke 0 y 20 use - n smali d V ke l focusin0 neutroD bea15 e f th gmo n generatorse polTh .e fac ediametem m apertur 2 8 rf o e beafit th s m tube f almoso s l al t commercial neutron generators. The properties of this lens allow the D beams to e focusebeaV b ke m 0 windin 0 de coilenergy 85 15 th dow r e sf o go fo 40-5t nTh . m c 0 were designed for a 50 cm focal length. Changing the standard coils of the quadrupole lens pair will ensure a different focal length. Increasing the current

150 tx) O 2 3

150 f mm)

Fig.107 Mechanical diagram quadrupolea f o magnet r neutronfo generators flow through the coil will shorten the focal length. The f focal length of this A quadrupole lens can be calculated by:

2x 10 4[A. turns.cm] fjcrn = —————————————] — (38) [A-turnsN x I ] x

e focath whers l i e quadrupollengt th f e f o h e efocusin th len n i s g s i plane N , A e numbe th e curren th f a turn singlo s r i n f amperesi o tsI e e powed coiTh an l. r supply of the quadrupole doublet is based on a simple standard 24 V AC secondary transformer e e quadrupolsupplth Th .f o y s i asymmetricale , usin e biaseth g d quadrupole principle. A diagra f mhalo f thio f s quadrupol s i showe n i Fign . 107e materiaTh . f o l the iron core is manufactured from low carbon content iron (C-10 type). The pole piece e plasma-weldear s d onte planeth o d iro e nfou th sheet r d iroan s n sheete ar s screwed together by Alien type bolts. The pole pieces were shaped on a lathe and the connecting surfaces were manufactured on a milling machine. The wiring diagram of the quadrupole lens pair is shown in Fig. 105. Note e perpendicula th e thacoil th f o t s r segment e connectear s n i series de coilsTh . , a doubldi m em emaile8 o0. f d copper wires, each coil wit0 turns85 h , were wound. The surface of each coil is large enough for the dissipation of 10-15 VA. The coile connectear s a 12-blado t d e socket fixe o aluminiut d m sheet that fastens the pairs. The rectifier circuit serves for the supply of the voltage regulation of the e magnetdissipatioth e coils th s A i low .sf o ne chang,th f o resistancee th f o e coil can be ignored, which means that this magnet does not need a constant curren o ttransformer tw supply e Th . so ful tw suppll e wavo th ytw e e rectifierth d an s Zener diodes regulatin- in g voltagC e circuitD th V s i e,4 witvoltageC +3 A h e Th . put e voltagvoltagth f o ee regulator C e -5.D th 6V d san (7805V 4 ) +2 whil e th e source e needee operationaar sth r fo d l amplifiers (Fig. 108). The regulator of the quadrupole lenses is a voltage regulator based on the three-terminal 7805 circuit, which gives adjustable output voltage througe th h first 741 operational amplifier. The center tap of the series-connected coils is poweree th "electroni y b d c potentiometer" built fro n a memitte r followere Th . center tap of the two coils and the currents flowing through the coils can be changed by the symmetry potentiometer. The basic voltage regulator is shown in Fig. 106. A quadrupole doublet requires four regulator units. e quadrupolTh ee mounteb lense n ca s d ont e beath o m liny removinb ep to e th g e quadratith f lio d c shaped lense beaTh .e mneutroth linf o en generator should

152 800mA —— 0I 0 22x15k CO- IN9K

24V- 2AV~ 380C5000 380C800

lOOOOuF -HO- 63V

0 it ov + 34V + 2AV -5.6V

Fig. 108 Rectifier circuit of the biased quadrupole lens pair

be placed between the pole pieces and returned to the upper part of the lens. The pole pieces are vertical and horizontal, respectively. A quartz or Pyrex glass target 5-8 mm thick can be used to observe the focusing properties of the lenses. As the beam of a commercial neutron generator with about 100-200 W power can heat up the glass, it is recommended that a copper mesh scree e placeb n d e quartzjus th e n fronbeai tth s f sooo A mts . a n glown o s the quartz, increase the lens current by the "current" potentiometers of one of the lenses. If the lens current seems to be at optimum, a 45° line will appear on the glass target. The position of the line focused beam can be changed, on the biased quadrupole principle, by the "symmetry" potentiometers. These potentiometers contro e emitteth l r follower e shuntoptimaTh . l positioe th f o n e centerth bean i . s mi Whe e firsth n t s lenbeeha s n tested, turn e dowth n "current" potentiometers and activate the second quadrupole lens. A line perpendicular to the previous one will show the focusing of the second lens in the perpendicular plane. Whe e th optiman l positio r botfo n h lense s beeha s n found, the focus of both lenses can be activated. If the user does not like the two perpendicular 45° plane focusing feature of this quadrupole doublet e doubleth , t shoul e turneb d d 45°. This means thae th t stan - witd h horizontal tabl - e shoul e replaceb d a stan y b d holdin e outeth g r magne e horizontath t o planet ° 45 ls plane.

153 12.3 TROUBLESHOOTIN MAGNETIA F GO C QUADRUPOLE LENS

The effects of improper operation of a quadrupole lens can be as follows:

a) The beam is not focused or is deflected in spite of the activation of the curren d symmetran t y potentiometers. This show e th currentless s statf o e the coils. Test: e interconnectioth - e powee magneth th d f rno an tsupply , - the output voltage of the power supplies, e walth - l e mainsoutleth f o t, - the continuity of the coils. The repair of the faulty component depends on the fault found. b) The doublet lens does not focus in one plane. Check the corresponding circuits. e lenTh sc ) doet focu no sn bot i s h t i workplane d san s only horizontall r vero y - tically. Activatio f boto n h quadrupole lenses resulta bea n i ms which cannot be focused by the two lenses together. This can cause a faulty rectifier. Test: e buffeth - r condenser e rectifierth f o s , - the rectifier diodes, - the voltage regulator 1C. d) Bad focusing conditions can also be caused by poor vacuum conditions. Tesd improvan t e vacuu e systemth e th mn i . ) e Strange magnetic fieldy caus e ma sth focusine g propertie a quadrupol f o s e leno t deteriorates e th quadrupol f I . e lens (paire th s i beaclos ) o t me analyzing magnet e deflectioth , n magne y magnetizma t e th quadrupole e lens. This effect can be observed from the strange focusing behaviour of the quadrupole lens while the power supplies and coils work normally. In this case a magnetic shiel hely o improvt pma d e focusinth e g properties.

154 13. HIGH VOLTAGE POWER SUPPLIES

High voltages are extensively used in physics (accelerators, electron micro- scopy, etc.), for electromechanical (X-ray) equipment and industrial applications (precipitation and filtering of exhaust), and in communication (radio stations). e requirementTh r voltagfo s e level, current rating d shor an r so tlon g term stability for every DC HV power supply may differ widely. High voltage power supplies are used in neutron generators for extraction, focusing, acceleration, etc. e deuterooth f n ions. Rectificatio e alternatinth f o n g e currenmosth s ti t usual metho o t obtaid n DC high voltages. (A typical circuit is shown in Fig. 112). Most of the rectifier diodes nowadays adopt Si-type d althougan , e peath h k reverse voltag s i limitee d , rectifie kV o t lesn 5 ca s2. r d thahundred. V o t tenan stackk ca n p sf u o s s be made by series connection of more diodes. The use of the old selenium rectifiers has the advantage of serial connection without resistor chain, but the voltage drop and the power consumption are high during the conducting period. As the serial connectio a e higdiode th r h fo f o nvoltags ecauseV k 0 oves10 rmor e problems, the high voltage power supplies over 100 kV use mainly voltage multiplier circuits. The electrostatic high voltage generators, manufactured by SAMES (Societe Anonym e Machined e s Electrostatiques, Grenoble, Frances successorsit r o )D AI , (Assistance Industrielle Dauphinoise, Zirst-Meylan, France) and ENERTECH, France, are electrostatic machines converting the mechanical energy into DC high voltages. Such electrostatic generator e callear s d Felici generators- in , e afteth r ventor (as the insulating ribbon electrostatic HV machines are called Van de Graaff generators).

13.1 ELECTROSTATIC (FELICI) HIGH VOLTAGE GENERATOR e FelicTh i generato e rGraafd [94n f] Va replace generatore th e belf th o s t s by a rotating insulating cylinder, which can sustain a perfectly stable movement agains e rubbeth t r belt, which tend o vibratt s e eve t a hign h speeds e princiTh . - e Felicth plf io e generato s showi r n Figi n . 109e maiTh . n - componentFe e th f o s lici generator are:

The rotor, which is a tube-like cylinder, made of insulating material. The rotor is driven by an electric motor and charges are deposited on the surface of the rotore rotatinTh . g cylindee onlth ys i rmovin g parf o thit s electrostatic generator.

155 HV out

Fig. 109 Diagram cross section of the Felici generator

Discharging ionizer

Metallic inductor Stator _ Glass cylinder Insulating cylindrical rotor HV Output (-200KV) Charging ionizer

Compressed hydrogen

Load

Voltage regulator

Exciter (+30KV) out

Fig. 110 Principle diagram of a regulated two-pole Felici HV generator

156 The ionizer electrodes, which are very thin metallic needles (blades) placed in close proximity to the rotating cylinder. The charging needles spray the electric charges by corona discharge onto the surface of the rotor while the discharging electrodes (needles) collec e chargeth t y drawinb s e gth f theof m surface rotorth f o .e e segmentTh r inductorso s , which induc a strone g electri e cth shar fieln o pd e th ionizersedg f o e inductoTh . r electrode e ar places d behin a d slightly conductive (ca. 120 Q/cm) special glass cylinder. The excitation inductors lay the electric charges onto the surface of the rotor whilst the extracting ting inductor withdraws them. The charge collecting (ionizer) electrode and the inducto e oppositrth pain o r e e callesid ar e eth machinea pold f o ee Th . diagram cross section in Fig.109 shows a single-pole machine, and Fig.110 show a two-pols e machine e maximuTh . m numbe f poleo r s usuallyi s . 8 The inductor and the conductive glass cylinder are together called the stator. e e statorth ionizeTh e rotod th ,an r electrode e closear s d hermeticallly ina tan k under pressur f compresseo e d hydrogen.

The U excitation voltage in Fig.110 produces a potential difference be- CXC twee e excitatioth n n e inductorcharginth d an gs electrodes (inductors) sufficient to get an electric field on the edge of the charging ionizer high enough to cre- e locath le at ionizatio e e rotorsurface th th insulatin Th f n .o o en g rotor trans- e ferth electris c charge toward e collectinth s g electrodes which collece th t chargescapacitancC e Th . e betwee e th discharginn e gth inducto ionized an s i r several times less thae capacitancth n e betwee e th charginn g ionizer electrode . outpu U t inductor its voltag the and so e: wilbe lsevera l times higher than the U charging potential. t^A.i*' The U . potential of the collecting ionizer at any instant is about C abov Q/ e ground th e= e charg, th e capaciwherU s th i e s i Q ecollecte - C d an d V electrod H e ground V electrodth e H e th tancpotentiao t e Th f .o th ee f o (i.el . the voltage of the HV terminal of the neutron generator) rises at a rate given by the expression dU/dt = I/C where I is the net charging current to the terminal, I = sbv where s is the charge density on the the surface of the cylinder in Cou- 2 e tangentiath s i e e cylindeheighv th th lomb/c l f d s o i velocite tan rth b m f , o y cylinder in m and m/s respectively. As the loading behaviour of the Felici generator depends on the dU/dt charging capability of the construction, the larger size of the insulating cylinder ensure a larges r loading generatoV possibilitH e th s well a f r o y . The output high voltage will be controlled by the regulation of the

Ufixc charging voltage usin a gfeedbac j kvoltag R loo- - ^ pdi e R consistin n a f o g vider and an operational amplifier.

157 The maximum charge density on the rotor surface is about 0.01 - 0.02 ,aC/cm ; the tangential field on the surface of the rotor is maximum 15 kV/cm. The output current of the HV output varies between 100 /uA and 15 mA depending on the speed e rotationth of e surfac th e ,numbe th e rotoe d th f chargo an sizrf r o e e collecting inductor-ionizer pairs (poles). Figure 110 shows a two-pole machine. The maximum achievable output currene Felicth f io t generator depende th numbe n o sf o r poles; its maximum output voltage is approximately proportional to the distance betwee a e givepoles th r n nFo . diamete e rotor th e f th outpu,o r t voltag- in s i e versely proportiona e numbeth o t f lpole o r s [95]. The motor driving the HV generator rotor rotates at a speed below 3000 rpm. This usually corresponds to a velocity of 15-25 m/s of the surface of the rotor. generatoV H e efficienc Th e s betweeth i r f o y n 80-90% .a remarkabl This i s y high value compared with the efficiency of the usual voltage multipliers. e utilizatioTh f puro y nhydroge dr e n (10-2 r pressure5ba ) give n a sexcel - lent insulation, ion transfer onto the surface of the rotor, and good thermal conductance (i.e. cooling) e Felic th e tan f Th io .k generator doet neey no san d extra cooling facility n i som: e Felici generators utilize t a neutrod n generators the HV tank is sometimes cooled by the same cooling water as the vacuum pumps and the target. e rotor e Th ionizersth , e stator th e ,drivin th , e gprecisio th moto d - an re nr sistor chain (R e enclose)e ar th air-tigh n i d t tank, e calleth hermeticalld y

sealed unit. The unih t does not require any maintenance and it can be repaired only at the manufacturer's. In case of any difficulties related to the hermetically sealed unit, the whole unit should be replaced. The only duty related to the hermetically seale e dregula th uni s i tr (say monthly) checkine hydrogeth f o g n pressure in the tank. The pressure gauge can be found usually at the bottom of the pressure tank.

Uout \—- without 1 excitation control \ ' 1 with excitation 1 control 1 I \ \ !out

Fig.lll Load characteristics e Felicith f o generatortypeV H

158 The Felici generator has an ideal low short circuit current and ideal CV-CC (Constant Voltag - eConstan t Current) output characteristics e fluctuatioTh . n i n voltag s i lese s belo% thaI e criticath wO. n l loading current e loadinTh . g characteristic f thio sV generatos H typ f o e eshow ar r n i Fig.Illn , d witwithouan h t excitation charge control.

i —— V~< . P U

V QR == L(load h.K transformer

Fig. 112 The half wave rectifier

13.2 AC-DC CONVERSION HIGH VOLTAGE POWER SUPPLIES

13.2. e singl1Th e phase half wave rectifier The single phase half wave rectifier with capacitive smoothing shown in Fig. 112 is of basic interest. Neglecting the reactance of the HV transformer and the interna ldiodD impedance th durin f o e conductioneth g , capacitos i C r charged to the maximum voltage of the AC voltage V ~ (t) of the transformer if thdiodD e e conductsdiodD e Th .mus e dimensioneb t o t dwithstan a d pea- re k . versThi estransformeV V 2 H voltagwoul e f e o th cas th edf s i ei r als e b o IlldX grounde e terminath t e a outpudb insteae termina Th lth f . to da o voltagln V e longer remains constant if the circuit is loaded. During one period, T = 1/f, of the AC voltage a charge Q is transferred to the load R, which is represented as

= iQ ¥(t) V (t) dt = IT = I/f (38)

s therefori I e meaoutpuC th e D (t)C d y n i D V(te an , tvalue th th f o )es i voltage which includes a ripple, as shown in Fig. 113. If we introduce the ripple

factor aV we may easily see that V(t) now varies between VTV1Qv and V • and v 2 aV mm = - - The charge Q is also supplied from the transformer within a

159 "max

Fig. 113 The output voltage and charging current of the half wave rectifier with buffer condenser loadand Rj C

short conduction time t = «T of the diode D during each cycle. Therefore, Q also equals

(39)

As « « 1, the transformer current i(t) is pulsed as shown for an idealized s e givei rippl Th y forV b n . a e n Figm i 3 .11 IT 1 T I = 2ffVC = Q CTV = (40) 2C 2fC This relation shows the interaction between the ripple, the load current and the circuit parameter design values f and C. As, according to the ripple, the mean output voltage will also be influenced by oV, even with a constant AC voltage V ~ (t) and a loss-free rectifier D, no load-independent output voltage can be reached. The product of fC, is therefore an important design factor. For neutron generator circuits (oscillator, focus, etc., power supplies), a sudden voltage breakdown at the load (Ry —> 0) must always be taken into account e disadvantagTh . e singlth f eo e phase half wave rectifier concern e posth s - sible saturation of the HV transformer if the amplitude of the direct current is comparabl e th nomina o t e l alternating e currentransformerth f o e tbiphas Th . e half wave rectifier shown in Fig. 114 overcomes this disadvantage, and does not chang e fundamentath e l efficiency V windinge ,H transformeth sinco f o tw es e ar r

160 v,~m w\

-= V ——————————— (

h.t. transformer

°2 ...... „„,... Jfcl, ...... V2~m

Fig. 114 Full wave single phase rectifier

now available. With referenc e frequencth o t e f ydurin e cyclew eacon gf no o h, s conductini e ~ halD on thf r ed cyclfo gdiodean e, D wits a htim e dela f T/2o y . e ripplTh e facto s therefori r e halved. Thus single phase full wave circuits can only be used for HV applications C D e transformee th th coiV e e f midpoinearthe H b o th f li t e n a d th ca ifdran t outpu s single-endei t d grounded.

13.2.2 Cascade generators The first voltage multiplying circuit was published by Greinacher [97] in s improvewa d 192 n an i 193d0 y Cockcrofb 2 d e Waltofirsan th t t n i naccelerator . Such cascade circuits are known as Cockcroft-Walton [98] generators. An n-stage cascade circui f o Cockcroft-Waltot n typ s i showe n i Fign5 togethe.11 r wite th h main working parameters e see.b Fronn ca thatm t i Fig :6 .11

- the potential at all nodes l',2',...n' are oscillating due to the voltage oscillation of V(t); - the potential at the nodes l,2,...n remain constant with reference to the ground potential; - the voltages across all capacitors are of DC type, the magnitude of which is 2V across each capacitor stage, except the capacitor C , which is stressed IlldX il with Vmmun x only; n s i stresse , r , ;...o n D~ D D d ; ; ~ , -V wit D 2 L ever ; h yv , D L rectifie ; 1 D m£Lr x i i twice AC peak voltage; and - the HV output will reach a maximum voltage of 2nV IllctX

161 vm;vmnx J_

Ideal output voltage: V = 2nVmax

Output voltage ripple: V =

f = frequency = loadinI g current

Voltage drop: 12/(n3

I2n V 2n = Loade V outputV H d : o max

5 CircuitFig11 . diagram f Cockcroft-Waltono generatorV H

162 1 = H.V. output

w 2

0,....Dn conductin2 I' g t, o;....on conducting

Fig. 116 Waveforms of the potentials at the nodes of the Cockcroft-Walton cascade circuit

Loaded HV output (I>0): If the generator supplies any load current I, the output voltage will never s showa n Figi reacn e valu.V th 116h 2n e. There wile la th rippl als e n b o e IIluA. voltage, and therefore we have to deal with two quantities: the voltage drop V and the peak-to-peak ripple of 2aV. The sketch in Fig. 116 shows the shape of the output voltage and the definitions of V and 2aV. The peak-to-peak ripple is giv- y b n e rT(l/C.= V 2a ) (41)

The total ripple will be [96]

^ (1/C = 2/Cn/Cj+ V + . 3/Cn+ 2) .. 3+ (42)

For a given load, V may rise initially with the number of stages n, but reaches an optimum value and even decays if n is too large. For constant values

163 of I, V f and C the "optimum" number of stages is obtained by

dV differentiating the equation for VQm with respect to n, i.e. from —-^ [96]:

omax n ~max " nopt (43) d an a Cockcroft-Waltoz r H Fo 0 50 = f , n kV generato0 10 = r _ witV h max I = 500 mA n = 10. It is, however, not desirable to use the optimum number of stages, as then V is reduced to 2/3 of its maximum value (2nVmaxm ') and the voltage variations for different loads will increase too much. voltageC D e sTh produced wit a hcascad e circui rangy ma t e frop u m V somk 0 1 e o mort , wite thahMV curren2 n e t Th rating . mA o somst 0 frop 10 eu m A som// 0 1 e supply frequencies of 50/60 Hz heavily limit the efficiency; therefore, higher frequencies up to some 10 kHz are dominating. A 50 Hz transformer circuit needs a much larger capacitor than 10 kHz, so the breakdown along the smoothing column much more likely to damage the acceleration tube or other ion optical components valuw lo tha ee capacitonth r staca mediu f o k m frequency cascade circuit.

13.2.3 Improved cascade circuits A number of improved HV power supplies have been developed: e.g. the insulated core transformer (ICT) e Allibonth , e voltage multiplier e cascadth , e circuit with cascaded transformers, the "Deltatron" and the parallel, capacitively powered "Dynamitron".

Y////////////,

o*nV0

Fig.117 Insulated core transformer (ICT) high voltage power supply

164 ) Insulate(a d core transformer typpoweV H e r supply e insulateTh d core transformer based power supply (Fig .a serie117 s i ) s connection of single (or full wave) rectifiers on different - well insulated from the transformer cord froan em each othe - secondarr y a transformercoil f o s . Each rectifier circuit will produce a V output voltage, so the n stage of rectifiers will produc V n outpue t voltage e advantage e Th loadabili .th e circuith e f ar o ts - d e thahighestyan th ,t s secondarit t stacd an k y coil shoul e b insulated d only fro e transformemth r cord primaran e y coi V loutpun ove e th rt voltage.

;;c, w jio,

v i

in Fig.118 Voltage multiplier circuit according to Allibone

a f J - Further stages

(up to n ) .

i k 3 Stage 2 2 i i 1 '' = n 2V

2 i 3 Stage 1 2 i, t i = V 1 HLi1 J~

Fig. 119 HV cascade circuit with cascaded transformers

165 (b) Allibone voltage multiplier e AllibonTh e voltage multiplie s i rbase n higheo d r voltage transformers. Each stage comprises one HV transformer which feeds two half wave rectifiers. As the storage capacitors of these half wave rectifiers are series connected, the HV secondary coi\ T cannol e groundedb t . This mean e mainth s insulation between the primary and the secondary coils of T-, has to be insulated for a DC voltage of

V , the peak value of the secondary coil of Tv The same is necessary for T9, IllcL/C A £* but here the HV secondary coil is at a potential of 3V Increasing the number ill.cU\ of stages will increas e insulatioth e n problem e secondarrelateth o t do t priy - mar e insulatey th e cas insulationth f o en i d s cora , e transformer e AllibonTh . e HV multiplier circuit is shown in Fig. 118.

cascadC (cD ) e with cascaded transformers The increasingly better insulation between the transformer core and the sec- generatorV H e moderate ondarb T n e cascadeIC th ca sye y th coilb d n di s transformer method. In this circuit, every transformer per stage consists of a low volt- e primaryag a hig, h voltag evoltagw lo secondar a e d tertiaran y y e coilTh thir. d low voltage coil excites the primary coil of the next stage. The operation of the understoobe circui may t d fro circuimthe t diagram show Figin n . 119. The advantage of this circuit is that each stage is identical: there are no higher insulation problems betwee e primarth n d secondaran y y coil- s Al tha . V n though there are limitations on the number of stages, as the lower transformers have to supply the energy for the upper ones, this circuit, excited with mains frequency, provides an economical DC power supply with moderate ripple factors and high output power capability.

(d) The Deltatron HV generator A sophisticated cascade transformer systee poweC Deltatroth D s mi r V H n supply. These generators might be limited in power output up to about 1 MV and e verTh y . somsmalmA e l ripple factor e higth , h stability e fasth ,t regulation e anth dsmal l stored energie e ar sessentia l capabilitie f o this s circuite Th . circuit shown in Fig. 120 consists of a series connection of transformers, which t havd no on iro a e n core. Connecte o t everd ye usuath stag s i le Cockcroft-Walton cascade circuit, which has onl a ysmal l input voltag t eproduce bu (som ) kV es output voltage f somo s 0 1 e r stage ke pe Vstorag Th . e column f theso s e cascade e thear s n connecte n direci d t series, providin e higC outputh gD h t e wholvoltagth r efo e cascad V generatoH e r unit. Typically, up to about 25 stages can be used, every stage being modular- constructed as indicated in Fig. 120. These modules are quite small; they can be

166 Terminatio' : n HV d.c. -output-

Further stagesi

Module

Stage 2

Voltage multipliers

Stage 1

Fig. 120 The Deltatron high voltage generator

out

Fig. 121 Circuit diagram of the Dynamitron HV cascade

stacke a cylindrica n i d l unit, sometimes unde^ pressureSF r e outpuTh . t voltage is usually controlled by regulation of the primary side. The stages of the multipliers usually have an HV voltage divider for feedback. As the supply frequency ranges from several kHz up to 100 kHz, the voltage divider for regulation of the output is a combined R-C voltage divider. The modular power supplies of the WALLIS HIVOLT (United Kingdom), GLASSMANN Highvolt (USA) and the Technical Uni- versity Buadapest (Hungary) hav a similae r structure. Thes V generatorH e e ar s used as acceleration, extraction and focus power supplies in neutron generators, e supplth r f electrostati o yfo s wel a s a l c quadrupole lenses.

167 (e) The Dynamitron HV power supply The Dynamitron power supply is a parallel fed cascade circuit using high frequency of a few MHz. The MHz range has the advantage of using the parallel charging capacitor n Figi .2 121(€ s) constructe s simpla d e spray capacitie- be s tween the nodes of n and n' of the original Cockcroft-Walton voltage multiplier. e loadabilitTh e Dynamitroth f o y s severai n V witA loutpu m a hten f o st voltage of MV range. The voltage multiplier system of the Dynamitron is usually placed

in the SFft pressure vessel of the accelerator, and the stages sometimes power the homogeneous field acceleration tube directly A Dynamitro. n type high voltage power supply utilized in a neutron generator is in operation for 14 MeV neutron therapy at the Eppendorf Hospital in Hamburg. A compariso e maith f no n parameter e singlth r efo s wave, full wave Cockcroft-Walton cascadee Dynamitroth d an s s showi n Tabln i n 9 [99]1 e .

TABLE 19. Main parameter f Cockcroft-Waltoo s n cascade d Dynamitroan s n

Circuit Single wave Full wave Dynamitron

NU - ^ T T V U -= 2NUQ U == 2mU u ° o o 1+4C1/C2 (unloaded)

I0(N+1)M I N Ripple IuT - ° u ZTC 2TU ~

l l 1 ( N -i i J Voltage drop ^ a\J(2^/3 = Q^ (N = +/Q 6 N/3 +U N/3)U) ^ = Q 1 +

Usual VQut 0.1 - 1 MV 0.1 - 2.5 MVV M 7 - 7 0.

A 1 ~ Imax A m 0 50 ~ ~ 50 - 100 mA

13.3 TROUBLESHOOTIN HIGF GO H VOLTAGE POWER SUPPLIES e malfunctioTh a hig f ho n voltage power suppl s usualli y y cause y impropeb d r e operatioth components f o n , e.g. transformer, rectifier(s), buffer condensers, filter (resistive or inductive), regulating system (electronics), contacts or in- sulations.

168 In troubleshootin a neutrof o g n generato n sourceio r F witn H a analo,h g (without active semiconductors) multimete s recommendedi r e outpuTh . t n powea f o r whic W n easil0 ca h 10 y F w oscillatokiln H a fe LSI-basel a s i r dT FE digita r o l transistor, amplifier based electronic multimeter. Similarly, every possible high energy discharge (e.g. buffer condense a main f o s r frequency Cockcroft-Walton circuit) can induce enough power in the circuit of an electronic multimeter to e kilactivth l e components. In troubleshooting and repair of high voltage units, maximum precautions shoul e takeb d n durin e work poweV th e gvoltagH Th .e d energr th an esupplie f o y s a oneutrof n generato e higar r h enoug e o t kilservichth l e personnel; therefore, troubleshooting and repair should be carried out with great care and never alone. r wor Fo n o khig h voltage power supplies, insulated-handled tools, test pind othean s r proper instruments shoul e usedb e drecommende Th . meteV H r d fo r measuremen e th outpu f o t t voltage usua th V servic T s i le e voltmeter. a Thi s i s cheap, commercially available meter, measurin , wit kV a proh g0 3 -voltage o t p u s perly insulated voltaghandleAC eFor . measurements similarlthe , y constructeHV d voltage dividers ( HV test probe ) are recommended for use with analog multimeters. For HV measurements, the grounding of the ground contact of the HV meter or probV H e shoul e testeb d d carefully before switch-o e poweth f o nr supply. seconA d important instrumen poweV H r fo supplt y testinn a ohmmetes i g r working with a few hundred volts for the measurement of high resistances over r testinfo d gan rectifyin, 1MQ 0 g diodeV condenser H s wel a s a l r insulationso s . As every neutron generator laborator s ha ynuclea r electronic instruments, recti- d fieran higs h value e testeb resistor y b dusinn ca sa g nuclea r detector high voltage power supply. A component can only be tested under working conditions: the totacircuitn ow la componens tes .f it o te carrie n b i t n ou ca dt e housine higTh th hf o gvoltag e power supplies should alway e b sopene a n i d switched-of d an output-groundef d condition. Eve a nswitched-of d outputan f - shortened power supply may contain charged high voltage capacitors: precautions should be taken in dismantling the covers of the HV power supplies; the body of the power supply shoul e groundee b repairind th d an dg personnel should wear shoes with insulated toes. After removal of the covers, try to discharge every high voltage condenser. The troubleshooting of a high voltage power supply usually starts with visual or small inspections. If there is a lack of output voltage, the output wiring and/or connectors should be inspected. The measurement of the continuity between e e poweoutputh th f o re tloa th suppld d an yshoul e donb d e without output voltage and the unit should be grounded (if it floats at the terminal voltage of the neutron generator). The conductivity can be measured with the usual multimeter

169 (digital or analog). Discharging of the output buffer condenser is advisable. If a ther lacf outpuo s i ke t voltage (an e dbuilt-ith V voltmeteH n r doet indino s - cate any output voltage), discharge, shorten the output and inspect the fuse (or e circuitth fuses n i .) Blown fuses often indicate some malfunctio e cirth n -i n cuit. Do not use higher rating fuses than the original value; this could cause some problems, and as the condensers in the HV power supplies usually store high a highe f o re energyratinus e gth , n fuseveca en caus a efire ! Recommended tests are described below.

(a) Testing the main components Testing a high voltage power supply should start from the power (mains) side. Chec e th primarkV trans transformerV H H e e -y th th e sid tesf f o o Th et. former starts with measurements on the unloaded transformer. Remove the load from the secondary side of the transformer. Test the primary current versus primary voltagee meantimth n I . e secondarth e y voltage V shoulH e C b dmeasureA n a y b d meter. The primary current should be in the interval indicated by the parameters, and the ratio of the secondary to primary coil voltage should be almost constant f I ther. s i some e short circuit betwee e windingth n ( primars r o secondy - ary), the temperature of the transformer will rise. If the HV transformer is e operatinfounb o t d g normally, tese rectifiere buffeth th t d r an scondensers .

(b) Testin rectifierV gH s The data sheet usually contains all of the parameters which should be tested and measured to find the trouble with the rectifier. The HV rectifiers are silicon diode d usuallan s n seriei y s connected controlled avalanche diodes A . rectifier diod n junctio p ey withstan wite ma on hn a d 1000-150 0V reverse avalanche breakdown voltage, whil e controlleth e d avalanche rectifier stacks withstand vol- e rangth tage f 100-15o n ei s . Dependin kV 0e diodese speeth th f o n do , g then ca y operate at up to 20-30 kHz frequencies if the charge storage in the pn junction is low enough. Testing these rectifiers (diodes) in the forward direction is sim- e plereversth ; e voltage bias require a reverss e bias voltag ee operath equa o t -l tional voltage of the HV power supply itself or higher. The forward characteristicf o sthes e emeasureb diode n a ca slightly sb d y higher voltage (300-50) 0V power supply, and the reverse current of the rectifiers can only be checked by a nuclear detector power supply.

(c) Testing forward characteristics The forward characteristics and the reverse bias current of the HV rectifiers can be tested by a circuit shown in Fig. 122.

170 SxIMQ 2W DIODE STACK + . r^~\ st

DETECTOR BIAS P. S. £/fkV METER 0-SkV

*-^ FFORWART D REVERSE nA-mA METER

Fig.122 rectifierCircuitV H r fo testing

r forwarFo d characteristic a 100-15 f o sV cresk 0 t working reverse voltage rectifier stack, the forward voltage drop of the rectifier stack at the nominal ) forwarmA (sa0 5 dy curren s i about t 100-18 , indicatin0V n equaa g l voltage drop at all of the rectifier tablets. e rectifieTh re testeb stack n n i theidca sr original circuiy wa n i suc ta h that the rectifier diode should be loaded only resistively and the shape of the rectified current shoul e observeb n da oscilloscope y b d . This n tesonle ca b ty carried out in simple (variac controlled, mains powered transformer) HV power supplies e rectifiee th shapTh f . o e de reversth voltag d ean e conducting (not only e chargth dun o junctions t p ee observe b e screee storag th n th f ca n o f n)o o d e the oscilloscope s thiA . s method requires some experienc n electrii e d an elecc - tronic measurements it is recommended only for personnel well trained and experienced in electronics.

(d) Testing the buffer condensers The capacitance of the condensers should be in the + 30 % and - 20 % range of the nominal one. As the low capacitance of a buffer condenser may cause a higher ripple in the output voltage, the correct value of the buffer condenser is important. In high voltage condensers the stored energy is usually high, so a breakdown insid e condenseth e y causma er blowin d evaporatioan g e outeth f ro n e condensercontacth f o t . Therefor e capacitancth e n ea importan tess i t t duty. e storeTh d e energcondenseth n i yn als e a ca testeb discharginro y b d g resistor: the resisto- re r e shoul th o arm e f fixetw b do s de th ont n insulatina d o an d ro g sistor shoul o terminald tw e th higtouc e f ho th s h voltage condenser f I ther. e e "healthyar " spark e condenseth s n storca r e higher energies. This discharging method always gives an indication on the condition of the condenser, especially when the person troubleshooting is experienced in discharging condensers in normally operating power supplies. The discharging resistor should be long enough

171 (correspondin e voltagth o V poweet g H condition re supplyth f d o havs an ) e high enough wattage to prevent the power from dissipating during the condenser discharge.

(e) Testing the resistors Testin e resistorth g , y othelikan e r test, should start with visual inspec- e tionbrownish-blackisTh . h e resistorcolouth f o r s indicate n overloaa s r buro d n off of the resistors. In the case of a high voltage drop on the serial resistor, the interruption in the resistors leads to sparks; these can be observed on the e surfacepainteth f o sd resistors. High voltage resistors covered with epoxy resin layers also show some coloured e regioncas th f o malfunction en i s . Lower value e resistancee b e testeth rangusua th n f y ca o b n del1-2 (i ) multis 0MQ - meters, but the higher (10-100 MQ) resistors requires megaohmmeters or a few hundred volts power supply and a suitable micro-nanoammeter. The usual 3-digit hand held multimeter e use b r thi fo dn s ca spurpose e ammeteTh . r should alwaye b s connected betwee e grounth n d termina e poweth f o e routpul th suppl d t an y voltage shoul e connecteb e dresistor th e voltagTh o t .d e acros e e th th resistos d an r measured current determine e resistancth s e resistorth f o e . The resistance of the insulators, connectors and other insulating components can be measured by the same method, utilizing a nuclear detector HV power supply and suitable nanoammeters.

172 . BEA14 M LINE COMPONENTS

14.1 BEAM STOPS fasA t switch-o d switch-of an e nneutro th f o fn productio a neutro t a n n gen- n sourcey e pulsinio carrieb b erato t y e n b th deflectioou ,gca d re accelth f o -n erated beam from the target or by a mechanical shutter closing the path of the accelerated beam to the target. This mechanical shutter is called a beam stop. The beam stop has a metal sheet - usually water cooled - which shuts off the beam in front of the target. This shutter can be operated electromagnetically or pneu- matically. The response time of the beam stops is relatively fast: they close and ope e beath n m within e extractioalmos th n source second io d on te an e nTh . voltage pulsation or the electrostatic beam deflection are faster. The principles of the two main type f beao s m 4 [50] stoe indicate12 .ar p d an n Fig3 i d 12 s

14.2 BEAM SCANNERS In general, the scanners consist of two wires, one rotating around the hor- irizontal axis, scanning the beam vertically, while the other moves perpendicu- larly to the first and gives the beam a horizontal profile. The planes scanned o wire e tw perpendiculaar s e be th deuteroyth o t rn beao mtw direction e th f I . wires move (rotate) synchronously, the shape of the accelerated beam can be determined. The principle of the single-axis rotating beam scanner is shown in Fig. 125 and of the two-axis beam scanner in Fig. 126.

14.2.1 Determinatio e beath mf no profile Let us suppose a homogeneous beam of radius r along the x axis in the Carte- sian coordinate system e rotatinTh . g wire rotates axialony e s gth wit h radiu. R s e thicknes e Th th scannin f o s g wir s muci e h lesse th thaf e radiuo th n r s beam. Similarly the radius r of the beam is much less than the radius R of the rotating wire. If the wire rotates with an angular velocity CD, the shape of the i(t) ion current in time will be described by

(cos2o>t-l)+1 (44)

which is based on the usual geometric formula. As the wire intercepts the beam twice during a single cycle, as in Fig. 127, the oscilloscope connected to the wire shows the shapes demonstrated in the positions I - V described below.

173 Beam line

Eleclromagne

Spring Soft iron core

Stainless steel housing

Fig. 123 Schematic diagram of an electromagneticalfy activated beam stop without water cooling

Beam line

! i 0*

Water inlet Water outlet

Wafer cooled beam shutter Feedthrough

4 FigSchematic12 . diagram watera f o cooled beam stop utilizing a vacuum feedthrough

Rotating wire

Fig5 Principle12 . rotatinga f o beam scanner

174 Fig. 126 The principle of a two-axis beam scanner

BEAM

Fig7 Arrangement.12 beama f o scanner

If the beam is centered along the time intervals b and c between the up-to- down and the down-to-up, the currents will be equal. In case of a circular beam: b = c eccentric beam: b = c

the diameter of the beam can be calculated from the following expression:

n si R 2 = d (45) e - replaceb —p n r ca y narro - b dFo —• wn . si (45beamseq e n )th i , , b whe« a n JTJ1 and so

d = (46)

175 e singlTh e rotating wire scane whole th beath st emge Z axisalono e T .th g e picturth bea f o me profile a secon, d wire, rotatin a perpendiculan i g r plane around the Z axis is needed. This means that two synchronously rotating wires can scan the beam shape and position [87]. Thes o wiretw e s scae beath n m curren Y d alon an e t th ) gZ axi alonI ( e s th g £j ^ y axis (Iy ). The horizontally (I ) and the vertically (I ) scanned beam currents can be displayed together on the screen of a double trace oscilloscope. If the rotating mechanism ensures a n/2 (90°) phase difference between the two wires, the exact positioe determine b e bea th n mf ca o n d [54]. e followinTh g oscillographi n indicatca e behaviouI th e r co picture I r f o s the beam:

I. Unstable beam current (improper operation of ion source, ion optics)

II. Stable beam current

III. In the case of perpendicular, synchronous two-wire beam scanning:

Horizontal scanning

Vertical scanning

For an absolutely concentric beam in the target tube we have bz (47)

e welTh l focused beam e oscilloscopshowth n o s a esmal l dut1 y« cycle v : o

176 rv.

Horizonta r verticao l l scanning

a 1 a whilwid r efo e beam

It shoul e notedb d , however e replaceb o wire a singln tw , y ca sb e thad e th t wire with a special curvature [100] for scanning the beam simultaneously in the horizonta d verticaan l l planes.

14.2.2 Problems with rotating beam scanners

1) Vacuum problems e rotatinth s A :g beam scanners require rotating feedthroughs: the usual seal testing shoul e b carried d out. e Similarlyth e statoth d an f i r , a synchronouroto f o r e atmosphereth s n i n motoi vacuum e r ar o r, , respectively, the problems observe e sometimear d e electrith t a sc feedthroug e scanninth f o h g wires.

2) Mechanical problems s thiA : s scanner require a smoothls y rotating wire, every mechanical problem (stain, dirt in the bearings) may make the operation of the scanner unnstable. These mechanical instabilities can be observed on the screen of the oscilloscope:

VI.

177 e perioth e pulsf tim o dd th e lengt ean e Th f o wilh l appeare th jerk n o y screen o f pulsI ther n .e y wir havs th i eema e e e stoppedth wir ef I . intercepts the beam, some DC current can be observed by the oscilloscope. Dismantling and cleanin e mechanicath g e scannerl th par f o t s shoul e donb d e carefully.

) 3 Electrical problems e e isolatioTh wirth : e ef conductivit th o noutpu d an t y between the scanning wire and the off-vacuum connector should be tested periodi- callye ceramiTh . c insulatio e wireth s i usuallf so n y covere a sputtere y b d d met- l layera : they shoul e cleaneb d d both mechanicall d chemicallyyan . The usual synchronous motors rotating the wire scanners should stopped off- bea o mavoit de beam th sputterin o t . Thie s i assuredu n gselectria y b dc contact e synchronouth e axif th o s fitte o t ds motor (end position switch) f I thi. s contact does not work properly the synchronization of the y and z scanning wires wile unstableb l e propeTh . - r os operatioe checkeb e e contac th th n y f ca b o dn t cilloscope observing the signals of the beam scanner while the synchronous motors f severaof d e switchear an l n times o e d phasTh . eY positiod an f Z o n bot e th h signals n shouli contact e b e doscilloscop Th . e shoul t C leveindicatD no d ly an e in the case of a stopped beam scanner.

14.3 WIRE ELECTRODE (MATRIX) BEAM SCANNERS e developmenTh f o microelectronict e sprea th f personao dd an s l computers have made beam scanning easy even at the lower end of accelerators. The typical multiwire sensors of a beam monitor - used earlier in the expensive machines - consisa vertica f o d horizontat an l l grif thio d n wires. These wires collece th t ion beam and the current is converted into voltages. These voltages are multi- plexed intn analoa o o digitat g l converter e multiplexinTh e .th processin d an g g e oth measuref d current daty e computecarrieb ar a t ou d r [101] n thiI . s Manual, only a general description of such systems will be given. e wirTh e chambe a bea f o mr scanning electrode o consistnormatw f o ls glass laminated epoxy printed circuit boards which hold the vertical and the horizontal wires e wire e Th mad.ar s e mainl f tungsteno y , electroplated with golde usuaTh . l diametee golth df o rplate d tungsten wires, e b whicsoldere n ca h d easilys i , abou 0 ftm.2 t Dependin e e analonumbee th inputh th f n o f go tgo r multiplexere th , wires cover an active area of about 50 mm x 50 mm square. The vertical and horizontal wires are mounted through the square using conventional soft soldering technique directly onto the printed circuit board. The wire holding the printed circuit board will be connected to the screening diaphragms and vacuum system flange holders by epoxy resin. Without epoxy resin bonding, the printed circuit e vacuub board n ca ms seale y b dTeflo n (PTFE) gaskets e printeTh . d circuite ar s

178 Al Endcap

Wireplanes Survey Markers

Spacer

Spacers

Endcap

8 AssemblyFig12 . drawing multiwirea f o beam scanner [102]

conventionally etched and connected by the usual ribbon cables to the individual e multiplexeth amplifier e o datt th ad f o an rprocessin s g e unitassemblTh . y draw- a multiwir inf o g e beam scanne s showi r n Fig i n8 [102] .12 .

14.4 THE FARADAY CUP Most experiments with neutron generators and charged particle accelerators require knowledg e beath mf o ee positiointensite th target th e d t sima Th nan y. - plest method for detection of a charged particle beam is to intercept the beam witn a insulateh d metal plat f o sufficiene t n thicknesscurrenio e tTh . striking this plate is measured by an analog meter if the mean current is greater than , 10e A whic"th casr s fo i neutroeh n generators. Complications arise because the ion beam to be detected is accompanied by a diffuse secondary electron beam. The utilization of beam line components like deflection magnets and quadrupoles will remove the electron component of the beam, but this can be rebuilt near the target. e uppeTh e renerg th limi f o yt e spectrsecondarth f o a y electrons inducey b d the accelerate de targe th ion n i ts doet exceeno s ddependinV e aboue 0 th 10 tn o g target material. However, the X-rays produced in the target when it is bombarded by protons or deuterons can contribute significantly to the high energy tail of the electron spectrum e problemTh . s e relateth secondar o t d y electrone b n ca s overcom y obtaininb e g reliable measurement e beath mf o se currensuppresth y b t - sion of the high flux of the fast primary electrons, as well as the emitted slow secondary electrons.

179 Amplifiers

L r— l!_ Horizontal wires ——— H r Computer MultiplexeC dAO Vertical wires

Fig. 129 Electronic block diagram of a multiwire beam scanner [101]

-HV GRID

- OUTPUT

GRID — VACUUM JACKET INSULATOR HIGH ATOMIC NUMBER METAL

Fig. 130 Schematic representation of a Faraday cup for measurement of the beam current

Starting froe facth mt thae mainth t electron componen s i slod tthae wan th t number of fast electrons reaching the current collectors (target), either from the residual gas or from the target, is small, it is possible to suppress the electrons by surrounding the target with a negatively biased mesh screenmaintaned at a voltage of 200 V. In addition, the collector itself can be biased positively and thus re-collect all the slow secondary electrons. The schematic representation of an electrically suppressed beam current collector (Faraday cup) is shown in Fig. 130. The grid placed in front of the beam collector cup is usually also replaced by diaphragms. The external cylindrical mesh screens secondary electrons emitte e outedth e froe Faradar th th m sidf o ey o X-rat e cuydu p n beaproductio io a m e neutro n th I [87] y .b nn generatore th , e targeth Farada s i t p holdercu y .

180 RELAY.

REFERENCE CAPACITOR

INPUT

Fig. 131 The principle of the target current integrator

10Hz-10kHz fOUT

^Nmax 0.1mA 1mA 10mA RS 100k 10k 1k

Fig.132 Circuit diagram of a V/f converter-based current integrator

181 14.4.1 Beam current integration In all of the experiments carried out with charged particle beams, the measuremene totath l f beao t m e chargeth numbe r o f ,o rinciden t particles that reached the target during a given time, is required. This requires an integration of the beam current detected in the given time interval. For the integration of the beam current the Faraday cup (target) is fed to a large high-quality capacitor. The voltage across the capacitor increases with the collected charge and reaches some predetermined level At .thi s voltag a e fast acting flip-flop will be triggered. This trigger operates an electromechanical relay for counting the charge, or in a more advanced form it is fed an accurately known quantity of charge of inverse polarity to the input capacitor to discharge it. In the circuit shown in Fig. 131, the input capacitor is charged by the input current (bea ms i discharge e t currenti referencth d y b an d) e voltage sourcen I . principle all of the voltage-to-frequency converters work similarly. The voltage-to-frequency converter chips can be used in simple target current integrators in the range of 10 nA to a couple of mA. A simple target current integrator utilize t a neutrod n generators with beam curren f severao t A l /* ten f o s is shown in Fig. 132. This integrator is a useful device for the observation of the burn-off of the tritium targets [103].

14.5 TARGET ASSEMBLIES The target assemblies of neutron generators are constructed to fulfil the following requirements: target holding; target cooling; target current measurement; to suppress the secondary electrons; to achieve the shortest distance between the target spot to the sample to be irradiated. The usual target holders are cooled with water or air. Water cooled target holders are used at beam current higher than 500 [iA when more than 100 W target acceleratioV k loa0 s preseni d20 e cas th f o en i t voltage. Water cooled target holder e usear s d mainl r fo activatioy n analysis, where distortioneutroV e originaMe th 4 f n1 o n spectrul m doet influencno s o mucto eh e accurace th elementath f o y l analysis using reference sampl f similao e r composi- tion.

Bombarding . . Evaporated occluding metal (Ti,Er,Sc....)

s*SSSfsssssi Backing , (CuAl,...Mo , ) Fig. 133 The construction of a target

182 The tritium or deuterium target used at the neutron generators consists of a good heat conductin , etc.Al g r backin)o covere o M g , d(mad Cu wit f ho e thin vacuum evaporated layer of hydrogen-occluding metal (usually Ti, Er, Sc, etc.) (see Fig. 133). The heavy hydrogen isotopes (tritium or deuterium) form a quasichemical compound with the occluding metal. In principle the tritium concentration can achiev e stoichiometrith e c ratio TLT. . ThiQ . s ratio e dependtemperatureth n o s , becaus e disintegratioth e T i "moleculesT f o n s i significan" t over 400°Ce Th . x y loss of tritium (or deuterium) from the targets in neutron generators is propor- e energtionath o t yl dissipatio e accelerateth f o n D d beam e tritiuTh . m implantation, using a mixed D , T beam, as is usual in sealed tube neutron generators, n dela e ca burn-u e th tritiuy th f o mp target a facto y f b abouso r t 1.5-1.7. In low voltage generators the most common targets are deuterium or tritium absorbe n thii d n metal layers. BesideU sd an titaniu Y d zirconium , man Sc , Er , e alsar o use o product d e intermetallic compounds with deuteriu r tritiummo . Theo- retically e rati th f ,tritiuo o o t mtitaniu m atom s i e abouscas th etn i 1.9:1t bu , of commercial targets it is about 1.5:1. To produce a thin target a 0.2 to 2.5 mg/cm s evaporatei backin W r Z r thic o r o gkg di A T laye metalont, f o rCu o . 2 A 1 mg/cm titanium layer loaded with tritium or deuterium up to this atomic ra- 2 tio corresponds to 0.23 cm3/cm2, or 0.6 Ci/cm2 for tritium only. For a thick tar- + 3 2 a TiTcme 2 ion rangD ges absorben i 1- i Th f st /cm , o s e. 7 ga Zr dr o int i T o target is between 0.12 and 1.05 mg/cm for incident deuteron energies from 25 to 2 40respectivelV 0ke y [1]. Therefore T targeTi f o t 5-1 a , 0 mg/cm s i abou2 0 1 t times thicker than the range of the deuteron ions of a few 100 keV. For the cal- culatio e rangth f a esimplo n e approximatio s givei n Refn i n . [104].

Cu backing

Thermal target tube insulator

Stainless steel

Fig 4 Heavy.13 deuteriume ic target r neutronfo generators

183 Heavy water (D2O n froze)i a e targeusen b s a fordy t mma material e prinTh . - ciple of heavy ice target formation in the vacuum system of a neutron generator is shown in Fig. 134. The target holder is made of a thermally well conducting backing directly cooled by liquid nitrogen. Control of the amount of heavy water vapour will balanc e amoun e evaporateth e th f o t d heavy ice.

beam Collimator

Fig.135 The off-axis target arrangement to utilize the whole surface of targetsTiT

Water outlet- Insulating Target / ring cup , Ttarge- t

Insulated, water- tight electrical contact for target current __. measurement

Rubber 0-rings Cooling 1 Water water inlet

Fig6 Schematic.13 diagram watera f o cooled target assembly

184 Various methods to achieve higher source strengths and target lifetimes have been developed [1] A smal. l rotary targe s i mort e economical tha a nstationar y one [61] and is ideal for applications with medium ion beams, especially to maintain constant high neutron flux ove a lonr g perio f irradiationo d e maiTh . n char-

acteristic f suco s h targets [105] are: maximum bea m, rotatin W powe 0 60 gr speed 60 rev/min, active target 0 cmare, 2 10 a half-lif 0 mA-30 he [106]e coss Th i .t about 40 % of that of the respective number of stationary disc targets. e targeTh t e increaseb lifetim n ca e d significantly witn a hoff-axi s deuteron beam, by which an annular surface of the target will be used. This target-beam geometry is sketched in Fig. 135. A water cooled target assembly, shown in Fig. 136, is able to dissipate a few hundred W/cm2. The target assembly is one of the most important parts of the neutron generator, containing a large amount of tritium and induced radioactivity. Therefore the target and the target assembly should be handled with caution!

14.5.1 Target replacement e e targeth th exhaus Switc f n o o t roomht . Openin a gtarge t assembly involves the exposure of tritium to atmosphere. Close the gate valve of the tar- e hig th t tub hf ge (i evacuu m pump(s s i stil) l working). Slowly expos e vacuuth e m e targeth par f o t t assembl o t atmosphery e usin y nitrogen y dr nitrogeg dr f I s i . n not available, use the vent valve of the target tube. The following description of target exchange is based on the hypothetical target holder shown in Fig. 136. Clos e wateth e e rtarge th inle f o t t cooling e surB . e thae wate th s left ha tr the target cooling cup: blow out the water with compressed air or a blower. Dis- connect the target current meter cable. Opening the target cup should be done with the same tools, used only for target assembling. These tools should be stored separately in a vented glove box: they are expected to be polluted with tritium. During the dismantling of the target assembly rubber gloves should be worn. Put polyethylene foil on the floor under the target assembly. Two polyethylene bags should be kept in the vicinity of the target assembly: when the vacuum system has been opened, the target assembly should be put into the first plastic bag o t avoid further tritium pollution e seconTh . d plasti g shoulba ce pulleb d d onto the open end of the target (beam line) tube. A tritium gas monitor should be read o monitot y e tritiuth r m contamination durin e th assembling e th targe f o gt holder.

185 Hold the tritium target with tweezers or with medical rubber glove covered s replacemen it fingersd targe ol d e e removashoul an e on Th t .th w df to ne l wit a h e dona persob y b e n wearing glove d moutan s h mask. When replacing the old target with a new one, make sure that the gray side (the titanium layer) is on the side of the vacuum! When a new target is placed in the seat of the target holder, switch on the forevacuum pump of the target tube. The vacuum will hold the target: the O-ring on the vacuum side should seal properly. This can be observed by the Pirani or thermo-pair vacuum e targemeteth f o t r tubee e vacuuO-rinth Th . f o gm side should be cleaned with organic solvent and greased slightly with high vacuum silicon grease e tissueTh . s user cleaninfo de targed th O-ringol g te th assembls d an y should be handled as radioactive litter and stored in the special bin for radioactive litter. Whe e targeth n t e vacuusealinth n o mg side seats properly, - assemblwa e th e ter sealing O-ring and the water cooling cup. As the tritium target backing seats in the isolated seat between the two O-rings, the cup should be tightened carefully. During the tightening of the water cooling cup, observe the vacuum meter of the target tube. The sealing, both on the vacuum side and on the water side, should be done properly. When the cup of the target holder has been fixed, connect the water line. Check the target assembly. If the water drops from the targe t p shoulholdercu e tightenee db th , d even more. Attention! When the target separates the water from the vacuum (tritium !!) side, every e targeoperatioth n to assembln y shoul e b carried t carefullyou d . e radiatioDetailth f o s n protection procedures relate o t neutrod n generatorn ca s be found in Refs [107,108]. e electricaTh l connectio a spring s e i propen Figi Th n6 . .13 r connectio- be n tween the target backing and the connector (the figure shows a female BNC connec- e e ensureb propeth n tory ca b rd) contact. After reassemblin e targeth g t holder, the target connector should be tested for conductivity between the target connector and the target housing. A couple of MQ resistance - especially with cooling wate - r doet influencno s e accurac th e etarge th f o ty current measuremene th f i t target current meter circuit has an input resistance in the range of kQ. (The cooling water usually gives a specific resistance of MQ cm between the target and the target housing; if the resistance is measured by the usual multimeter, which uses a couple of volts at resistance measurements.) This water resistance between the target and its housing should be taken into account when a target current integrator is used. The lower input impedance of the integrator can ensure the higher accuracy of the target current measurements.

186 Whe e targeth n t exchang s beeha e n complete e seald contacth an sd an dt have been tested and found to be normal, close the valve of the forevacuum pump, suck- ing down the target tube, and open the gate valve isolating the high vacuum part e neutrooth f n generator froe targemth t tubee readine higTh th . hf o gvacuu - mme ter should reach the normal value in about ten minutes. If the reading of the high vacuum does not reach the normal value, it is an indication of some leaking. Whe e targeth n t exchang s i finishede , collec e th plastit c foil froe th m t i inte t plasti th opu g whic ba c d floo previousld ha han r y covere e targeth d t tube. Plac e th plastie g (anba c d other probably tritium contaminated litter) into the radioactive litter bin. Put back the used target - with its container - into the vented target storage glove box, in the place for used targets. Put back the toold glovean s e targesth user tfo d exchange inte samth o e vented glovx bo e that stores the tritium targets. Wash your hands, and test with a tritium monitor whether they were properly e washedrequireth o D d. administration e relateth o t d targets! Attention! Tritium targets should be handled carefully. The biological half- f lifn tritiuo i ehuma s mga n organ e s th i abousmals1 days t 1 bu lt , chipd an s powder from the titanium layer absorbing the tritium has a much longer biological half-life in the human body and they will act as "hot spot" intense beta radiation sources.

To SAMES beam tube

PTFE RING

Fig. 137 Thin wall tube, air cooled target holder for SAMES neutron generators

187 14.5.2 Air cooled target holder The air cooled target holder shown in Fig. 137 can be connected to the beam a SAMElin f o e S neutron generator e targeTh . t holde s madi r f AlMgSo e i alloy having a relatively low inelastic scattering cross section. The target is held by e thith n wal s i isolatel tubd an ed electricall a Teflo y b yn rin e vacgth sean -o l uum side and a thin Teflon ring on the backing. The small amount of scattering material aroun e targe th dV neutro tMe doet 4 ndistur1 no s fiele th bd aroune th d target spot o this , s arrangemen s i tpropose r accuratfo d e neutron data measurements rinA . g shaped sampl s usualli e y fixed aroun e th dtarge t spor fo altert - ation of the neutron energy. The cooling of the target needs compressed air to bloe targeth w t throug a hnozzl e (air jet) e sam.Th e nozzl s i utilizede o t hol d the spring contact of the target current measuring meter. This arrangement with e samplth e holde s showi r Fig.13n i n 8 [1].

Samples

Sample holder

Compressed air

Target current Fig.13 coolingr 8Ai cooledr ai n oa f target holder with ring shape sample holder

14.5.3 Replacemen r coolee targeai th t f a dto t target holders Attentione th compresse s A s i !sometime r ai d t dehumidifiedno sr ai e th , cooling shoul e checkeb d - speciall e dbeginnin th t a y - g becaus e compresseth e d n bloca w r e targeai wateth o t t r holder. Durin e operatio th ge neutro th f o nn generators the high voltage power supplies should be interlocked by a flow switch of the target cooling water or compressed air. A well collimated ion beam may melt e targeth t backing without cooling, releasin e wholth g e tritium content e intth o vacuum system of the neutron generator. The beam heated target can melt the vacuum seals or a well collimated beam can even punch the target disc itself. Both a faslea o tt d pressure increase vacuuth n i em syste - mi.e e destructio.th f o n the high vacuum - which can be fatal for the high voltage power supply or for the whole neutron generator.

188 To rget Roc>m Control Room

I——— L ; 1—— —1 '""J» ' A ———I — I ^N ——————' "1 1' Thermal Image 1 Analyzer D'9'tal Processor Gas T~ Color MonitoV T r / Telescope L^y^L mml^ Sapphire 4 Sapphin -tt- Window ' "^* Aperture l

Oeuterono ^SkMirrnr / H Beam 1 Target Assembly n 1 ^~\t A, CT*

S (JAERI)Fig.13FN e Th 9 target surface temperature d beaman profile monitor

Bad electrical contact betwee e targeth e targe th d nt an current t meten ca r cause an electrical breakdown between the target and the ground. n additioI e thermocouplth o t n r thermistoro e e videth , o equipment wite th h necessary electronics is sometimes useful equipment for monitoring the temperature of the target. However, a CCD device or similar radiation-sensitive semicon- ducto e rplaceb must n closi dno t e contact wite targeth h t assembly n ordei , o t r avoid radiation damage to these components. It is therefore essential that a suitable distanc s i ensureen a optica y b d l connection betwee e th ntarge t obser- vin ge infrare th mirro d dan r sensitive camera A monitorin. g system complying with this requirement is shown in Fig.139. This system is utilized at the 80° beam t JAERa S IFN [109]e th lin f .o e This equipmen s i basicallt a ysorf o scannint g infrared telescope camera. The infrared radiation originating from the deuteron beam bombarded target surfac s takei e n out y mean b ,a a mirrosapphir f d o s an r e window e outsidth o f t o ,e e vacuuth m beam duct. The thermal image is transformed by the analyzer into video signals and they are transmitted to the digital processor of the analyzer in the control room, aftee necessarth r y data processin e commerciath n i g l model e thermaTh . l imagf o e th ee targeobjecth - t t surfacs displayei a colou- e n o d r monitor a scree n i n

189 OFHC COPPER DISC

TiT COATING

REMOTE DRIVE

RADIATOR SHAPED COOLING SYSTEM

WATER

OUT

PLAC SAMPLER EFO S

e schematicsTh Fige MULTIVOLT0 th .14 f o rotating target with water coolingd an magnetic fluid seal feedthrough

Shaft r\

OJ c Target E a ai CQ

Wobbling target tube Ball bearing

Cogged ring

Fig 1 Operation.14 principle wobblinga f o target holder

190 16-colour temperatur f thio se e us equipmenscale e e highepoweTh th .W k t a tr r beams is important to increase the target life and decrease the tritium pollution roune neutroth d ne vacuu th generato n i m d exhaustan r . Similar equipmen s usei t d in Dresden [110].

14.5.4 Rotatin wobblind gan g target holders The thermal load of the tritium targets can be decreased by rotating or wobbling the targets around the beam. The targets rotate at a speed of several hundred revolution r minutepe s e vacuuTh . m feedthroug f theso h e rotating target- al s low a maximus m 1000 rpme commerciallTh . y available rotatin d watean g r cooled target holder e manufacturear s MULTIVOLTy b d , Crawley, United Kingdom. The schematics of a MULTIVOLT rotating target are shown in Fig. 140. The wobbling target holders do not require vacuum feedthrough. Below 200 W target load the wobbling system is recommended for use at the whole target surface and in order to increase the target lifetime. As the target moves around a circle, the beam will utilize an annular surface of the tritium target. The speed of the rotation is a few revolutions per minute, so the mechanical and vacuum problems of the target rotatioe solveb n simply b dca n e bellow r otheo s r elastic tube [105]. A typical wobbling target is presented in Fig. 141. The circular rotation of the target is carried out by a cam-ring. This cam is fixed onto a cogged ring drivea cogwheel y b n e coggeTh . d ring s rotatei d aroun a dbal l e bearinth n o g standing par f thao t t ball bearing e targeTh . t tub s i forcee y springb d s onte th o e connectiocamTh e .wobblin th f o n g targe e tbea th tub mo t e line tub s madi e f o e stainless steel bellows. This bellows allow a slighs t circular movement causey b d e rotatinth g cam e excentricit.Th e targeth f o ty - rotatioad e - b nwobblin n ca - g justed by the position of the wobbling target tube by holding the flange of the target holder n I thi. s way e tota,th l e tritiusurfacth f mo ee utib target n - ca s e targeth lized t an dlifetim e wile increaseb l a facto y f b 5-10o dr e manufac.Th - tur f wobblino e g target holders require a wels l equipped mechanical workshod an p well trained staff.

191 15. CLOSED CIRCUIT COOLING SYSTEMS

The cooling water consumption of the usual neutron generator is about 0.5 to , whic /h n 1.e supplhig5m ca th h e targe h th d y vacuuan t m (diffusio r turbomoo n - lecular) e publipumpsth s cA . water suppl n i many y developing countrie- un s i s reliable, closed circuit cooling systems are recommended. In tropical or subtropical regions e circulateth , d water shoul e cooleb dd chilledan d . However, commer- cialal water chiller f senergyo consumt o lo s ,othe a e r type f heao s t exchanger may sometimes be preferable.

15.1 THE KAMAN COOLING SYSTEM The A-711 sealed tube generator and the A-1254 pumped neutron generator have the same combined cooling system e PenninTh . g typn sourceio e f theso s e genera- e torlocatear s n higo d h voltage termina d neean l d external cooling e insulaTh . - tion problems related to high voltage are solved by circulating the electrically insulating FREON-113 coolant r otheFo . r neutron generators n sourcio e th e, r higo h voltage terminal cooling uses petroleu r transformemo l coolanoi r t [110]. e coolinTh g e systeKAMAth f mo N neutron generato a compac s i r ts unitha t i : two closed circuit coolant loops with circulating pumps and a refrigerator. The heat exchanger of the cooling unit chills both the circulated target cooling water and the ion source cooling FREON-113. The operation of the cooling system is controllee temperaturth e th y circulateb df e coolano th e y b dt wate flod wan r detectors. A schematic representation of the KAMAN cooling unit is shown in o coolantw Fige .t Th 142flo. w switches e coolantdetecth e losf th d o interts an s - loce operatioth k e neutroth f o n n generator.

HEAT EXCHANGER

| COOLING MACHINE

FREON-113 FLOW SWITCH • - , . . . r * " — * LOOP r I .^ I ^... --.» -* f~ I jii^r™"""? *=•==* I Cc=: I FAN ^ ) FREON-12 1 in LOOP

WATER LOOP

Fig. 142 Schematic diagram of the KAMAN cooling unit

192 A: water pump : wateB r filter : wateC r pressure gauge D: water bypass valve : wateE r sump F: FREON 113 pump G: FREO filte3 N11 r H: FREO bypas3 N11 s valve J: FREON 113 sump K: temperature control (RAMCO) : liquid-eyL e sight glass

Fig. 143 Location of the main components of the KAMAN cooling unit

15.1.1 Maintenance The smooth operation of the cooling system and the neutron generator requires regular daild monthlan y y inspectio e th coolan f o n t e levelth FREON n i s - 113 and the heat exchanger water sumps. The lost of FREON-12 in the cooling ma- e observeb chinn ca e d throug e liquid-eyth h e sigh e tcompressor th glas f o s . Regu- r inspectiola e coolanth f o nt tubin d jointan gs i recommendeds e built-iTh . n filters and flow switches must be cleaned at monthly intervals depending on working hourd conditionsan s a coolan f I . t becomes coloured (especially FREON-113- re ) place it with a fresh one. e wholTh e syste s i mrelativel y simple e coolinTh . g e machinrepaireb n ca ed ba techniciay n fro a mcommercia l refrigerato r conditioneai r o r r service. Trou- bleshooting and repair of the two-coolant circulating system can be done by the operator of the neutron generator. The location of the main components of the cooling unit is shown in Fig. 143. The monthly maintenance routine - based on the description in the Instruc- tion e ManuaA-71th r fo 1l neutron generator, KAMAN, February 1976 editions a s i , follows:

193 sourcn IIo . e cooling loop - Chec e FREON-11th k 3 filter screen - Check the flow rate and FREON-113 either monthly, if the atmosphere is humid (rainy t season)a leas r o t , every three months under normal conditionf I . FREON-113 becomes contaminate r o ddirt t i shouly e b dreplace d even sooner. (Don't forget to find the source of dirt and stop it!) f necessaryI - , adjus e FREON-11th t 3 pump bypass valv r correcfo e t flow rate.

II. Target cooling loop - Clean the water filter screen. If the water is extremely dirty and rusty, chec e floth kw ratd changan e e waterth e . Unless, during regular cleaninf go filter the screethe , n show necesnot be wate the dirt srustyis to -and ry it , sary to drain and refill. - Check the water pressure. If necessary, adjust the water pump bypass valve. - Check the temperature rise of the cooling water during operation of the neutron generator.

III. Check the oil level in the compressor of the cooling unit.

NOTE: Durin l thesal g e maintenance procedures e mainth , power cabl s disi e - connected from the junction box as a safety precaution to prevent inadvertent energizin e systeth f mo g whil e personneth e e e workinequipmentar th l n o ge Th . maintenance operator should hang a "DO NOT SWITCH ON" board on the control conso- e generatorth f o e l . o Therpossibilitn s i e f neutrono y s being generated whee th n main powe r e centracablth o t el control consol s disconnectedi e . The test of the cooling unit should be carried out after maintenance with the service jumper cable, without switching on the whole neutron generator. The maintenance procedure should follow the instruction in the Manual of the given neutron generator. The normal flow rates of the coolants of the cooling system (they should be measured periodically) are:

FREON-113: 2 gallons/min (8.9 1/min) gallons/mi5 WATE7. : 1/min8 Rn (2 )

The temperature of the water in the water sump after 40-60 minutes of operation r 40°Fo C .i5° s

194 Flow indicator

Neutron generator building Isolated water tank (1.5m3|

Neutron generator

Fig4 Closed.14 circuit water cooling system with buried soil-cooled pipes

15.2 CLOSED CIRCUIT COOLING SYSTEM WITH SOIL HEAT EXCHANGER A simple closed circuit water cooling system can be constructed in tropical and subtropical regions wit a heah t exchanger, consistino watetw f r o gpipe s e soilburieth , n i dtakin g advantage e th coolinth o t f o ege e soith du dowl f o n evaporation of the water from the soil. Such a system has been built, tested and utilizee Physicth t a d s Departmen f Chiano t i UniversityMa g , Thailand [112]. The schematic diagram of the system is shown in Fig. 144. The water tank of 1500 litres volume is thermally isolated and placed on the roof of next building. The circulation pump is also placed there and connected to the bottom of the tanke wateTh . r pipe e ar thermalls d buriean yr d ai withouisolatee th y an n i td isolation in the soil between the neutron generator building and the building housing the tank. The two 1/2" dia pipes were buried 10 cm deep in the normal soil. There was no grass or other plants over the tubes. The temperature of the soil remained always around 20°C even durin e hottesy seasonth gdr e t th , day f o s when the temperature of the air rose to 28°C during the night and 39°C during the temperaturw lo s e e dayachievesoith wa Th l . f o ey sprayin b d e soith gl with water twr threo o e time a days . Evaporatio f wateo n r consumes e energydryinth f o o s g, the soil cools it down. This "heat exchanger" worked satisfactorily when sprayed with wate a couplr f timeo e a days e syste.Th m also worked satisfactorily durin e lonth g g working days, withou y decreasan t n pumpini e g speea 200 s f 1/ o 0d diffusion pump.

195 REMARK: As the evaporation of the water from the soil is utilized for chill- e inth coolang t circulatee closeth n i d d circuit e burie th e are th f ,o d a pipes should be sprayed with water about one hour before the operation of the cooling system so as to start with water chilled in the soil. This syste s sommha e extra tap n casi sf emergencyo e e neutroth s A .n generator needs only electric energy, the cutoff of the electric power and the utiliza- a diffusiotio f o n n pump require some precautions e normaTh . e watel th flo f rwo is detected in the return pipe to the water tank by a flow-actuated switch as shown in Fig. 145.

Water return Microswitch

Closed circuit water tank or sink

Fig5 Water14 . flow detection switch r waterfo cooling systems

The water from the return pipe of the cooling water loops flows through a hollow cylinder e weigh Th d .pressur e an tth flowin f o e g water pushes dowe th n cylinder and, througe th contac a , it hmicroswitchf o t e Th threshol.e th f o d switching depends on the water flow rate of the coolant and it can be adjusted by the balancing weight of the lever. The tank of the cooling water should be protected against dirt wit a filteh r mesh. In case of emergency (cutoff of the electric power), when the diffusion pump needs further water cooling, the water can flow through the circulation pump owing to the height difference between the tank and the diffusion pump. In this case the operator of the system should close the V~ valve in the return leg and ope e th nemergenc y water outlet periodicall e valv , th V- (sey e b y e Fig. 144)- De . pending on the water consumption of the diffusion pump, the water should be removed from the diffusion pump every 2-3 minutes. The cooling time of the diffusion pump is about 30 min. This means that the V-, valve should be opened and close 0 timer abou1 dfo s t 20-30 seconds. Usin a fasg t cooling loop aroune th d diffusion pump, in case of emergency the cooling down of the diffusion pump will e faster b d watean , r consumption fro e tanmth k wil e lessb l .

196 16. PNEUMATIC SAMPLE TRANSFER SYSTEMS

A pneumatic transfer syste s i mrequire o transport d t samples betwee- ir e th n radiatio d measurinan n g sites. Such system e user ar activatiosfo d n analysiy b s reactory intensb r o se radioactive neutron sources (Cf,AmBe,PuBe,etc)- re a n I . actor, the neutron field is homogeneous, while for neutron sources or neutron generators in a nonmoderated arrangement, the irradiation is carried out by point sources. Thes caseo e demonstratetw e ar s Fign i d. 146.

Neutron source

Fig. 146 Irradiation of samples in isotropic and nonisotropic neutron fields (a) Reactor irradiation of thin, nonabsorbing sample isotropican in neutron field (b) Irradiation of a scattering-free sample in a neutron field produced by a point source

a reactor r e Fo averagth , e activating flux e dependfluth x n o distortios n d self-absorptioan n e samplecauseth y b da , poin r whilfo t esourc f faso e t neutron e averagth s e fluenc s i determinee e th source-sampl y b d e geometry n I .addi - e geometryth tio o t ne neutroth , n attenuatio a thic n i nk sampl n influencca e e th e activity distribution, demanding a careful evaluation of the measured gamma-line intensities. Therefore, during irradiation and measurement, the cylindrical sam- plee ar rotates d aroun o axetw ds perpendicula o t reac h other. This solutios i n e KAMAth use t a d N twin tube pneumatic system. Unfortunately, this method requires complicated sample holders at the irradiation and measuring sites. In the case of periodical irradiatio d measuremenan n t (require r shorfo d t half-life isotopes) the position n changca s e randomly, resultin n a averag n i g e e activth valu r - fo e ity. A proper pneumatic transfer system for activation analysis with neutron generators should be a twin tube with:

197 a) Sample rotating system for cylindrical samples, or b) Rectangular tubes for disc shaped samples (irradiated in the direction e sample th e axi of th fo s )t neewhicno do d hrotatio e sampleth f o n , c) Neutron production (e.g. ion source) switch on and off facility, d) Accurate time controller for the irradiation and measurement, ) e Sample position detector e irradiatioth t a s n (neutron generatord an ) measurement (gamma detector)a sided an , f) Loading (ejecting e samples) th por r fo t .

SOURCN IO E Cooling-1 MEASUREMENT Cooling-2 START STOP START STOP

I< — T(A) — > I< •I<——TC2——>I

sample transfer sample transfer

Fig7 Timing.14 diagram pneumatica f o sample transfer system r periodicfo irradiation and measurements

::3

Fig8 Twin.14 rectangular tube pneumatic transfer system r activationfo analysis neutronby generator

198 e timTh e programmer should ensur e e changirradiatioth th e f o e n coolind an g measuring times in a wide range. A typical cyclic timing diagram of a pneumatic sample transfer syste s i mshow n i Fign . e 147cyclTh . e starts wite activatioth h n , tim followee firsTA eth t y b coolind g TC, e seconth , measurind dan coolinM T g g

TC2 times [113]. The schematics of a pneumatic transfer system with twin rectangular tubes, develope e speciath r fo ld requirement f activatioo s n analysis with neutron generators e showar , n Figi n . 148. o separattw e eTh rectangular tubes transportin e standare th samplth gd an de sample are controlled by two electropneumatic valves (1). The slowing down or breaking of the samples at the irradiation or detector position is carried out d (3) an e valvebth .y) Thes(2 s e valves ensur e "softth e " e samplearrivath f o l s in the irradiation or measurement positions. The microswitches (4) detect the ar- rival of the samples in the correct positions. The compressed air supplied by the compressor (5) is buffered by the buffer tank (6). The whole system [114] is controlled by a microprocessor controller. The timer-controller can be made from a parallel input/output card of a personal computer or other independent industrial timer.

199 17. NANOSECOND PULSED NEUTRON GENERATORS

The nanosecond bunching of steady deuteron beams for the production of short 14 MeV pulses can be done before or after the acceleration. Two typical systems will be described here: a pre- and a post-acceleration bunched neutron generator. A descriptio e th principle f o n f theio s r operatio r fa beyon s i ne scopth df o e this Manual.

17.1 PRE-ACCELERATION NANOSECOND BUNCHED ION BEAM NEUTRON GENERATOR

The block diagram of a pre-acceleration bunched nanosecond neutron generator is shown in Fig. 149 [115]. This neutron generator has the following characteristics: Q - Average neutros nn/ output 0 1 < : - Neutron yield in pulses: 4 x 10 n/s - Average target current: 10-30 - Beam diameter: 8 mm - Beam current during the pulses: > 1 mA - Acceleration voltage: maximuV k 0 m30 - Target on the acceleration high voltage - HV power supply: single wave Cockcroft-Walton circuit - Ion source: RF (200 W push-pull oscillator) - Extraction voltage: 0-15 kV - Focus voltage:V k 0-20 s consumptio sourcen Ga - io 2 ml/hou5 e D 4- th : P f o n NT r - Vacuum system l diffusiooi s : 1/ 120n0 pump with boosterh mechanicam/ 0 2 , l duplex pump, liquid nitrogen trap - Fina " lmBa 10 vacuum x r 2 wit < :h liquid nitrogen trap - Pulsing: twin gap klystron bunching with chopper and selector plates and X steerers - Compression factor0 1 > : - Pulse e neutrowidtth f o h n s pulsesn 1 ~ : - Repetitio ne neutro th rat f o e nz pulsesMH 0 1 : - 1.25 - Targe t aluminum holderm 3 0. : m tube with aluminium backin T targegTi t - Target cooling: by compressed air - Power consumptio e A generatorth kV f o n5 . ca : - Water consumption of the vacuum system: ca. 5 1/min - Compresse r consumptionai d . 500ca h : 01/

200 n HSourcIo F e Focus I Focus I Pulsing Unit Acceleration Tube Target Head —

1 1 | 1 1 1 1 1 | 1—

FT 10MHz 20MHz HV Power I-20KV Power Powtr Supply Amplifier Amplifier Pickup -200V 4- k- Ampl. P.S Integrator ^^ r~C'l Phase AC | Voltage 1 1 1 Shifter Regulator 1 1 1 •*- I 1- ————————— ^^ ——— - i I ^ 1 n Io Source Focus Base I Control Panel Control Panel Oscilldtor —— — — •—I 1

Units settled I Units placed e th Unit n o s e th Closo t e I in the ion Control desk Ion optics 1 Source block

Fig.149 Block diagram pre-accelerationa f o nanosecond pulsed neutron generator

oNJ e peculiaritTh f o thiy s generato s i e thartargeth te s i higplaceth t ht a d voltage. As the shadow bar and the neutron detector or gamma detector in time-of- flight measurements shoul e placeb a dcertai t a d n distance froe sampleth m d an , may be placed also at the acceleration high voltage, the benefit of this system is the easy pulse control of the extracted and pre-accelerated deuteron beam. The use of the deuteron beam pulse pick-up signal needs optical insulation. The target curren s i similarlt y converted into frequenc d opticallan y y connectee th o t d integrator placed at the ground potential. As the beam chopper, the beam buncher and the selector electrodes are at the ground potentialsourcion is floatinvoltagfirsethe the the , t at of gefocu s (pre-accelerating) power supply. This focus lens is an immersion type cylindrical len , sfocusin e extracteth g d pre-acceleratean d d deuteron beam throug e maith h n vacuum manifold inte entrance th oseconth f o de Einzel focus lense Focus-ITh . I unit consist f threo s e diaphragms wit a hshor t focus. This lens focuse e deuth s - teron beam inte p twibuncherth oga n e chopper-steereTh . r deflector plates (two pairse bunche th n froni )f o t r chop e steadth s y deuteron beam wit a frequench y steerin~ UD d g an MHz0 o1 1 f e positioe controllee b UD beaTh . th e n mf th ca o n y b d voltages. The second deflector plate pair deflects and chops the pre-accelerated deuteron beam - whose energy is determined by the sum of the extraction voltage and the voltage of the first focus electrode. As the chopping frequency is 10 MHz, the buncher electrode works at a 20 MHz frequency. The phase control be- z signal MH e pulss mad i sth 0 2 en i e d controan en z e l MH th unit 0 s 1 tweeA . e th n ergy of the pre-accelerated deuteron beam is relatively low, the diaphragm between the chopper and the buncher electrode does not need any extra (water) cooling. This diaphragm is made of stainless steel and its chopper side is covered by a titanium sheet. Similarly, a titanium sheet covers the selector side of the post-selector diaphragm e z repetitioe pulseTh th bunche. MH f o 0 s2 nd ratn io e beam can be selected at 1.25-10 MHz. As this selection is controlled by normal frequency dividers and phase shifting units, the additional use of a pseudo- random pulse selecting unit is easy [116]. The geometry of the ion optical system of this nanosecond pulsed accelerator is shown in Fig. 150. The first focus lens was necessary to get pre-accelerated deuteron beam (an d o decreast n beaio me th escatterin e residuath f o g l vacuu- mal ong the 80 cm long path in the main vacuum manifold). The Einzel lens - as it does not change the energy of the focused beam - easily focuses the pre- accelerated deuteron beam into the chopper-buncher-selector-acceleratortube ion optical line.

202 Pulsing unit

Ion Sourc I eImmersio n | Einzelens l lens Chopper Buncher Selector Ace. tube Pick 80cm I 20cm I 25cm m c 0 9 up Suppr. TiT Target

o «• U pxtr. -o •30kV

Fig.150 Geometry pre-acceleratione th f o bunched nanosecond neutron generator BEAM LINE s 2n 30ns 1st SLIT DEFLECTOR 7nd , "nL~ BUNCHER ""if ... PICK-OFF ^...... :<...... k...... jr.^.*.A^k I,. .. I LAMi- I™ TAR"T j ^ ,ii——•••••••,•••""""""• •<—0 • • • •

ELECTRIC FIELD SOOns

\ TIME

DEUTERON INTENSITY | I I I 111 I TIME

TIME

J DEUTERON -JUlHl INTENSITY

1 i 1 I TIME Fig.151 Typical post-acceleration klystron bunched neutron generator baseda n o commercial neutron generator

17.2 POST-ACCELERATION KLYSTRON BUNCHING OF A COMMERCIAL NEUTRON GENERATOR A post-acceleration klystron bunching nanosecond neutron generator was constructe n i Chiand i (ThailandMa g ) a basecommercia n o d l SAMES J-25 neutron generator. The extracted and accelerated beam is chopped by a 2 MHz electrostatic deflector and bunched by a 4 MHz two-gap klystron. The schematic representation of the post-acceleration nanosecond pulsing is shown in Fig.151. The operation of e systeth e s mshowcorrespondini th y b n g waveforms [111]. energV ke - 0 ypreviousl 15 e Th y selecte - ddeutero n beam enter e bunchinth s g e sectiogeneratoth f o n r throug e firsth h t water cooled e slitTh horizonta. - de l

204 d SLIT-^-j»-n 2 30ns BUNCHER

X-STEERER SUPPLY

WIDTH CONTROLLER

AND PHASE CONTROL PANEL STEERER CONTROL PANEL CHOPPE D BUNCHEAN R R

CONTROL PANEL

BEAK CURRENT

MONITOR BOARD

Fig. 152 Block diagram of klystron bunched J-25 neutron generator

fleeter plates chop the deuteron beam by 2 MHz sinusoidal voltage ( (a) in Fig. 151)e selectioe Th pulses.th y f changinb o n, e th grepetitio n frequencys i , carried out in front of the klystron buncher by the second vertical deflection pairse Th horizonta. l deflection result s wid n n i abou0 se5 triangulat r pulses (b). These deuteron pulses are selected by the Y deflector plates (c). The deflected pulses hit the second slit in the beam line before the two-gap klystron. The klystron itself works at a frequency of 4 MHz (d) and bunches the chopped deuteron beam onte target th oe deutero Th . e detecte ar ns widt n ) pulse2 (e h df o s ba capacitivy e pick-off electrod e vicinite targeth th n f i eo ty [118]. e steerine deuteroTh th f o g n bea s mi solve y additionab d l steering voltages on the deflector plates X and Y. The control of the nanosecond mode of the neutron generator is monitored by oscilloscopic observation of the M2 , M~ and M. e slitspulsTh . e forms give good informatio e problemth n o n s e relateoperth o -t d e systemth atio f o nblocA . k diagra e th nanosecon f mo d pulsin f thio g s J-25 commercial neutron generato s i showr n i Fign . e 152horizontaTh . l dashed line divides e blocth k diagram e acceleratointth o e controth r d halan ll room e beaTh . m current and beam monitor panel is the unit where the beam shapes and beam currents at the diaphragms and at the target are monitored.

205 18. THE ASSOCIATED PARTICLE METHOD

The associated particle method (APM) is widely used for determination of the absolute neutron emission rates and to ensure the electronic collimation of the neutron beam in TOP (time-of-flight) and other coincidence measurements. The alpha particles emitted in the T(d,n) He reaction are detected in a well defined solid angle around a given emission angle to the incident deuteron beam, the e neutroth erro n i r n yield being minimize n alpha r afo d detectoo t r ° place90 t a d the deuteron beam n I .thi s geometr e e th spreay th neutro n i d n e energth s i y lowest. Generally a surfac, e barrie i detectoS r s i use ro t detecd e alphth t a particles. a fase casf th to e n chargeI d particle detector (usually thin plastic scintillators e timinth ) g featur s i excellene o t e giv startth P meae tTO -signan i l surements, and the solid angle of the alpha detector - due to the reaction kine- matic - wils l determin a ecoincidenc e fas th cont f o neutrone s froe targetth m . These fast neutrons will be utilized later as the primary neutrons for the study of neutron induced reactions. typicaA targeM AP lt hea - utilize da commercia t a d l neutron generator is show n i Fign e . deutero153Th . n bea s mi collimate e d targefocuseth an d n to d and the alphas are detected by a scintillation detector. The scintillation detector foil (0.0 thicm 5E 102Am N k s )mountei d insid e targeth e t chamber, while th e photomultiplier is optically connected by a light-guide Perspex cone. The solid angle in which the coincidence 14 MeV neutrons can be detected is also indicated [119]. The surface barrier (SB) detectors have excellent energy resolution (compared to the scintillators) which allows the measurement of the deuteron or helium buildup withi e tritiuth n m target e tritiuTh . s i mradioactiv d decayan e o t s He by beta emission. During one year about 6% of the tritium will be replaced by 3 , causinHe a gbackgroun e associateth n i d d alpha particle detecto- rre froe th m action e e alphenergth He(d,pTh f a . o y particleHe ) s froe d th H(d,nman e H ) 3He(d,p ) 4He reactions are almost the same and can only be separated by a detector of high resolution. a Fortunatelresonanc keV0 s 44 ha , t e e a reactioeH th y n o n while the reaction on tritium is at 110 keV. On the other hand, the He particles can escape from the target, depending on the temperature. At about 400 keV incident deuteron energy (this is not rare at neutron generators) the cross sections o processetw e th ofs become equal, whic n causca h e large e determinaerrorth n i s - tion of the neutron yield if this effect is neglected [120]. neutroV k 0 40 n a generato r Fo r e yead tritiuusinon ol r a gm - targetex n a , cess count of about 6% will come from the helium reaction. As the energy of the

206 Scintillotor Aluminium sheet foil Ti-T target

Oeuteron beam

3 AssociatedFig15 . alpha particle target head using thin plastic scintillation detector

HO 1SO CHANNEL

Fig.154 Associated alpha particle spectrum showing e He(d,p)th effectf o e H 3 4 and 2H(d,p) 3H reactions

207 incident deuteron energy is lowered the effect is reduced, but even at energies of 200 keV in a fairly new target this kind of error could result in an uncer- . % 5 taint0. f o y Fig. 154 shows an associated particle spectrum around the alpha peak for an old tritium target at 400 keV bombarding deuteron energy. Peaks 1 and 2 corre- e reactionse H alph- sponth re e o at e dth particle th . d s Peai an 3 kH s froe mth suit of the proton buildup from the 2H(d,p ) 3H reaction. The energy difference of e energ th o alph tht tw ey bu a resolutioV peak e detecto ke th s i abou s 0 f o n35 t r does not allow a better separation [120].

18.1 SELF-TARGET FORMATIO DEUTEROY NB N DRIVE-IN The associated particle target head with semiconductor detector is a good toor studyinfo l e e buildue th beath th deuterone g mth f targed o p an n i t s aperture materials e presence Th self-targeth . f o e t will V contaminatMe 4 1 e th e neutron spectru M targe mreactionD AP D- througt n heasketca A e .f th do hh - made for the study of self-target formation and (with plastic scintillators) time-of-flight measurements - is shown in Fig. 155. heade M neutroth AP , n Basena d contaminatioyieln o dan neuT d D- -e th f o n tron spectrue determineb n ca m y observatiob d e chargeth f o nd particle spectrum. The charged particle spectrum measured by a surface barrier semiconductor detec- n givca e r to informatio e conditioth e tritiuth n o f ne o m backnth targed - an t ground neutrons typicaA M .spectru AP l s mshowi n i nFig . 156, which showe th s

FTrv ^ "&' jT,////; ; ss / /•/; s-//;; •> ss.r^, ^ o* Si SB Detector

Fig. 155 Associated particle target head used for self-target formation and fast neutron time-of-flight spectrometry

208 PeaB k

3H(d,n)4He : 2.74 MeV

3 4 He(d.p) He PeaC k £„,: 2.87 MeV 2H(d.p)3H E :2.6V 1Me p \

Fig. 156 Typical pulse height distribution of the charged particles detected by surface barrier detector usingtargetAPM head

0(d.p)3H REACTION 10H

'H PROTONS TRITONS 500-

50 0 35 0 10 400 N

Fig.157 Charged particle spectrum measured from H(d,p)the APM reactionby H 2 3

—O — STARTP T-

Fig. 158 Block diagram of the electronics needed for studies with a simple associated particle target head

209 e alphpeakth f ao s particle d protonan s s froe e H(d,nmth th , d He(d,pHe ) an e H ) OO*2 A H(d,p) H, He(d,p) He reactions, respectively. The tritons originating from the D-D reaction can be observed at low channel numbers [121]. Observatio e peath A givekf o na possibilits r correctiofo ye countth f o sn detected in the peak B originating from the alphas. The peak of the protons from e th H(d,p 2H reactio) n3 give a goos d possibilit o monitot y e amounth r f drive-io t n deuterons in the tritium target. The charged particle pulse height spectrum from D-D reactions is shown in Fig. 157. The electronics of the APM technique depends on the purpose of the measurement r monitorinFo . r o gself-targe t buildup studies, a simpl e linear amplifier line shown in Fig. 158 can be utilized. The MCA can be replaced by a single channel analyzea counte d r recordinan fo r r e alphth g a particle d protonan s s froe th m 3H(d,n) 24d H(d,p)Han e 3H reactions, respectively. n electroniA c system require r producinfo d e starth g F tmeasure signaTO t a l - ment s presentei s Fign i d . 158. This syste n als me useca ob r monitoring fo d .

210 19. NEUTRON MONITORS

Monitoring the neutron yield of neutron generators during the irradiation of sample s i importans r determinatiofo t f o nuclean r cross sectionn i activatio r o s n analysis e neutroTh . e monitorenb fieln ca d: dby

a) Alpha particle detection, using the associated particle method; ) Protob n recoil detector (e.g. hydrogen filled proportional counters, organic scintillators, etc); c) Long counter; d) Fission chambers.

Monitoring neutron production by alpha particles has been treated in the pre- previous section. The use of a proton recoil detector is not discussed in this s treatei Manual t i t n Refi d bu ,s [1,2,123].

19.1 MONITORIN LONY GB G COUNTER The long counter is the simplest neutron monitor, with a flat-shaped energy efficiency a curvwid n i e energy range. These broad, flat energy efficiency curves allow easy absolute calibration by a standard neutron source. As the long

, 1/2counteBF " r

S$^~3-^

metal tube

paraffin

Fig. 159 Schematic diagram of a long counter (sizes in mm)

211 LONG COUNTER

Fig.160 Electronics neutrona r fo long counter

counter detect e neutrone th sth y B(n,ab s i L reactio) n (after thermalization), the pulses froe mth gamm a rays associated wite neutroth h e welb n ln fielca d separated e lonTh .g counte a s direction-sensitivi r e detector s constructioit ; n is show n Fig i n9 [122] .15 . The setup of a long counter's electronics is shown in Fig.160. As can be seen in Fig. 160, setting the optimal discrimination level can be helped by a multichannel analyzer. For monitoring and recording the neutron flux, a multichannel analyzer may be used in the multiscaler mode [123]. e mechanicaTh l constructio a lon f go n counter depende neutroth n o sn modera- tor materiale casf th o polyethylene n I . r o similae r good machineable materials containing high concentration f o hydrogens e internath , le exterth cylinde d - an r nal shield can be machined by lathe.

19.2 FISSION CHAMBER MONITORING The fission chamber is a cheap and reliable piece of equipment for recording fast neutrons throug e detectioth h f o fission n fragments e frofissioth mf o n . ThoriuU r mo 9^ dioxidh 9 T r uraniuo e 9^m8 tertafluoride layer 2 mg/c0. s m thick 9 and 15-20 mm dia are used. The chamber shown in Fig.161 was constructed from a thin-wall aluminium cosmetic cream box. The pressure of the counting gas (methane or argon) is just slightly over the atmospheric pressure [124]. e pressure fissioTh th n i ne chambee regulateb a n regulato ca y rb d r valve. e thiTh n wall aluminiu s i simpl x ymbo seale y self-adhesivb d s inleega ttapee Th . and outlet tube e thiar sn polyethylene tubes e detectioTh . f countino n s floga gw d over-pressuran a smal s carriey i eb lt silicoou d l bubblinoi n g glass vessel. e . vesseWitth cm s i e aboun e hsilicoi lheighth 4 Th l methane2- f toi o nt , the exhaust of the fission chamber should be led outside to prevent the formation n explosiva of e methane-air mixture e collectoTh .a wel s i lr polished thin metal disc held by a metal rod soldered to a high voltage BNC socket. The usual voltage e appliefissioth o t nd chamber throug e simplth h e preamplifie. V s i abou0 r 60 t

212 Pressure regulator Self adhesive tape

Methane x bo Thi l nA or argos ga n bottle

238 BNC socket (to preamplifier)

5cm

Exhaust

Silicon oiU C-

300

CHANNEL

Fig. 161 Schematics of a fission chamber and typical pulse height distribution of the fragments

213 The insulating feedthrough is made of Teflon (PTFE) or similar good insulator. The whole monitor counter consists of the fission chamber, preamplifier, high voltage power supply, main (spectroscopy) amplifier, multichannel analyzed an r single channel analyzer-counter. For longer irradiations a monitor counter with time-programmed printer or MCA in multiscaling mode is recommended. A typical pulse height distribution of the fragments is shown in Fig. 161. If the shape of the fission fragments spectrum has changed, check the over- pressure in the detector box. The bubbling gas flow indicator indicates the proper pressure in the chamber. If there are no bubbles, check the built-in manome- s cylinder ga e cylindet th emptye no f th e I manometes . i th f r,o r te r shows over- pressure. Clos e polyethylenth e ea clip tuby )(b e betwee e fissioth n n chambed an r the regulato e d fissiochec th an e rlea th n ki nk chambe r betweeo r e chambeth n r ane bubblth d e vessel.

214 . SAFETY20 : HAZARDS RELATE NEUTROO DT N GENERATORS

e maiTh n source f hazardo s s e relateoperationth o t d , maintenance, troubleshooting and repair of neutron generators are as follows:

Neutro d X-ranan y radiation; Open radioactive source (tritium); Residual radiation of the neutron-irradiated construction materials; High voltage power n suppliesourceio r fo ,s extraction, focud an s acceleration; Mains voltages; Pneumatic pressure vessels and tubes; Vacuum vessel d chambersan s ; Poisonou s (SF^ga s ) pressure;

Flammable deuterium gas (D2); Flammable insulating transformer oil.

20.1 Radiation hazard Maintenanc d repaian ef o rneutro n generators mus e b performet d eithey b r personnel trained by the manufacturers or other by experts in high vacuum and high voltage techniques and in handling tritium. In designing the shielding, consideration should be given to the fact that radiation levels from neutron 2 Sv/s higa ss h a h (200 rem/h e commom ar ) 1 t a n from a neutron generator [107]. e shieldinTh g materia a laborator r fo l y often r costmor o s muca s es a htha n the actual neutron generator, and care is needed in designing a building to house such equipment e maximuTh . m recommended weekly dos o personnet e l al type r sfo l

Tabl. e20 Maximum permissible neutron fluxed an s fluences to personnel

En(MeV) Flux ) s (n/cm Fluence (n/c) m

Thermal neutrons 268 38.8 x 106 neutronV Me 1 s0. 32 4.8 x 106 0.2 MeV neutrons 8 1.16 x 106 neutronV 10-3Me 0 s 4 6.0 x 105

215 of radiation is 4 x 10~4 Sv/40 h (40 mrem/40 h) or 1.0 x 10"5 Sv/h (1.0 mrem/h). Particle fluxes (n/cm2 s) equivalent . to 1.0 x 10 5 Sv/h and fluences (n/cm ) delivering 4 x 10" Sv are given in Table 20. The maximum recommended dose for the public is 0.5 x 10 Sv/h; that is, 2 abou n/c2 0. t m. Neutros V neutron Me 4 n1 n i sgenerator d an producV n ca sMe 3 e T reactionsD- d an , D respectivelyD- D reactio e D- yiel f Th s .o i lowe dn r than that of D-T by a factor of 100, and the energy of neutrons is also lower for D-D. Therefore, shielding shoul e constructeb d a radiatio r fo dV neunMe 4 hazar1 - f o d trons e flu f neutrono Th x. e followins givei th s y nb g expression:

-x x e^ (48)

e sourcth s wheri e S intensite e distancth y 0 n/s)1 s (e.gi x e R 5 ,.betwee2. n th e give th sourc nd an poine t wher e radiatioth e n exposure shoul e determinedb d , x is the thickness of the shielding, and 2 is the removal cross section,

Table 21. Removal cross sections of shielding materials for fission and 14 MeV neutrons

Fission neutrons En = V 14.Me 5 Material 2 2 A (cm'1) (cm) (cm'1) (cm)

Water 0.103 9.7 0.079 12.7 Iron 0.158 6.3 0.112 8.9 Concrete (2.4 g/cm) 0.077 13.0 Concrete (3.5 g/cm ) 0.094 10.6 0.08 12.5 Barytes concrete 0.095 10.6 0.083 12.1 (3.5 g/cm3) Ironshot concrete _ . 0.071 14.3 Gravel (1.83 g/cm3) 0.052 19.2 Sand (1.6 g/cm3) 0.047 21.3 Brick (1.83 g/cm3) 0.048 20.8 Graphite (1.83 g/cm3) 0.079 12.7 0.058 17.2 Paraffin - - 0.071 14.1 Aluminium - - 0.063 15.9 Lead - - 0.088 11.4

216 £ = I/A, where A is the relaxation length of fast neutrons in the shielding mate- e remova th n riaTabl (i , l 21 el cross section e give ar r ssom fo n e shielding materials) e dosTh . e leve s ln/ indi obtainea sourc 0 -r 1 fo x e d 5 intensit2. f o y cates that, at a distance of at least 2 m from the target, an additional 1.5 m •2 wall of concrete (p = 2.3 g/cm ) is needed. On the basis of the removal cross sections summarized in Table 21, the doses can be estimated both for D-D and D-T neutrons. The macroscopic removal cross section is given by 0.602o/> . 1 2r = A ^ (cm' ) (49) e microscopith s wheri a e c removal cross sectioe densityth n barnsd i s ni an , p , A is the atomic weight. The macroscopic removal cross section for a material com- perising several elements is obtained by simple summation over its constituents:

+ 50 2r compound = - ^ PI + - />2 - ( >

where (2 /p)- is calculated by the data in Table 21 of the i-th compound 2 r i 3 (cm /g), and p- is the effective density of the i-th element (g/cm ). Neutron n contributsca scatterer ai ey b dsignificantl e totath lo t ydose f I . a barrier shield were r usedscatterinai , g abov e sourcth e e ("skyshine") would contribute more to the dose than to the transmitted radiation. Thus, there is no poin n i attemptint o t shiel ge directh d t radiation unless skyshin s i - alsere o e provisioth duce a y shieldinb f do n g e roofpresencTh . f o ductse , voids, pas- sageways, and safety doors all require careful consideration in the design of the final facility [1].

Tabl. e22 Expected level f induceo s d radioactivit aftem c r0 1 t a y 1 h of operation with a neutron generator yield of 2.5 x 10 n/s

Exposurm c e0 1 rat t a e Reaction ———— fmR \ C /* f [iqTTiJ Half-life

0 2720 Al(n,p)27Mg 5 9. 52 min 27Al(n,a)24Na 30 7.7 14.h 9 630 6 Cu(n,2n)62Cu 15 9.8 min 0 656 Cu(n,2n)64Cu 15 12.h 8

217 Gamma radiation produced by the scattered and captured neutrons in the shield can contribute significantly to the radiation hazard. Calculation and measurements show that for thick (a 1 cm) concrete shields, the gamma dose rate is about hal o epicadmiut f e thadu t m neutrons n waterI . e gammth , a dose exceede th s fast neutro ne thicknesth dos f i e s e greatei spromp Th . r tcm tha gamm5 7 n a emission of 4 to 7.6 MeV from the Fe(n,y) Fe reaction is a problem if the shielding contains iron. The neutron-induced radioactivity of structural materials can become a sig- e coolinth d gnificanan water e ai th t d r N(n,2n an e problemth N )n I . 14 13 O(n,a) N reactions will take place. Some expected exposur f m followino c operatio e 0 1 h 1 t rate a t ga n s 2.5 x 10 n/s produced by the target material are summarized in Table 22. Backstreaming electron n producca s e bremsstrahlun e higth h t a gvoltag e terminal, wit a h dosa nonanalyzemR/ 0 r e80 fo hrato t f abou o e0 60 dt beam e dosTh .e ratf o bremsstrahlune e b decrease n ca g d significantl n a electro f i y n suppressor is applied close to the target. During the operation of D-T neutron sources, contamination or radiological hazard can occur in the environment caused by tritium outgassed from the target. A potential tritium hazard exists also when the system is opened to the atmosphere for any reason (i.e. repair, maintenance, target replacement, deuterium (leak) replacement n sourcio F e H ,balloo n replacement, breakage, pump exhausting inte generatoth o r hall, etc.).

It has been proven that the tritium in gas and T9O vapour forms takes part in chemical reactions in the same manner as the hydrogen gas and H^O vapour. The half-life of tritium is relatively long: it is about 12 y. The specific activity 0 s i 9.91 Bq/g x 9O .foT2 r Tritiu s i muniforml y distribute e bodth yn i dwithi n 90 e minbiologicaTh . l half-lif f tritiuo e s i mshor - about days2 e 1 gaseoutTh . s O vapouJ^ T - ~ r atmospheri cr day pe exchang . % Tritiu e ordeI th ef o n mri rat s i e in gaseous form is absorbed by the lungs at a rate of 0.1 %. The maximum permis- siblr concentratioai e r w hour0 tritiufe fo 1 Bq/cn ss A i mm afte . r exposure, -2 3

the body fluids contai e samth n e d Tconcentratio2an O ; O thereforHT f o n e urine analysis indicate e leve th f so incorporation l A tritiu. m concentratio Bq/1 1f o n in urine represent e maximuth s m permissible dose. All components of the ion source, accelerating tube, beam transport system, target system, vacuum equipment, exhaust linesd targean , t cooling water system become tritium-contaminated. A small amount, about 5 %, of the total tritium released from the target is deposited on the static components of the generator. e tritiu th e bule elementf Th th s o n maccumulateki io d f an o e spum th l oi n pi d pump or vented to the atmosphere by the roughing pump. In the forepump oil the

218 0 e timediffu2 th contaminatio o f st -o highe0 l 1 oi e s e founrb th wa nthao t dn i n sion pump. During repai d maintenancean r , contaminatios na w mus lo e kep b ts a t possible. For example, components in the vacuum system of RTNS-II near the target 7 " hav 7 e 10f \ to 10 dpm/cm of surface contamination which is detectable by wiping. Routine target change n causca s e tritium incorporatio y personneb n a leve t a ll whic s i observablh n urinei ee handlinTh . d replacemenan g f tritiuo t m target- re s quire the use of a well ventilated glove box to reduce the hazards from inhala- tion of tritium gas, especially the large tritium-carrying chunks created by surface erosion. For an intense neutron source, the amount of tritium released in a o greayeato o exhaust s ti r t inte atmosphereoth . There are two capture techniques to retain the tritium released from the target: (1) trapping in ion pumps and bulk sublimators or (2) using a tritium scrubber that convert e th tritius a o watem t moleculan d bind o an rt i s r sieve bed. It was found that tritium released from the gas-in-metal target with an out- 0 1 Bq/ o t puh througt 0 1 valu e f scrubbeth o he r syste11 s mi typicall 12y less than 7 3 x 10 Bq/d. The tritium evolution rate during storage depends on the type of a lowe s ha r valud carrier argoan fo es nga r tha r airnfo . Special attention shoul e paib t da neutrod n generators with titanium getter pumps. The amount of tritium (-10 Bq/h) will be captured mainly by the ion getter pump. The estimated amount of tritium found in a gaseous state at any time durin e operatiogth e neutroth f no n generato. Bq s aboui r 0 1 t n getteio e rTh pump goes throug a hheatin g stage s whestartedi t i d n an , consequently tritium is released into the vacuum system.

20.2 RADIOACTIVE MATERIAL STORAGE AND WASTE DISPOSAL HAZARD Neutron generators utilize tritium targets of 3-37 x 10 Bq (approx. 1-10 curie) activity. Spare targets may also be stocked as replacements. All of this radioactive material must be used or stored in an exclusion or storage area in accordance with the regulations. During the operation of neutron generators, a certain amount of radioactive waste will be accumulated (used targets, components sorption and getter pumps, tritium contaminated oils, etc.), and disposal must be in accordance with the regulations.

20.3 HIGH VOLTAGE HAZARD e higTh h voltage power supply housing should alway e connecteb s a goo o t d ground. A neutron generator uses 150 - 200 kV; therefore, care should be taken when working in its vicinity without protective covers or devices. Discharge al l unitV capacitorH se beforth f o se attempting power supply maintenanc- re r o e n isolatea f o e d us handl paire Th . e discharg s recommendei d ro e r fo sucd h pur-

219 poses. Ground the HV terminal with the same rod if the generator doesn't operate e 1-2 Th poweV 0! k wit HV hr supplien sorce io d extractio f an o ss d focuan n s system and ion getter pump are lethal voltages, necessitating extreme caution, when terminaV pumn H workin io e r pth o l suppliesn go .

20.4 IMPLOSION HAZARD The neutron generator is basically a vacuum vessel and presents the same implosion hazar o accidentat d l breakag picturV s doea T e a s e tube.

20.5 PRESSURE HAZARD Some neutron generators have e abouhigth - 2 hbarisolation t o SR s s ga n voltage and in the HV power supplies. The pressure must be reduced to atmospheric before openin e pressurth g e dome. Removin e domth g e without reducin e pressurth g e could caus a serioue s accident. Similarly e pressurth , e use n pneumatii d c systems should be reduced to atmospheric pressure before opening the system. The SF^ tanks should be opened in well ventilated rooms. Handle with care the hermetic e SAMEunitth f (Felicio sV SH ) generators!

20.6 FIRE HAZARD e deuteriud Th soman es mga transforme r e oilflammablear s . Avoid electric sparks and avoid open fire while opening V^ g38 vessels and oil transforms. The organic solvents (acetone, benzene, alcohol, n-hexane, etc.) used in cleaning the neutron generator components are flammable; use them with caution. In the case of n accidentaa e neutroth l firn i en generator e halcarbous l n dioxide fire extinguishers. Water, foam or powder extinguishers can cause fatal damage in the neutron generator e relateth d dan s instrumentation.

220 21. CONSTRUCTION OF A NEUTRON GENERATOR LABORATORY

21.1 CONSTRUCTION DETAILS Many factors mus e takeb t n into accoun n establishini t a gneutro n generator laboratory; the most important topics are listed (based on the proposal for the local a developinstaf n i f g country) below: (1) The thickness of the biological shielding can be calculated. In general, around the source at a distance of at least 2 m an additional 1.5 m wall of poured 3 g/cs neededconcreti 2. m) = p ( .e (2) The inner surface of the neutron generator room (NGR) must be covered with washabl r o plasti l oi ce paint A thi. n plastic fois i l advise r coverinfo d g surface constructed of concrete blocks. (3) The door between the NGR and the control room must be airtight and protect e measurinth g laboratories from contaminatio y b n radioactive gaseso T . increase the efficiency of the ventilation in the target storage glove box and target d controareas an e dooth , lR r roobetweeNG m e constructeth n f o d paraffin or polyethylene blocks should be covered by plastic foils. (4) A crane with a 2.0 - 2.5 t capacity is recommended for transporting large parts and heavy equipment (e.g. source container, lead spectrometer, subcritical tank, etc.). ) (5 Three chimney e needear s r ventilationfo d : e closeabouR th n NG d5 i timet r r houo refrespe t ai so ensur t re e th hOn e ) a that the concentration of tritium does not exceed the value of 5 x 10"6 fiCi/m3. b) One to ventilate the glove-box when it is in use: this can be done by placing a small fan in the appropriate chimney. c) One (a 10 cm diameter tube) which can be connected to the exhaust of the forevacuum pump. At the top of this chimney the tritium content should e controlleb d continuously. (6) Fresh air for ventilation can be supplied through the measuring room by a tube system (e.g. usin e channelth g r cable fo d s an pneumatis c transfer). (7) In humid climates dehumidifiers are needed: at least one in the NGR and one e measurinith n g room. (8) To avoid noise and contamination in the NGR, the compressor and the ventila- tors shoul e placed b da separatn i e room e compresseTh . r ai requiremend t h dependin/ varie e m dimension th s0 1 between o gd d distancan e an sth 4 n f o e •3 transfer system tubes. Working pressure of the network must not be less than 6 bar. A buffering air-tank of about 100 1 should be placed close to the pneumatic transfer system.

221 (9) A pipeline for cooling water in the NGR with a flow-rate of about 100 1/h is needed o f outlea I thern canalize. o s t i et d water network, forced water circulation is necessary through a cooling system. Another pipeline at normal pressure should be introduced into the measuring center. The same pipeline can be used for the whole neutron generator laboratory (NGL). (10) Neutron generators n generali , e ar operate, d controllean d a centray b d l Generator Control Unit (control desk )e locate b whica distancn t a ca df h o e 10-15 m far from the machine. The power requirement of the whole labotory is t morincludinA no ekV thae 0 1 operatioth gn a compressof o n r which needs about 4 kVA, 220/380 V, 3-phase. For the generator 3-phase, 5 kVA, 220/380V, 50 Hz are required. Considering the further development of a final NGL as well as the energy consumption of the measuring and operating equipment, it is advisabl o t desige e supportinth n g cable d an transformers r fo abous t 25 kVA. Every room in the NGL should be supplied with a one- and three-phase network system. The cables from the control room to the NGR can be placed either along the wall or through channels in the wall. For electric cables, about four 5-c ma di channels shoul e b dconstructed f steeo l e tubeth n i s wall close to the ceiling of the control room. The unused channels can be closed with iron bars. (11) For the pneumatic transfer system, 5x3 cm and 8x4 cm channels should be constructe y steeb d l profil e e walle th unusetube Th .n i s d channele clob n - ca s d wit n se iroa h n plug. Althoug e pneumatith h c tubes coul n froe generu dth m - e measurinth rato o t r g rooms alon s e i nevertheleswallth gt i , s advisablo t e construc e proposeth t d channels. s i necessar t (12I ) o r t construcstoragyfo R e tNG leae d th e wal f th boxeo l n i s of radioactive source d tritiuan s m targets e boxeTh . s should x measurm c 0 3 e e completelb d an m yc surroundethic0 m 3 c x k5 layea m c f y leado b d0 r 3 . (13) Warning signd an lighs t beacons shoul e installeb e dcontro th n i dl rooo mt warn of high voltage hazards. (14) The generation of neutrons should be indicated by a warning light or rotating beacon connected to the HV on switch and to the neutron flux monitor. interlocV H n (15A k) system shouldoorR NG e installe. b de th t a d (16) Water-safe lamp d plugan s s e lightinmusth e used b r electricat an fo gd l connections in the NGR. 2 (17) The floor of the NGR must be able to carry about 20 t/m and have an outlet tube to a water sink. n (18tropicaI ) l climater conditioninai n a s g syste s i mrequire l roomal f o n i ds the NGL. (19wateA ) r tank wit a pumh s needepi r closefo d d circuit cooling.

222 (20) It is advisable to construct a channel for water pipelines and cables with a removable cover plate belo e floowth r fro e R contromth NG e l th roo mo t 15 cm deep and 20 cm wide. (21combineA ) d ventilatio d coolinan n g syste s mi recommende e NGRth r fo .d (22) Further improvement of a neutron generator would make possible to use a TOP spectrometer; therefore, a channel of about 10 cm dia between the NGR and e controth l room shoul e constructedb d .

21.2 WORKSHOPS It is recommended to complete the neutron generator laboratory with a mechanical workshop containin e th gusua l locksmith's tools, p drilltablto ,e lathe d millin(50an - 0100 ) g 0mm machines. Arc, plasm r acetyleno a e weldin s somei g - times useful t thesbu , e facilitie e usuallar s y availabl t a othee e r th unit t a s e laboratoryth sit f o e . It is recommended that frequently needed materials be stored in the workshop. s followsa Thes e ar e: steel, aluminium, brass, plasti d an Perspec o t x 5 rod( s 50 mm dia), screws, nuts; steel, aluminium and bakelite or Perspex sheets up to a . Standarthicknesmm 0 1 d f o s profile r stand fo sd holderan s e alsar so recommended. e usuaTh l toold instrumentan s e electronicth n i s s workshop are: pliers, soldering irons, multimeters, oscilloscopes s i als t oI . advisabl o t complete e th e list with some high voltage meters (as used for TV repair) and some home-made high frequency test instruments and tools (e.g. resonant circuit with incandescent lamp, etc). The short-lived components of the neutron generator or their equivalents shoul e procuredb d e usuaTh . l stoc f activo kd passivan e e components e recommendear F componentH d an d insulatino complett dV an sH e th ge materials, such as epoxy resin. s laboratorieA - especiall se developinth n i y t gno usuall worlo d y- d have staff skille vacuun i d m technology d sincan , e vacuum material d componentan s s e difficular o obtaint t e persoth , chargn e i nneutro th f o e n generator shoule b d carefu o t lmaintai e supplth nf frequentl o y y needed component d materialsan s . e diffusioth r d mechanicafo Sparan n l oi e l pumps, silicon high vacuum grease solvents, the most important O-rings, rubber sheets, spare Pirani and other vacuum gauges should be kept in store. e store-rooTh m attache e mechanicath o t d d electronian l c workshop shoule b d clead equippean n d with shelve s i usualld cabinets an t i s ys A .locked e storeth , - room is the best place to store radioactive materials, targets and radioactive litter n lockei , d d bincabinetsan s e e operatoneutroth Th . f no r generator

223 - usually a technician - is a skilled worker, so he may use both the workshop e e anradioactivstore th e th dkey f Th o .s e material storage cabinet d an ventis - lated glove box in this store-room must not be available to the general personnel e workshop ith ne th neutro t bu , n generator operator t shoulintge oe ablb do t e the workshops at any time, especially during the long irradiations (at night), to perform maintenanc d repairs an r etritiu fo glovA .x mbo e target e b storag y ma e constructed by experienced local personnel.

21.3 LABORATOR BOOG YLO K The operation of neutron generators, like all complicated equipment, requires regular maintenance dutie d an regulas r exchang e th short-live f o e d components. The operator of the neutron generator should record the working parameters in a log book. This log book should show the time of every operation and record the exchang f o short-livee d components (e.g. target n io sourc, e balloon, quartz sleeve). The neutron generator log book is a good basis for planning maintenance and repai s wel a rs operation a l .

224 A typical laboratory log sheet:

Date: Supported by: Project: Project leader: Operator: Beam branch: Direct: Deflected: Target: None Deuterium: Tritium: Others: Dose durin e operationgth : Vacuum: Gauge No.l: Gauge No.2: Gauge No3.:

10,-3 mbar 10~6 mbar Starting time: hours minutes Ion source gas: (relative unit) High frequency: V mA Magnet current: (relativ) e A uni r o t Extraction: V mA (relative unit) Focus: V mA (relative unit) Accelerating high voltage: kV mA Deflecting magnet: A (relative unit) Quadrupole lenses: A A A A r relativo ( e unit) s TargeA (directm t )current: mA (deflected) Target and target holder activation: (relative unit) Beam curren n isolateo t d slit: mA Monitor counter: counts Remarks:

NEXT left 225 REFERENCES

. CsikaiJ ] [I , Handboo f Faso k t Neutron Generators PresC CR s, Inc., Boca Raton, Florida (1987) Vol MI [2] S.S. Nargowalla, E.P Przybylowicz, Activation Analysis with Neutron Generators, John Wile d SonsYorw an y Ne k, (1973) f . o Smal[3J Csikai] e l Us ,Neutro n Generator n i Sciencs d Technologyan e , Atomic Energy Review 11 (1973) 415 [4] H. Liskien, A. Paulsen, Nuclear Data Tables 1 (1973) 569 [5] J.D. Seagrave, E.R. Graves, S.J. Hipwood, C.J. McDole, D(dn) 3d T(dn)Han e 4H Neutron Source Handbook, LAMS-2162, Los Alamos Scientific Lab., NM (1958) [6] J. Csikai, Zs. Lantos, Cs.M. Buczk6, IAEA-TECDOC-410, Vienna (1987) p.296 [7] V.E. Lewis K.J. Zieba, Nucl. Instr. Meth.174 (1980) p.141 [8. CsikaiJ ] Procn i , . Int. Conf. Nuclear Dat r Sciencfo a d Technologyan e , Antwerp, Reidel, Dordrecht (1983) p.414 . Wagner[9M ] . WinklerG , . VonachH , , Cs.M. Buczko . J Csikai, , Ann. Nucl. Energy 16 (1989) p.623 [10] A. Pavlik, G. Winkler, Report INDC(AUS)-011/LI, IAEA, Vienna (1986) [II] K. Kudo, T. Kinoshito, Fusion Energy and Design 10 (1989) p. 145 [12] J. Csikai, Cs.M. Buczko, R. Pepelnik, H.M. Agrawal, Ann. Nucl. Energy, 18 (1991) p.l . B6dyZ [13 ] Handboon i , Nuclean o k r Activation Data, Technical Report Series, No.273, IAEA, Vienna (1987) p.29 [14] Cs.M. Buczko . J Csikai, . ChimoyeT , , B.W. Jimba, Proc. Conf. Nuclear Data for Science and Technology, Jiilich, Springer Verlag, Berlin (1992) p.656 [15] J. Csikai, R. Raics, in Progress Report (HUN), IAEA, Vienna (1991) [16] K. Kudo, N. Takeda, Matiullah, A. Fukada, Proc. Conf. Nuclear Data for Science and Technology, Jiilich, Springer Verlag, Berlin (1992) p.533 [17] N. Jarmie, R.E. Brown, Nucl. Instr. Meth. D10/11 (1985) p.405 Energw Lo Usew r yNe fo s[18Accelerators ] , National Academ f Scienceso y , Washington, D.C. 1968 . FreyeT [19 ]. Lorenz-WirzbaA , . CleffB , , H.P. Trautvetter . RolfsC , , Z.Physik, A281 (1977) p.211 . J Szabo[20] . RaicsP , n Proci , . XlX-th Int. Symp n Nucleao . r Physics, ZfK-733, Dresden, 1989, p.266 [21] G.M. Griffiths, E.A. Larson, L.P. Robertson, Can. J. Phys. A40 (1962) p.402 . RolfsC [22 . ]Rodney S , . DurranceS , . WinklerH , , Nucl.Phys. A240 (1975) p.221

227 [23] X-ray and Gamma-Ray Standards for Detector Calibration, IAEA-TECDOC-619, Vienna (1991) [24] Z.E. Switkowski, P.D. Parker, Nucí. Instr. Meth (19753 1 . ) p.263 [25] D.F. Hebbard, J.L. Voglh, Nucl.Phys. 21 (1960) p.652 [26] P.M. Penning, J.H.A. Moubis, Physica 4 (1937) p.1190 [27] C.W. Elenga, O. Reifenschweiler, Proc.Symp.Pulsed Neutron Research, Karlsruhe, May 10-14, 1965, Vol.2, IAEA, Vienna (1965) p.609 [28] C.M. Gordon, C.W. Peters, Nucl.Geophys. 2 (1988) p. 123 [29] SODERN Data Sheet Sealen o s d Tube Neutron Generators, France (1990) [30] C.M. Gordon, C.W. Peters, T.K. Olson, Conf. Accuracy in Trace Analysis, Gaithersburg, MD,USA, 28 Sept. - 1 Oct. 1988 [31. SztaricskaiT ] , ATOMKI Kozl (19802 2 . ) p.4 Hungariann 7(i ) [32] T. Sztaricskai, Neutron Generators, Proc. IAEA Advisory Group Meeting on Small Accelerators June5 1- , , 1992, Debrece e publishedb o (t n ) [33. SztaricskaiT ] , Proc. XIV-th Int. Symp. Interactio f Faso n t Neutrons with Nuclei, Nov. 19-23, 1984, Gaussig, ZfK-652, INDC/GDR/-40/G0 1 . p , [34] H.H. Barshall, Proc.Int. Conf. Fast Neutron Physics, May 26-31, 1986, Dubrovnik, Yugoslavia, p. 188 [35] J.C. Crawford . BauerW , , ANL-CTR-75 , Argonn4 , e Nat. Lab. (1975) p.227 [36. MaekawaH ] , JAERI-M-83-219, Japan Atomic Energy Res. Inst. (1983) [37] CD. Moak, H. Reese, W.M. Good, Nucleonics, 9 (3) (1951) p. 18 [38] Z. Szabó, Nucl.Instr.Meth 78 (1970) p.199 [39. PásztorE ] , Isotopenpraxi (19653 s ) p.259 [40] K.W. Ehlers, B.F. Gavin, E.L. Hubbard, Nucí. Instr. Meth (19632 2 . ) p.87 [41] J.L. Nagy, Nucí. Instr. Meth 2 (19653 . ) p.229 [42] Sun Biehe, Wang Fulin, Chen Quin, Proc. XIV-th Int.Symp. Interaction of Fast Neutrons with Nuclei, Gaussig, 19-23 Nov, 1984, ZfK-652 [43] D.D Armstrong et al., Nucí. Instr. Meth. 145 (1977) -127 [44] L. Vályi, Atom and Ion Sources, John Wiley & Sons, London (1977) [45] P.C. Thonemann et al., Proc. Phys. Soc., London, 61 (1948) p.483 [46 . WehnerG ] , Phys. Rev (19543 .9 ) p.653 [47] D.P. Smith, Hydrogen in Metal, Chicago University Press, Chicago (1948) . [48JostW ] , Diffusio n Solidsi n , Liquid d Gasesan s , Academic Press, New York (1960) [49] Manua f Needlo l e Valve EVNO1H1, Balzer , LiechtensteiAG s n [50] Instruction Manual of TOSHIBA NT-200 Neutron Generator, Tokyo [51] Instruction Manual of KFKI NA-4C Neutron Generator, Budapest [52] T. Sztaricskai, Remote control of Pd leak-electrolyzer (unpublished) [53] S. Bederka, J. Král, Nucí. Instr. Meth. 145 (1977) p.441

228 [54] Instruction Manual of J-25 Neutron Generator, SAMES, Grenoble [55] Instruction Manua NeutroT f o l n Generator, SAMES, Grenoble [56] J. Pivarc, Proc. XIV-th Int Symp. Interaction of Fast Neutrons with Nuclei, Nov. 19-23 1984, Gaussig, ZfK 652, p. 16 [57] P. Eckstein et al., Proc. XlV-th Int. Symp. Interaction of Fast Neutrons with Nuclei, Nov. 19-23, 1984, Gaussig 652K Zf ,, p.24 [58] M. Pothig, P. Eckstein (personal communication) . SandorM [59 ] . BolyanD , . SztaricskaT , e publishedb o (t i ) [60] Product Catalogue, BALZERS AG, Lichtenstein [61] Vacuum Technology, its Foundation, Formulae and Tables, Leybold-Hereus GmbH, Germany [62] High Vacuum Products, EDWARDS, Crawley, England (1984) . HarrisS . N [63,] Vacuum Technology, EDWARDS High Vacuum, Crawley, Sussex, England [64] High Vacuum Technology, Vacuum Components, ALCATEL, France (1980) [65] High Vacuum Components, TUNGSRAM, Budapest, Hungary . DushmanS [66] , Scientific Foundatio f Vacuuo n m Technique, John Wiled an y Sons, New York (1968) [67] A. Chambers, R.K. Fitch, B.S.Halliday, Basic Vacuum Technology, Adam Hilger Yorw , BristoNe k d (1980an l ) [68] J.F.O'Hanlon, A User's Guide to Vacuum Technology, Wiley Interscience, New York (1975) Ardennen vo . M [69], Tabelle r Elektronenphysiknde , lonenphysi d Uebermikkun - roskopie, VEB Deutscher Verlag der Wissenschaften, Berlin (1956) (in German) [70] J. Kliman (personal communication) [71] V.M. Kellmann, S.J. Javor, Eletronoptics, Akademiai Kiad6, Budapest (1965 Hungariann (i ) ) [72] M.D. Gabovich, Physic d Technologan s n SourcesIo f yo , Atomizdat, Moscow (1974 Russiann (i ) ) [73] M.R. Shubaly, M.S. de Jong, IEEE Trans. Nucl. Sci. NS-30 (1983) p.1399 [74] J.B. Hurst, M. Roche, Acta Physica Slovaca 30 (1980) p. 137 . GalejsA [75] , P.H. Rose, Focusin f Chargego d Particles, Academic Press, New York (1967) p. 297 [76] B. Gyarmati, E. Koltay, Nucl. Instr. Meth. 66 (1969) p.253 [77] G. Liebmann, Proc. Phys. Soc. London Sect. B-62, part 4 (1949) p.352 [78] J.B. Hours t al.e t , Nucl. Instr. Meth 5 (19779 14 .1 . p ) . KoltaE [79] t al.e y , ATOMKI Kozl (19802 2 . 5 15 ) [80] J.B. Hourst . RocheM , , Nucl. Instr. Meth (19712 9 . ) p.589

229 [81] M.R. Cleland, B.P. Offermann, Nucí. Instr. Meth. 145 (1977) 41 [82] J.C. Davi t al.e s , IEEE Trans. Nucí. Sci. NS-2 ) (1979(3 6 ) p.3058 [83] R. Booth et al., Nucí. Instr. Meth. 145 (1977) p.25 [84] K. Sumita et al., Proc. 12th Symp. Fusion Technology, Vol!, Pergamon Press Yorw Ne k, 5 (198267 . p ) [85 . KissA ]. KoltayE , . SzabóGy , , Nucí. Instr. Meth 7 (197411 . ) p.325 [86] A. Kiss, E. Koltay, Gy. Szabó, Nucí. Instr. Meth. 212 (1983) p.81 [87] J.B. England, Techniques in Nuclear Structure Physics, MacMillan (1974) [88] R.G. Wilson, G.R. Brewer Beamsn Io , , Wiley Interscience, London (1973) [89 . HumphriesS ] , Principle f Chargeo s d Particle Acceleration, Wiley Interscience Yorw Ne k, (1978) [90 . SztaricskaiT ] , Regulated analyzin gs (unpublishedmagneNG r fo t ) [91] L. Csánky (personal communication) [92 . WilliamS ] t al.e s , Nucí. Instr. Meth 5 (197118 . ) p.353 [93] T. Sztaricskai, Quadrupole doublet for Thailand (unpublished) [94] N.J. Felici, Direct Curren (19531 2 t 12 . p ) [95] E.G. Komar, Fundamental f Acceleratoo s r Technique, Atomizdat, Moscow (1975), (in Russian) [96] E. Kuffel, W.S. Zaeng, High Voltage Engineering, Pergamon Press (1972) [97] Handbuch der Elektrotechnik, Siemens, Müchen (1971), (in German) [98] J.D. Cockcroft, E.T.S. Wallon, Proc. Roy. Soc. Ser.A-129 (1930) p.477 [99] G. Peto, Izotóptechnika 16 (1973) p.561 ( in Hungarian ) [100] J. Takács, IEEE Trans.Nucl.Sci. NS-12 (3) (1965) p.980 [101 . SztaricskaT ] t al.e i , IAEA Int.Data Committee, INDC/GDR/-21/6/Spec, ZfK-476 (1982) p.95 [102] D.G. Shurley et al, Nucí. Instr. Meth. 187 (1981) p.347 [103 . SztaricskaT ] i (unpublished) [104] C.S. Zaidins, Nucí. Instr. Meth 0 (19745 12 .12 . p ) [105] D.D. Cossutta, Proc.Conf f Smalo e l Us .Accelerator n Technicai s l Research, Oak Ridge, Tennessee, CONF-700322 (1970) [106] G!. Primenko et al., Izv. Vuzov, Physica No.5 (1985) p. 17 (in Russian) [107] Radiation Protection Procedures, Safety Series No.38, IAEA, Vienna (1973) [108] R.F. Boggs, Radiological Safety Aspect e Operatioth f o s f Neutrono n Generators, Safety Series No.42, IAEA, Vienna (1976) [109 . MaekawH ] , JAERI-al t e a M 83-219, Japan Atomic Energy Res. Inst. (1983) [110] M. Buttig, INDC/GDR/-40/6, ZfK-562 (1985) p.56 [111] Instruction Manual of KAMAN-711 Sealed Tube Neutron Generator, Colorado Springs, USA [112] T. Sztaricskai, B. Yotsombat (to be published)

230 [113. SztaricskaiT ] , ATOMKI Kozl (19713 1 . ) p.117 . [114Pet6G ] , Pneumatic Sample Transfer Syste r Neutromfo n Generators (to be published) [115. SztaricskaT ] t al.e i , ATOMKI Kozl (19855 7 2 . 10 p ) [116] B.V. Anufrienk t al.e o , FEI-307, Fiziko-Energ. Inst., Obninsk (1971) (in Russian) . [117ChirapatpimoN ] t al.e l , INDC(TAI)-004/GI, IAEA, Vienna (1986) . [118VilaithonT ] t e al.g , Proc. Conf. Nuclear Dat r Sciencfo a d Technologyan e , Jiilich, Springer Verlag, Berlin (1992) p.486 [119] S. Singkarat et al., Proc. Conf. Nuclear Data for Science and Technology, Myto, Japan (1988) [120] J.C. Robertson, K.J. Zieba, Nucl. Instr. Meth. 45 (1960) p.179 [121] C.M. Herbach et al., Techn. Univ. Dresden Report 05-06-85 (1985) [122] Neutron physic Hungarian)n (i s , Editors . QuittnerP . Kisd D : an s , Akademiai Kiado, Budapest (1971) [123] G.F. Knoll, Radiation Detection and Measurements, John Wiley and Sons, New York (1981) [124] S. Nagy, Thesis, Debrecen (1976) . [125SztaricskaiT ] , Troubleshootin d Maintenancan g f Neutroo e n Generators, Proc. IAEA Advisory Group Meetin Smaln go l Accelerators, 1-5 June 1992, Debrecen, Hungar e publishedb o (t y )

SIEXT PAQI(S) 1 lef23 t BLAN I K ANNEXA

LIS MANUFACTURERF TO COMPONEND SAN T DEALERS

t e responsiblauthorno Th e correctnese th ar r s fo e f thio s s list. Telephone (Fax, Telex) and postal code numbers may have changed and small firms may have closed. Reader f o this s Manua e advisear l o t contacd t local representativef o s multinational manufacturer r currenfo s t addresse d up-to-datan s e information o n the available products. The multinational manufacturers are marked in the list with an asterisk (*).

LEYBOLD Bonnerstrasse 489 D-5000 Kol1 n5 P.O.Box 510760 GERMANY Tel:(0221) 347-0 Telex: 888 481-20 Fax: (0221) 3471250 Products: vacuum components, pumps, materials, systems, technologies

EDWARDS High Vacuum Crawley West Sussex RH1W 21 0 ENGLAND Phone: 0293 28844 3 Telex12 7 8 : Fax: 0293 33453 Products: components, materials, pumps, systems, technologies

BALZER* S FL-9496 Balzers LIECHTENSTEIN Tel 075 44111 8 Telex78 9 88 : 2 76 2 4 Fax5 07 : Products: components, materials, pumps, systems, technologies

VARIAN * 12 Hartwell Avenue Lexington MA 02173 USA Phone: (617) 273 6146 Telex: 710 321-0019 Fax: (617) 273 6150 Products: pumps, components, materials, systems, technologies

233 PFEIFFERS * Arthur Pfeiffer Vacuumtechnik P.O.Box 1280 D-6334 Assiar GERMANY Phone:(06441) 802-0 Telex: 483 859 Fax: (06441) 802-202 Products: turbomolecular pumps, other components (member of BALZERS group)

VAT CH-9469 Haag SWITZERLAND Phone: (085) 70161 2 Telex16 5 85 : Fax: (085) 74 830 Products: Vacuum valves, gate valves ALCATEL* 41 rue Perkier F-92120 Montrose FRANCE Tel: (1657) 1100 1 Telex43 0 27 : Products: components, materials, systems, pumps, technologies HUNTINGTON LABORATORIES 1040 L'Avenida, Mountain View CA-94043 California USA Tel: (415 3324 96 )3 Fax: (415) 964 6153 Product: vacuum components MAC VACUUM 23842 Cabot Boulevard, Hayward CA-94545-1651 California USA Phone: (415) 887-6100 Telex: 910 383 2023 Fax: (415) 887 0626 Products: vacuum components, materials, systems SHADIER SCIENTIFIC 2976 Arf Avenue P.O.Box 57287 Hayward CA-94545 California USA Tel: (415) 783 0552 Fax:(415) 783 7245 Products: vacuum materials, components, systems

234 MEGAVOLT Cornhill Ilminster Somerset H OA TA19 ENGLAND Tel: (44) 460 57 458 Fax: (44) 460 54 972 Products/services: accelerator tubesn sourcesio , , beam handling devices, target assemblie ( repairs , maintenance, second hand accelerator component) s TECHNABEXPORT Starimonetni pereulo6 2 k 1091800 - Moscow RUSSIA Tel: 2392-885 Fax: 70-952-302-638 TelexU : S 4132 E 8TS Products: neutron generators, tritium and deuterium targets Contact person . Boroduli . BasovMr : Mr d , an nMr.Grigorje v (targets) ATOMKI Ber r 18/nte C H-4001 Debrecen P.O.Bo1 x5 HUNGARY Tel: (36) 52 317-266 0 Telex21 2 7 : Telefax: (36) 52-316-181 Products: Diffusion pumps, vacuum meters, components, quadrupole mass spectrometers, for leak testing, accelerator components, acceleration tubes, lenses, magnets, etc. Contact person: Mr. S. Bohatka, Mr.E. Koltay TUNGSRAM H-1125 Budapest P.O.Bo7 x Szilag6 2 u y HUNGARY Tel: (36)-l-169 2800 or(36)-l-169 3800 8 45 5 22 r o 8 Telex05 5 22 : Fax: (36)-l-169 286 r (36)-l-168o 9 1779 Products: mechanical pumps, components componentsV UH , , vacuum materials

NATIONAL ELECTROSTATIC CORPORATION Graber Rd P.O.Bo0 x31 Middleton Wisconsin 53562 USA Phone:(608) 831-7600 Telex 5436 2 : 0 Fax:(415 7243 78 )5 Products: accelerators, accelerator components, RF and duoplasmatron ion sources, acceleration tubes, beam line components, related equipment liks ga e leaks

235 KFKI Central Research Institut r Physicfo e s Dept. Material Sciences Budapest 114 P.O.Box 49 H-1525 HUNGARY Phone:(36)-l-1166 540 Telex: 224722 Fax: (36)-l-155 3894 Products: neutron generators, components, neutron generator laboratories Contact person: Dr Istvan KrafcsiK IRELEC (formerly AIDd previouslan , y SAMES) 20 rue du Tour de 1'Eau Postal adress: BP 316-38407 ST MARTIN D'HERES Cedex FRANCE Tel: 76 44 12 96 Fax: 76-63 19 68 Telex: 980167 Contact person . RechateMr : n Products: electrostatic HV (Felici) generators, neutron generators, neutron generator components, accessories AMERSHAM INTERNATIONAL Amersham Laboratories, White Lion Road Amersham, Buckingshamshire HP7 9LL ENGLAND Tel: (44) 494 543 488 Fax2 24 : 3 (4454 ) Telex: 83 141 Products: tritiu d deuteriuman m target r neutrofo s n generators, radioisotopes Contact person: Mr.Keit . FletcheL h r (subsidiaries in North and South America, Asia and Europe) NUKEM Indus triestrass3 1 e D-8755 Alzenau P.O.Box 1313 GERMANY Tel:(49) 6023 500 0 Fax: (49) 6023 500 222 Telex: 418 4123 Products: tritium technology, tritium targets, glove boxes MULTIVOLT 26 Loppets Rd Crawley, Sussex W RH15D 0 ENGLAND Tel:(44 22633 29 ) 0 Fax: (44)-273-747 100 Products: neutron generators,components, SAMES neutron generator components, Contact person . D.CossuttMr : a

236 KAMAN NUCLEAR 1500 Garde f Goodo n . Rd s Colorado Springs, Colorado 80933 P.O.Box 7463 Tel:(303) 599 1500 2 Telex41 2 45 : USA Products: neutron generators, sealed tube neutron generators,neutron generator components, related equipment Contact person: Mr. Frey MF PHYSICS 4720 Forge Rd. Suite 112 Colorado Springs Colorado 80907-3549 USA Tel: 719-598 9549 Fax: 719-598 2599 Products: KAMAN components, equipments, sealed tube neutron generators SGN Societe generate pou s techniquele r s nouvelles s Heronsde e 1ru , Montign e Bretonneuxl y , F-78182 Saint-Quentin en Yvelex Cedex FRANCE Tel: (33 3051 ) 8 6814 Fax: (33) 1 3058 6852 Telex: 698316 Products: neutron generators, neutron detectors Contact person. VigreuxB r M : , director INTERATOM Friedrich Ebert Strasse D-5060 Bergisch-Gladbach GERMANY Tel: (49) 2200 484 Fax: (495 ) 04 2200 3 484 7 Telex85 7 88 : Pruducts: accelerators, neutron generators, Contact person . K.HMr : . Weyers IMAGING AND SENSING TECHNOLOGY Westinghouse Circle Horsehead NY 1484A 5US Tel: (1) 607 796 3400 Fax: (1) 607 796 3279 Telex: 490 998 9073 Products: neutron generators, neutron detectors Contact person . WilieMr : m Todt

237 EFREMOV Scientific Research Institute of Electrophysical Apparatus 189631 St-Peterburg, Russia Tel.:(7)-81 5655 7915 26 8 226 r 5 o Fax: (7)-812 265 7974 or 463 9812 E-mail: [email protected] Products: Neutron generators, accelerators and their components, upgrading components Contact person: Nicolay Tolstun SODERN Societe anonyme d'etude t realisatione s s nucleares avenu1 e Descartes 94451 Limeil-Brevannes Cedex FRANCE Tel: 33(1) 45 69 96 00 Telex: 270 322 2 0 Fax4 1 : 9 33(16 5 4 ) Products: sealed tube neutron generator r boreholfo s e logging, related equipment Contact person: Mr Serge Chezeau PHILIPS Eindhoven P.O.Box 5600 90050 NETHERLAND Tel: 40 783 749 Products: sealed tube neutron generators CLASSMAN Rout # e (EAS) Salem Industrial Park P.O.Box 555 Whitehouse Station N.J. 08889 USA Tel:(201) 534 9007 2830 Telex48 9 0 71 : Products: High voltage power supplies (medium frequency 3-400W) 1-250 mA HAEFELY Lehenmattstrase 353 Basel CH-4028 SWITZERLAND 1 11 Tel 5 : 53 (41 1 6 ) Telex: 62469 Products: high voltage test equipments, insulating transformers, mains frequency HV power supplies, sealed tube neutron generators

238 TECHNICAL UNIVERSITY BUDAPEST Institute for Automation H-1117 Budapest HUNGARY Tel:(36)-l-166 4527 Fax:(36)-l-166 6808 Telex:221 593 Products: high voltage and high current stabilized power supplies for acceleratorequipmeny ra X d an ts Contact person: Dr. I. Ipsits PULSE ELECTRONIC ENGINEERING 3-15,Tatekawa 4-chome, ida-kum Su , Tokyo0 ,13 JAPAN Tel: (03) 633 6101 Fax: (03) 634 0636 Products: high voltag C powerD e s supplies, stabilized constant current power supplies, special HV equipment INSTITUT EXPERIMENTAF EO L PHYSICS KOSSUTH UNIVERSITY H-4001 Debrecen Bern ter 18/A HUNGARY 2 Tel22 :5 (3641 2 )5 7 Fax08 : 5 (3631 2 )5 univ0 Telex20 h k2 7 : Products: accessories, components for neutron generators and related equipment ( target holders, quadrupole lenses, associated particle target heads, wobbling target holders sourcn io , e components) neutron monitor d samplan s e holders, pneumatic sample transfer systems for neutron generators, design and manufactur- ing services for neutron generator laboratories Contact person : e InstitutHeath f do e

ALFAX Malmo Lundanvagen 143 4 2 2 21 SWEDEN Tel: 040-180 900 Telex: 32504 Products: heavy water, compressed s cylindedeuteriuga n i rs mga INSTITUTE OF NUCLEAR RESEARCH Prospekt Nauk9 11 i 252650 Kiev-28 UKRAINE Products: tritiu d deuteriuman m targets Contact person: Mr Kolomentzev

239 RADIOISOTOPE CENTRE POLAND Foreign trade office POLATOM Ottwock-Swierk 05-400 POLAND Tel5 : 43 4828 279 Telex: 812202 Fax: 4822 797 381 Products: deuteriu d tritiuman m targets, deuterium gas. CHINA INSTITUTE OF ATOMIC ENERGY P.O.Box:275 Beijing 102413 PEOPLES REPUBLI CHINF CO A Tel: Beijing, (86) 1 935 7312 Telex: 222 373 IAE CN Fax: 7005 (8693 3 )1 Products: radiochemicals, labelled compounds, targets for neutron generators Contact person: Mr Yu-shan Wang (neutron generator targets) WALLIS HIVOLT Dominioy nWa Worthing, Sussex BN1W 8N 4 ENGLAND Tel: (44)-903-211 241 Fax: (44)-903-207 801 Telex: 877 112 Products: medium frequency high voltage power supplies up to 200 kV (500-2000W) MARCONI AVIONICS Neutron Division Elstrey Wa e Borehamwood Hertfordshire WD6 1RX ENGLAND Tel: (44) 1-953 2030 7 Telex77 2 2 : Product: sealed tube neutron generators

240 ANNEX B

TROUBLESHOOTING FLOW CHART FOR NEUTRON GENERATORS WITH RF ION SOURCE

START YES YES YES YES YES YES S YE FINISH

color bright pink ">

Chec r forevacuukfo m targed ol N r oto j-»jChang targee eth [ t and high vacuum utron production components check the circuit of the meter NO Check the forevacuum Check the electric Check the circuits HV power supply higd an h vacuum pumps cooling water supply

Check the voltage source components source components iA x•** •* HI

Check the D_ tank ANNEX C

TROUBLESHOOTING FLOW CHAR SEALER TFO D TUBE NEUTRON GENERATORS

FINISH START YES YES YES YES YES YES YES

Check the interlocks d coolinan g circuits

DEAD SEALED TUBE ( CHANGE FOR A NEW ONE ) ANNEXD

TROUBLESHOOTING FLOW CHAR NEUTROR TFO N GENERATOR VACUUM SYSTEM

START NO

Is the pump switch Switcn o t i h in ON position ? YES NO Chec e forepumth k p Troubleshooting and related syste? K mO and repair YES NO Check the high vacuum Troubleshooting, pum related pan d system exchange or repair YES NO Chec e vacuukth m meters Troubleshooting ? K thee O Ar y d exchangan e

YES NO Check the parts of the Leak testing system, are they OK ? and repair

YES NO Are the seals and the Leak testing ? jointK O s and exchange

YES FINISH

GOOD SYSTEM

NEXT PA«3E(S) 245 left BLA&K CONTRIBUTORS TO DRAFTING AND REVIEW

Csikai, J. Institute of Experimental Physics, Kossuth University, Debrecen, Hungary

Darsono Yogyakarta Nuclear Research Centre Indonesian Atomic Energy Commission, Yogyakarta, Indonesia

Dolnicar. J , International Atomic Energy Agency, Vienna, Austria

Li Gwang Nyong Institute of Nuclear Physics, Pyongyang, Democratic People's Republic of Korea

Molla, N.I. Atomic Energy Establishment, Dacca, Bangladesh

Raics, P.P. Institute of Experimental Physics, Kossuth University, Debrecen, Hungary

Sanchez, A.A. Nuclear Engineering Department, National Polytechnic Institute of Mexico, Mexico City, Mexico

Szegedi. S , Institute of Experimental Physics, Kossuth University, Debrecen, Hungary

Sztaricskai, T. Institut f Experimentaeo l Physics, Kossuth University, Debrecen, Hungary

Walsh, R.L. (Scientific Secretary) International Atomic Energy Agency, Vienna, Austria

247